36th Bled eConference Digital Economy and Society: The Balancing Act for Digital Innovation in Times of Instability June 25 – 28, 2023, Bled, Slovenia Conference Proceedings Editors Andreja Pucihar Mirjana Kljajić Borštnar Roger Bons Guido Ongena Marikka Heikkilä Doroteja Vidmar June 2023 Title 36th Bled eConference – Digital Economy and Society: The Balancing Act Naslov for Digital Innovation in Times of Instability Subtitle Podnaslov June 25 – 28, 2023, Bled, Slovenia, Conference Proceedings Editors Andreja Pucihar Uredniki (University of Maribor, Faculty of Organizational Sciences, Slovenia) Mirjana Kljajić Borštnar (University of Maribor, Faculty of Organizational Sciences, Slovenia) Roger Bons (Open University, Netherlands) Guido Ongena (HU University of Applied Sciences Utrecht, Netherlands) Marikka Heikkilä (University of Turku, Finland) Doroteja Vidmar (University of Maribor, Faculty of Organizational Sciences, Slovenia) Review Antragama Ewa Abbas (Delft University of Technology), Waleed Akhtar Recenzija (University of Jyväskylä), Belal Alsinglawi (Western Sydney University & Victoria University), Amir Andargoli (Swinburne University of Technology), Alenka Baggia (University of Maribor), Richard Bakkers (Avans University of Applied Sciences), Matthias Baldauf (Eastern Switzerland University of Applied Sciences), Nedo Bartels (Fraunhofer IESE), Slobodan Beliga (University of Rijeka), Matthijs Berkhout (HU University of Applied Sciences Utrecht), Shahid Bhat (Lappeenranta-Lahti University of Technology), Roger Bons (Open University), Christer Carlsson (Institute for Advanced Management Systems Research & Åbo Akademi University), Yiran Chen (Åbo Akademi University), Roger Clarke (Xamax Consultancy Pty Ltd), Helen Cripps (Edith Cowan University), Mirjam De Haas (HU University of Applied Sciences Utrecht), Ander De Keijzer (Avans University of Applied Sciences), Mark De Reuver (Delft University of Technology), Vanessa Dirksen (Open Universiteit), Duc Duy Nguyen (Durham University), Michael Fruhwirth (Silicon Austria Labs GmbH), Blaž Gašperlin (University of Maribor), Pia Gebbing (Constructor University gGmbH), Jaap Gordijn (Vrije Universiteit), Linda Grogorick (Technische Universität Braunschweig), Maaike Harbers (Hogeschool Rotterdam), Vil e Harkke (Åbo Akademi University), Marikka Heikkilä (University of Turku), Markku Heikkilä (Åbo Akademi University), Ralf Hoechenberger (Universität Ulm), Manfred Hofmeier (Universität der Bundeswehr München), Vincent Homburg (Erasmus University Rotterdam), Anke Hutzschenreuter (Baden-Wuerttemberg Cooperative State University), Hugo Jonker (Open Universiteit), Marija Jović (Institute of Shipping Economics and Logistics), Eeva Kettunen (University of Jyväskylä), Arif Khan (University of Oulu), Mirjana Kljajic Borstnar (University of Maribor), Rob Kusters (Open Universiteit), Christoph Lattemann (Constructor University gGmbH), Ulrike Lechner (Universität der Bundeswehr München), Sam Leewis (HU University of Applied Sciences Utrecht), Gregor Lenart (University of Maribor), Carolina Lopez-Nicolas (University of Murcia), Ivan Luković (University of Belgrade), Pasi Luukka (LUT School of Business and Management), Mahinda Mailagaha Kumbure (Lappeenranta-Lahti University of Technology), Leif Marcusson (Linnaeus University), Xochitl Margarita Cruz Pérez (Autonomous University of Queretaro), Arjen Maris (HU University of Applied Sciences Utrecht), Marjeta Marolt (University of Maribor), Sanda Martinčić Ipšić (University of Rijeka), Florian Maurer (Vorarlberg University of Applied Sciences), Thomas Mejtoft (Umea University), Jan Merkus (Open Universiteit), Angel-Luis Meroño-Cerdan (University of Murcia), Francisco-Jose Molina-Castillo (University of Murcia), Francesco Molinari (University of Rijeka), Paul Morsch (HU University of Applied Sciences), Andrei Munteanu (Åbo Akademi University), Mark Neerincx (Delft University of Technology), Vesna Novakovic (Pan-European University Apeiron), Jussi Nyrhinen (University of Jyväskylä), Guido Ongena (HU University of Applied Sciences), Tiina Paananen (University of Jyväskylä), Rafael Palacios Bustamante (BSP Business & Law School Berlin), Viktoria Pammer-Schindler (Graz University of Technology & Know-Center GmbH), Xiao Peng (University of Applied Sciences Utrecht), Rob Peters (Province of Utrecht), Gabriele Piccoli (Louisiana State University), Rok Piltaver (Institut Jožef Stefan), Dino Pitoski (University of Rijeka), Andreja Pucihar (University of Maribor), Teade Punter (Fontys University of Applied Sciences), Uroš Rajkovič (University of Maribor), Pascal Ravesteyn (HU University of Applied Sciences Utrecht), Susanne Robra-Bissantz (Technische Universität Braunschweig), Joaquin Rodriguez (Grenoble Ecole de Management), Moritz Rohde (Technische Universität Braunschweig), Emma Rosenlind (Umeå University), Anna Sell (Åbo Akademi University), Navin Sewberath Misser (HU University of Applied Sciences), Anand Sheombar (HU University of Applied Sciences), Muhammad Shuakat (Deakin University), Koen Smit (HU University of Applied Sciences Utrecht), Aletta Smits (HU University of Applied Sciences Utrecht), Ulrik Söderström (Umeå University), Ronald Spanjers (Roessingh Rehabilitation), Jan Stoklasa (Lappeenranta-Lahti University of Technology), Tomáš Talášek (Palacký University Olomouc), Verena Tiefenbeck (Friedrich- Alexander-University Erlangen-Nuremberg), Edvard Tijan (University of Rijeka), Tijs Timmerman (HU University of Applied Sciences Utrecht), Oren Tirosh (Swinburne University of Technology), Nevena Toporova (Technical University of Munich), Nalika Ulapane (Swinburne University of Technology), Martin Van den Berg (HU University of Applied Sciences), Esther Van der Stappen (Avans University of Applied Sciences), Jeroen Van Grondelle (HU University of Applied Sciences Utrecht), Willemijn Van Haeften (HU University of Applied Sciences), John Van Meerten (HU University of Applied Sciences Utrecht), Marlies Van Steenbergen (HU University of Applied Sciences Utrecht), Koen Van Turnhout (University of Applied Sciences Utrecht), Doroteja Vidmar (University of Maribor), Irina Wahlström (Åbo Akademi University), Thorsten Weber (FOM Hochschule), Borut Werber (University of Maribor), Nannan Xi (University of Vaasa), Ran Zhang (University of Applied Sciences Utrecht), Hans-Dieter Zimmermann (Eastern Switzerland University of Applied Sciences), Anja Znidarsic (University of Maribor) Technical editors Aljaž Murko Tehnična urednika (University of Maribor, Faculty of Organizational Sciences) Jan Perša (University of Maribor, University Press) Cover designer Jan Perša Oblikovanje ovitka (University of Maribor, University Press) Graphics material Grafične priloge Authors & editors, 2023 Conference 36th Bled eConference – Digital Economy and Society: The Balancing Act for Konferenca Digital Innovation in Times of Instability Location and date Kraj in datum Bled, Slovenia, June 25 –28, 2023 Organizing committee Matthias Baldauf (Eastern Switzerland University of Applied Sciences, Campus Organiazacijski odbor St. Gallen), Roger Bons (FOM Hochschule), Christer Carlsson (IAMSR/Abo Akademi University), Helen Cripps (Edith Cowan University), Blaž Gašperlin (University of Maribor, Faculty of organizational sciences), Matt Glowatz (University College Dublin), Petra Gorjanc (University of Maribor, Faculty of organizational sciences), Marikka Heikkilä (University of Turku, Finland), Mirjana Kljajić Borštnar (University of Maribor, Faculty of organizational sciences), Klara Knific (University of Maribor, Faculty of organizational sciences), Ana Kranner Porenta (University of Maribor, Faculty of organizational sciences), Gregor Lenart (University of Maribor, Faculty of organizational sciences), Xander Lub (HU University of Applied Sciences Utrecht), Maruša Luštrik (University of Maribor, Faculty of organizational sciences), Marjeta Marolt (University of Maribor, Faculty of organizational sciences), Sanda Martinčić Ipšić (University of Rijeka), Axel Minten (German Coworking Association and FOM University of Applied Sciences), Aljaž Murko (University of Maribor, Faculty of organizational sciences), Guido Ongena (HU University of Applied Sciences Utrecht), Marijn Plomp (Vrije Universiteit Amsterdam), Andreja Pucihar (University of Maribor, Faculty of organizational sciences), Pascal Ravesteyn (HU University of Applied Sciences Utrecht), Juergen Seitz (Baden-Wuerttemberg Cooperative State University), Anand Sheombar (HU University of Applied Sciences Utrecht), Koen Smit (HU University of Applied Sciences Utrecht), Esther van der Stappen (Avans University of Applied Sciences), Johan Versendaal (HU University of Applied Sciences Utrecht), Pasi Luukka (LUT School of Business and Management, Finland), Doroteja Vidmar (University of Maribor, Faculty of organizational sciences), Doug Vogel (Harbin Institute of Technology), Nilmini Wickramasinghe (Swinburne University Of Technology / Epworth Healthcare), Hans-Dieter Zimmermann (Eastern Switzerland University of Applied Sciences, Campus St. Gallen) Published by University of Maribor Založnik University Press Slomškov trg 15, 2000 Maribor, Slovenia https://press.um.si, zalozba@um.si Co-published by University of Maribor Izdajatelj Faculty of Organizational Sciences Kidričeva cesta 55a, 4000 Kranj, Slovenia http://www.fov.um.si, dekanat.fov@um.si Edition Izdaja 1ST Publication type Vrsta publikacija E-book Available at Dostopno na https://press.um.si/index.php/ump/catalog/book/786 Published Izdano Maribor, June 2023 © University of Maribor, University Press Text / Besedilo /Univerza v Mariboru, Univerzitetna založba © Authors & editors, 2023 This book is published under a Creative Commons 4.0 International licence (CC BY 4.0). This license lets others remix, tweak, and build upon your work even for commercial purposes, as long as they credit you and license their new creations under the identical terms. This license is often compared to “copyleft” free and open source software licenses. Any third-party material in this book is published under the book’s Creative Commons licence unless indicated otherwise in the credit line to the material. If you would like to reuse any third-party material not covered by the book’s Creative Commons licence, you wil need to obtain permission directly from the copyright holder. https://creativecommons.org/licenses/by /4.0/ CIP - Kataložni zapis o publikaciji Univerzitetna knjižnica Maribor 659.23:004(082)(0.034.2) BLED eConference Digital Economy and Society (36 ; 2023 ; Bled) 36th Bled eConference Digital Economy and Society [Elektronski vir] : The balancing act for digital innovation in times of instability : June 25 - 28, Bled, Slovenia : conference proceedings / editors Andreja Pucihar ... [et al.]. - 1st ed. - Maribor : Univerza v Mariboru, Univerzitetna založba, 2023 Način dostopa (URL): https://press.um.si/index.php/ump/catalog/book/786 ISBN 978-961-286-751-5 doi: 10.18690/um.feri.6.2023 COBISS.SI-ID 156313603 ISBN 978-961-286-751-5 (pdf) DOI https://doi.org/10.18690/um.fov.4.2023 Price Cena Free copie For publisher Prof. Dr. Zdravko Kačič, Odgovorna oseba založnika Rector of University of Maribor Attribution Pucihar, A. et al. (2023). 36th Bled eConference – Digital Economy and Citiranje Society: The Balancing Act for Digital Innovation in Times of Instability: June 25 – 28, 2023, Online Conference Proceedings. Maribor: University Press. doi: 10.18690/um.fov.4.2023 36TH BLED ECONFERENCE DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY JUNE 25 – 28, 2023, BLED, SLOVENIA, CONFERENCE PROCEEDINGS A. Pucihar (ed. et al.) Table of Contents CONFERENCE PROCEEDINGS PAGE Developing a Taxonomy for Revenue Models of Platform Business Models 1 Nedo Bartels, Matthias Koch, Jaap Gordijn Hypertension Self-Management Success in 2 Weeks; 3 Pilot Studies Luuk PA Simons, Bas Wielaard, Mark Neerincx 19 Data Governance Capabilities; Empirical Validation in Case Studies of Large Organisations 35 Jan R. Merkus, Remko W. Helms, Rob J. Kusters The Role of Digitalisation in Changing the Business Models in Logistics: Case of RoPax Ports Yiran Chen, Anastasia Tsvetkova, Kristel Edelman, Irina Wahlström, 49 Marikka Heikkilä, Magnus Hellström The Data Analytic Capability Wheel: An Implementation Framework for Digitalization Alinda Kokkinou, Ton van Kollenburg, Albert Mandemakers, Eric 67 Hopstaken, Jan van Elderen Unlocking the Potential of Data-Driven Business Models: An Empirical Investigation into the Role of Ecosystems and Fair Data Use 85 Marikka Heikkilä, Farham Ahmad, Jukka Heikkilä Digitalisation in RoPax ports: the Typology of Available Solutions Anastasia Tsvetkova, Yiran Chen, Irina Wahlström, Andrei-Raoul Morariu, 107 Bogdan Iancu, Magnus Hel ström Malicious Insider Threat Types – An Empirical Analysis Manfred Hofmeier, Isabelle Haunschild, Ulrike Lechner 123 Quality Dimensions for Digital Twin Maturity in the Context of Dutch Public Spatial Planning Koen Smit, Rob Peters, Chaim de Gelder, Johan Versendaal 137 ii Table of Contents. Tools and Technologies Utilized in Data-Related Positions: An Empirical Study of Job Advertisements 155 Nurlan Musazade Customers’ QR Code Usage Barriers in a Brick-and-Mortar Store: A Qualitative Study Tiina Paananen, Matilda Holkkola, Markus Makkonen, Lauri Frank, Ti na 171 Kemppainen Using the Apple Watch to Teach and Learn About Heart Rate Variability While Vacationing 189 Ondrej Mitas, Marenna van Reijsen, Nadia Carreira Oliveira A Virtual Companion for Lifelong Learning – Design Principles for Motivation, Social Learning, and Exam Preparation Pia Gebbing, Bijan Khosrawi-Rad, Timo Strohmann, Charlotte Windolf, 199 Christoph Lattemann Empirical Verification of Different Rank Dynamic Roles in Informal Hierarchies 217 Peter Vatter, Andreas Weisbeck Enhancing the Understanding of E-commerce Reviews through Aspect Extraction Techniques: A BERT-Based Approach 233 Laleh Davoodi The Role of the Service Desk for eHealth Solutions in the Digital Transformation of Slovenian Healthcare 249 Živa Rant, Hajdi Kosednar, Dalibor Stanimirović Adaptive Learning Technologies In Blended Learning Design: How Do Students and Teachers Use This Technology in Practice? 263 Richard Bakkers, Mariola Gremmen, Esther van der Stappen Toward Sovereign Data Exchange Through a Meta-Platform for Data Marketplaces: A Preliminary Evaluation of the Perceived Efficacy of Control Mechanisms 285 Antragama Ewa Abbas, Wirawan Agahari, Hosea Ofe, Anneke Zuiderwijk, Mark de Reuver An Ethical Perspective on Loot Box Purchasing - Examining Psychosocial Antecedents and the Association with Indebtedness 301 Jussi Nyrhinen, Anu Sirola, Lauri Frank, Julia Nuckols, Terhi-Anna Wilska How Do QR Codes Enhance Customer Experience? Omnichannel Customer Experiences in a Brick-and-Mortar Fashion Store Matilda Holkkola, Ti na Paananen, Lauri Frank, Ti na Kemppainen, Markus 319 Makkonen Table of Contents iii. What Prevents Consumers from Making Responsible Online Purchases? Ti na Kemppainen, Markus Makkonen, Lauri Frank 335 Ethical Considerations of Augmented Reality in High-Tech Manufacturing 349 Sander van der Hoek, Marlies van Steenbergen, Pascal Ravesteijn The Effect of e-Leadership on Digital Transformation in the Dutch Public Sector 363 Pascal Ravesteijn, Ton Camu, Paul Morsch, Guido Ongena The Impact of Innovation Objectives on Industry-Academia Collaboration. A Look Towards Sustainability María de las Mercedes Gracia-Laborda, Carolina López-Nicolás, Gabriel 379 Lozano-Reina, Ángel Meroño-Cerdán, Francisco José Molina-Castillo Young Adults’ Attitude Towards Digital Payment Methods and Financial Responsibility Emma Rosenlind, Ulrik Söderström, Ole Norberg, Helen Cripps, Thomas 393 Mejtoft Digital Coaching Motivating Towards Physical Activity in Japan During the Covid-19 Pandemic Eeva Kettunen, Kristiina Jokinen, Gentiane Venture, Will Critchley, Lauri 407 Frank Mobile Clinical Decision Support Systems: A Pathway from Design to Commercialization Nalika Ulapane, Abdur Rahim Mohammad Forkan, Prem Prakash Jayaraman, 425 Penelope Schofield,1 Kate Burbury, Nilmini Wickramasinghe Fiction or Reality – Which Game Story Promotes Learning Outcome More? Linda Grogorick, Moritz Rohde , Bijan Khosrawi-Rad, Susanne Robra- 441 Bissantz Improving Completion Rate of Digital Skil Self-assessment Survey: an Empirical Study Xiao Peng, Ran Zhang, Tamara Marantika, Machil Deinum, Jet van der 457 Touw, Xander Lub, Koen van Turnhout Using Simulators to Assist with Healthcare Issues: The Impact of a Sailing Simulator on People with Adhd Gurdeep Sarai, Oren Tirosh, Prem Prakash Jayaraman, Peter Brooks, 471 Norman Saunders, Nilmini Wickramasinghe iv Table of Contents. The Anatomy of a Personal Service: the Eight Dimensions of 'Personal' Jeroen van Grondelle, Marlies van Steenbergen, Aletta Smits, Tijs 483 Timmerman, Koen van Turnhout, Harald Pol Deriving Decision Mining System Capabilities: A Research Agenda Koen Smit, Sam Leewis, Matthijs Berkhout, John van Meerten, Chaim de Gelder, Susan Bruggeling, Hanne de Deckere, Annemae van de Hoef 497 How to Design a Participation Companion: A Conversational Interface to Foster Motivation and Support Participation 515 Varinia Wittholz, Timo Strohmann, Susanne Robra-Bissantz Comparative Analysis of Market Structures of p2p Energy Trading in a Local Energy System 531 Mengqiu Deng, Xiao Peng, Yang Zhao Towards Principles For a Data-Driven Business Model Innovation Process – A Design Science Case Study 545 Michael Fruhwirth, Viktoria Pammer-Schindler Future Competences for the European Software Sector: A Mixed- Method Approach 561 Paul Morsch, Wil emijn van Haeften, Pascal Ravesteyn, Guido Ongena Towards a New Approach to Competences in Digital Transformation Xochitl Cruz, Rafael Palacios 577 The Cross-Channel Effects of In-Store Customer Experience in the Case of Omnichannel Fashion Retailing in Finland Markus Makkonen, Lauri Frank, Ti na Paananen, Matilda Holkkola, Ti na 609 Kemppainen The Role of Anticipated Guilt and its Neutralisation in Explaining Responsible Online Shopping 627 Markus Makkonen, Lauri Frank, Ti na Paananen Conceptualizing Information Systems as Biological Ecosystems - A ”New” Vocabulary for Speaking of Information Systems 647 Anna Sell, John Jeansson, Miralem Helmefalk, Hans Allmér, Leif Marcusson The Theory of Identity Management Extended to the Authentication of Identity Assertions 663 Roger Clarke Decision Analytics - A Position Paper Christer Carlsson 695 Table of Contents v. RESEARCH IN PROGRESS The Role of IT Identity in the Formation and Mitigation of Technostress Najma Saidani, Laura Ruiz Santiago 709 Conceptualizing the Impact of Digital Business Models on Privacy Concerns 721 Miranda Kajtazi, Erdelina Kurti Integration of the Sustainable Development Goals in Project-Based IT Education Joris Gresnigt, Marlies van Steenbergen 737 AutoML as Facilitator of AI Adoption in SMEs: An Analysis of AutoML Use Cases Anna K. Polzer, Johannes P. Zeiringer, Stefan Thalmann 745 Rapid Scaling of a Danish Public Health System Under COVID-19 Nicklas Frederiksen, Erik L. Møller, Jarl Tuxen, Sarah E. O’Neil , Morten Boesen 755 Strengths, Weaknesses, Opportunities and Threats to Blockchain in Industries Beyond Crypto Nilmini Wickramasinghe, Nalika Ulapane, Amir Eslami Andargoli 763 Organizations in Digital Transformation (ODT): A Literature Review on the Implications for Mal eable Organization and HRM Xiao Peng, Machiel Bouwmans 773 Improving Safety at Seas Jukka Heikkilä, Marikka Heikkilä 785 Human-AI Collaboration in the Contemporary Workplace: The Job Demands-Resources Model Perspective Tamilla Triantoro, Aleksandra Przegalinska, Anna Kovbasiuk 793 Towards an Adaptive Implementation Tool for Devices in Complex Hospital Departments Navin Sewberath Misser, Joris Jaspers, Johan Versendaal 799 Component Models for IoT Search Engine Vaidas Giedrimas 813 vi Table of Contents. An Examination of Antecedents and Consequences of Technostress among University Students: Task -Technology Fit Perspective Abdul Karim Khan, Samina Quratulain 827 DOCTORAL CONSORTIUM A Proposal for a Study of the Process Aspect of the Integrated Lifelong Treatment of Healthcare to Patients 835 Živa Rant, Tomaž Kern Understanding Psychological Contracts in the Context of Organizational Change Processes Related to Digital Government Transformation 849 Willemijn A van Haeften Presentation of the Disposition Draft Neja Samar Brenčič 863 A Quantitative Experiment: Inter-Team Retrospectives in Scaled Agile Programs Daniel Toegl, Tim Huygh, Steven de Haes 875 The Multi-Criteria Digital and Sustainable Maturity Assessment Model Katja Mohar Bastar 891 Validation of Data Maturity Criteria for Small and Medium-Sized Enterprises 905 Blaž Gašperlin DEVELOPING A TAXONOMY FOR REVENUE MODELS OF PLATFORM BUSINESS MODELS NEDO BARTELS,1 MATTHIAS KOCH,1 JAAP GORDIJN2 1 Fraunhofer IESE, Kaiserslautern, Germany nedo.bartels@iese.fraunhofer.de, matthias.koch@iese.fraunhofer.de 2 Vrije Universiteit, Amsterdam, Netherland j.gordijn@vu.nl Platform business models like Uber Ride or Airbnb Lodging enable innovative business models by operating digital platforms to connect providers and consumers of products and services in two-sided markets. A particular challenge with platform business models is designing an appropriate revenue model to capture value. This paper presents a taxonomy that classifies the different dimensions and characteristics of revenue models for platform business models. A proven taxonomy development method is used that includes a review of current literature related to Keywords: platform business models. The taxonomy provides a pl atform comprehensive classification of platform revenue models and is business models, applied to a real-life case. The results of this paper include a UML revenue class model and a final taxonomy with 14 dimensions and 64 model, characteristics. The paper contributes to the design process of digital platform, novel platform business models and expands the understanding taxonomy, of how digital platforms can generate revenues. dimensions DOI https://doi.org/10.18690/um.fov.4.2023.1 ISBN 978-961-286-751-5 2 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The significance of digital platforms continues to grow, and companies such as Uber Ride, Airbnb Lodging, Spotify Music, and eBay Marketplace have established innovative platform business models. Regardless of the industry, every company must make strategic decisions about their business model to stay competitive (Parker et al., 2016). The emergence of platform business models raises the question of what competitive advantages a company can achieve with its own business model and underscores the need for design knowledge to innovate novel (platform) business models. The motivation of this paper is based on a research preview from Bartels & Gordijn (2022) and addresses the design of systematic revenue models for platform business models. We provide a taxonomy that classifies relevant dimensions and characteristics of revenue models for platform business models. The research question for this paper is as follows: Which dimensions and characteristics can be used to describe revenue models of platform business models? To answer this research question, we used a taxonomy development process following Nickerson et al. (2013) and extracted relevant dimensions and characteristics from a literature review. We also present a use case of the Smarte.Land.Regionen (SLR) platform, a digital solution-brokering platform for German counties, where the proposed taxonomy was applied to design a possible revenue model. In follow-up research, the taxonomy wil be developed into a design tool to help practitioners create platform business models more systematical y. This paper is structured as fol ows. In section 2, we introduce key terms and relevant related work. Section 3 presents the research design of the taxonomy development process and section 4 shows the taxonomy we created. Section 5 presents the use case to which the taxonomy was applied. Final y, section 6 presents our discussion, limitations, and an outlook on future work. 2 Theoretical Background In our understanding, a platform business model is characterized by four aspects adapted from the definitions of Koch & Krohmer et al. (2022), Gordijn & Wieringa (2021), and Täuscher & Laudien (2017): (1) A platform business model describes the concept of how economic value is created, distributed, and consumed in a network of parties, called a digital ecosystem. (2) It creates value through a digital platform, N. Bartels, M. Koch, J. Gordijn: A Reconsideration of the Foundations of Identity Management 3 operated by a platform operator (i.e., asset broker), which connects at least two market sides – asset providers and asset consumers. (3) It brokers assets such as products or services via its digital platform. (4) A digital platform can serve as the hub of a digital ecosystem consisting of companies working collaboratively and competitively to meet customer needs (Moore, 1996). The revenue model is part of the value capture dimension of a business model and clarifies which monetization mechanisms are used to generate revenues. Accordingly, the revenue model of a platform business model, as we understand it, explains how revenue can be generated by enabling brokering services via a digital platform connecting asset providers and asset consumers. A taxonomy is defined as a structure above the technical terms of a subject area (Freichel et al., 2021a). In this paper, a taxonomy is considered a form of classification of relevant dimensions and characteristics for revenue models of platform business models. Van de Ven et al. (2021) presented a taxonomy for business models of data marketplaces with 17 business model dimensions and 59 business model characteristics. Springer & Petrik (2021) showed a taxonomy for platform pricing of digital platforms in the context of the Industrial Internet of Things (IIoT) with 13 impact factors and 38 characteristics. Staub et al. (2021) elaborated a taxonomy for digital platforms with 16 design dimensions and 44 characteristics. A similar taxonomy for digital platforms was elaborated by Freichel et al. (2021a) with 16 dimensions and 40 characteristics. Täuscher & Laudien (2018) presented a taxonomy for marketplace business models with 14 business model attributes and 43 specifications. They applied their taxonomy to a sample of 100 digital marketplaces and showed that there are recurring revenue models, meaning that about 74% of al platform business models studied use a commission model as the key revenue stream. This finding prompted us to investigate platform revenue models to gain a deeper understanding of crucial business model variations. Compared to existing taxonomies that conceptualize digital platforms and their business models holistically (see van de Ven et al. (2021), Freichel et al. (2021a), or Staub et al. (2021)), our work focuses on the dimensions and characteristics of revenue models for platform business models and aims to contribute to a better understanding of how digital platforms can generate revenues. 4 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3 Research Design: Taxonomy Development Process The development process of our taxonomy for revenue models of platform business models follows the guidance of Nickerson et al. (2013) as a well-structured methodology for researchers who intend to develop taxonomies step by step. The literature review, development process, and data presented in this paper are fully documented and can be found here: Bartels et al. (2023). Figure 1: Taxonomy development process adapted from Nickerson et al. (2013) As shown in Figure 1, the purpose and object of the taxonomy is defined in Step 1. In Step 2, the ending conditions are set, i.e., the criteria that the taxonomy must meet in order to be accepted. For the development of the taxonomy, Step 3, Step 4, and Step 5 are repeated in two conceptual-to-empirical iteration cycles. After applying the taxonomy to a real-life case, all ending conditions in Step 6 are met. N. Bartels, M. Koch, J. Gordijn: A Reconsideration of the Foundations of Identity Management 5 3.1 Determination of meta-characteristics and ending conditions The purpose of our taxonomy follows the research preview of Bartels & Gordijn (2022), and revenue models of platform business models such as the revenue models of eBay Marketplace, Airbnb Lodging, or Uber Ride form our object of interest. The configuration aspects of these revenue models (e.g., $0.35 insertion fee per listing on eBay) are determined as the relevant meta-characteristics of our taxonomy. Following Nickerson et al. (2013), we consider objective and subjective ending conditions that must be met for the taxonomy to be accepted: (1) The taxonomy must comprise the main dimensions and characteristics of a revenue model for platform business models, and (2) no new dimensions or characteristics should be added in the last iteration. Subjectively, the taxonomy must be (3) meaningful without being unwieldy or overwhelming and (4) extensible in order to add new dimensions or characteristics. Final y, (5) each dimension and characteristic must provide useful explanations about the object ( explanatory). 3.2 First cycle: Literature research and classification To get a data basis for the creation of the taxonomy, we conducted a literature review on revenue models of platform business models. The databases of Scopus, Web of Science, IEEE Xplore, ACM Digital Library, Google Scholar, and Dimensions were searched using the following string: (ecosystem OR platform) AND (business model OR value capture OR revenue model OR profit model). This resulted in a total of 930 papers. The screening process of titles, abstracts, and ful text was guided by the definition of inclusion and exclusion criteria. Of the total of 930 papers, 29 papers were included based on the following inclusion criterion: The paper focuses on relevant dimensions or characteristics of revenue models for platform business models (IC). In addition, five more papers were added to the included results, as we consider them relevant: Derave et al. (2022), Freichel et al. (2021a), Springer & Petrik (2021), Van de Ven et al. (2021), and Weking et al. (2020). A total of 34 papers were thus used for developing the taxonomy. The remaining 901 papers were excluded based on the following exclusion criteria: 204 papers were duplicates of another paper (EC1), 30 papers were not in English (EC2), six papers were less than three pages (EC3), 13 papers were not research papers (EC4), 41 papers were not accessible even after contacting the authors (EC5), and 607 papers did not meet the inclusion criteria (EC6). The full-text review of the 34 included papers resulted in a total of 68 dimensions and 258 6 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY characteristics for revenue models of platform business models. The review process of the literature search with each criterion is documented here: Bartels et al. (2023). To synthesize the data, a classification was created as a concept matrix according to Webster and Watson (2002). First, all dimensions were sorted alphabetical y by title, studied based on the descriptions, and coded using our own classifications. Of the 68 dimensions examined from the literature, nine dimensions could not be classified – the remaining 59 dimensions were grouped into nine self-coded dimensions. Figure 2 gives an overview of the selected revenue model dimensions derived from the literature. The concept matrix summarizes the comprehensive classifications for revenue models of platform business models on the left side (A) while showing relevant dimensions for revenue models on the right side (B). Figure 2 shows that nine dimensions could be extracted based on 27 papers. Here, “revenue model”, “revenue stream”, “revenue source”, and “pricing model” are frequently used as relevant dimensions. N. Bartels, M. Koch, J. Gordijn: A Reconsideration of the Foundations of Identity Management 7 (A) Meta-descriptions of the analyzed papers (B) Own derived and classified revenue model dimensions n ss apture- apture- ions atio y)? sine alue c alue c in classifictaion bu ens odel tream ource equency nom lated f v f v odel crim es a xo r of ions dim characteristics echanism iscovery is be l-re ent fr ens ber o ber o e d ovid N° Authors .g. ta um ode um lated um lated evenue m evenue s evenue s Pr (e N m dim N re N re R R R Paym Pricing m Price m Price d Pric 1 Curtis et al. (2020) Yes 16 5 28 x x x x 2 Derave et al. (2022) Yes 12 5 24 x x x x x 3 El Sawy & Pereira (2013) Yes 19 4 0 x x 4 Enders et al. (2008) No x 5 Freichel et al. (2021a) Yes 7 3 0 x x 6 Freichel et al. (2021b) Yes 16 2 4 x x x x 7 Ghezzi (2010) No 8 Giessmann et al. (2014) Yes 5 1 5 x 9 Helfat & Raubitschek (2018) No 10 Hoyer et al. (2009) No 11 Hyrynsalmi et al. (2012) No x 12 Immonen et al. (2014) No x 13 Janssen & Zuiderwijk (2014) No x 14 Kim (2016) No x 15 Kohler (2015) No x 16 Kübel & Zarnekow (2014) Yes 19 2 4 x x 17 Laczko et al. (2019) No x 18 Lin et al. (2020) No x 19 Mancha & Gordon (2022) No x 20 Park et al. (2020) No x 21 Rohn et al. (2021) Yes 5 2 7 x x x x 22 Ruggieri e al. (2018) No x 23 Schreieck et al. (2017) No 24 Springer & Petrik (2021) Yes 13 3 10 x x 25 Staub et al. (2021) Yes 16 3 10 x x x 26 Still et al. (2017) Yes 10 2 0 x 27 Täuscher & Laudien (2017) Yes 14 4 15 x x x x x 28 Täuscher & Laudien (2018) Yes 14 4 15 x x x x x 29 Teece & Linden (2017) No 30 Teece (2018) No 31 Ven et al. (2022) Yes 17 5 17 x x x 32 Verstegen & Doorneweert (2017) No 33 Weking et al. (2018) No x x 34 Weking et al. (2020) Yes 19 3 11 x x x x Average 13 3 10 ∑ 10 12 11 4 8 5 5 5 Deviation 5 1 8 Figure 2: Concept matrix of search results However, the initial taxonomy derived from the concept matrix did not meet the ending conditions, as the “pricing model” dimension had a strong overlap with “price mechanism”, “price discovery”, and “price discrimination”. Therefore, in the second iteration cycle, the dimension was deleted to avoid redundancy. 8 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3.3 Second cycle: Meta-model and taxonomy revision In the second iteration of taxonomy development, we created a UML class model to express the relationships of the revenue model dimensions for platform business models within the taxonomy in a transparent way. We consider this step to be useful for designing a taxonomy holistical y and ensuring its meaningfulness. The metamodel in Figure 3 il ustrates the relationships between eight classes depicting the dimensions of the taxonomy. N. Bartels, M. Koch, J. Gordijn: A Reconsideration of the Foundations of Identity Management 9 Figure 3: UML metamodel of the proposed taxonomy An asset broker and operator of a revenue model (e.g., the platform provider of the eBay marketplace) may have multiple “revenue model types”, each having a “revenue source” ( who is monetized?) and a “revenue stream” ( how to monetize?). This triangular relationship is crucial in our opinion and is also confirmed by the literature, as demonstrated in Figure 2. The pricing components, including “price 10 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY discovery”, “payment frequency”, “pricing mechanism”, and “price discrimination”, always refer to an individual “revenue stream”. The pricing model as a dimension is not explicitly included in the metamodel, as it is either redundant to the existing dimensions or can be considered as a combination. The classes shown in Figure 3 were adopted as dimensions in the second iteration. 4 Taxonomy An asset provider (e.g., Airbnb host) aims to generate revenues through a business model of its own (e.g., renting one’s own apartment to travelers), which should be viewed as a separate but relevant component for describing the overal platform business model of an asset broker (e.g., the operator of the Airbnb Lodging platform). For this, the use of a digital platform by asset providers depends on their ability to generate revenues. We concluded that a revenue model for a (two-sided) platform business model can only be described holistical y if both the asset broker's revenue model and the asset provider's revenue model are represented. Consequently, the final taxonomy includes 14 dimensions, with seven dimensions covering the asset broker's perspective and the other seven dimensions covering the asset provider's perspective. The taxonomy shown in Figure 4 satisfies al relevant ending conditions. A revenue model type of the asset broker (DB1) covers the revenue source and revenue stream through which the asset broker generates revenues. A revenue stream of the asset broker (DB2) describes how the asset broker generates revenues, i.e., the strategy the asset broker uses to monetize the revenue source through the platform. Access fees, commission fees, sale of platform services, advertising fees, listing fees, or donations may be used to generate revenue. The revenue source of the asset broker (DB3) describes who is monetized by the asset broker, i.e., the actor through whom the asset broker generates the revenue stream. Asset consumers, asset providers, or third parties can be monetized by the asset broker. The payment frequency of the platform price (DB4) describes how often payments recur for the asset broker, i.e., the frequency with which the revenue source is charged by the asset broker. Payments can appear as one-time, multiple-time, or usage-based. The price discovery of the platform price (DB5) describes who sets the platform price, i.e., whether the platform price is set by the asset broker, by asset providers, asset consumers, or by negotiations. The price mechanism of the platform price (DB6) describes the influence of supply and demand on N. Bartels, M. Koch, J. Gordijn: A Reconsideration of the Foundations of Identity Management 11 the platform price, i.e., whether the platform price is fixed or variable. A platform price can be fixed and static or variable and dependent on further factors. If the platform price is variable, it can be subject to price discrimination. The price discrimination of the platform price (DB7) describes different platform prices, i.e., whether discriminatory factors influence the platform price to be paid. Platform price discrimination can take the form of location-based, quantity-based, or feature-based price differences. Revenue model dimensions Revenue model characteristics of a platform business model of a platform business modelRevenue model type Access Commis- Pay per Sales Adver- Listing Donation Other B1 of the asset broker model sion model use model tising model model D model model Revenue stream Access Commis- Commis- Sales Adver- Listing Donations Other er B2 of the asset broker fees to sion fees on sion fees model of tising fees on D platform platform on usage platform fees for platform ok transactions services space sset br Revenue source Asset consumers Asset providers Third party Other a B3D of the asset broker the Payment frequency One-time Subscription-based frequency Usage-based frequency Other el of B4D of the platform price e mod Price discovery Platform price set by Platform price set Platform price set Platform Other B5 of the platform price asset providers by asset consumers by negotiation price set by D asset broker Revenu B6 Price mechanism Fixed platform pricing Variable platform pricing Other D of the platform price Price discrimination Feature-based price Quantity-based price Location-based price Other B7D of the platform price discrimination discrimination discrimination Sales model Rental model Pay per use model Other P1 Revenue model type D of the asset provider Sales of assets Rental fees for assets Usage fees for assets Other P2 Revenue stream D of the asset provider iderov Asset consumers Asset broker Third party Other P3 Revenue source D of the asset provider sset pr a One-time Subscription-based frequency Usage-based frequency Other P4 Payment frequency the D of the asset price el of Asset price set by asset Asset price set by Asset price set by Asset price Other P5 Price discovery D of the asset price providers asset consumers negotiation set by asset e mod broker Fixed asset pricing Variable asset pricing Other P6 Price mechanism Revenu D of the asset price Feature-based price Quantity-based price Location-based price Other P7 Price discrimination D of the asset price discrimination discrimination discrimination Figure 4: Taxonomy for revenue models of platform business models 12 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY A revenue model type of the asset provider (DP1) covers the revenue source and revenue stream by which the asset providers generate revenues. The revenue stream of the asset provider (DP2) describes how the asset providers generate revenues, i.e., the strategy the asset providers use to monetize the revenue source through the platform. The asset provider can generate revenue through the platform by sel ing, renting, or charging a usage-based fee for the asset. The revenue source of the asset provider (DP3) describes who is monetized by the asset providers, i.e., the actor through which asset providers generate their revenue stream. Asset consumers, the asset broker, or third parties can generate revenue for the asset provider. The payment frequency of the asset price (DP4) describes how often payments recur for asset providers, i.e., the frequency with which the revenue source is charged by the asset providers. Payments for an asset can appear as one-time, multiple times, or usage-based. The price discovery of the asset price (DP5) describes who sets asset prices on the platform, i.e., whether asset prices are set by the asset broker, by asset providers, or by asset consumers. The price mechanism of the asset price (DP6) describes the influence of supply and demand on asset prices, i.e., whether asset prices on the platform are fixed or variable. The price of an asset may be fixed or variable and depend on other factors. If the price of an asset is variable, it can be subject to price discrimination. The price discrimination of the asset price (DP7) describes different asset prices, i.e., whether discriminatory factors influence asset prices on the platform. Asset price discrimination can take the form of location-based, quantity-based, or feature-based price differences. 5 Taxonomy applied to the SLR Platform To ensure that our taxonomy wil be applicable, we used a revenue model of a digital platform in a research project as a real-life case. This research project, cal ed Smarte.Land.Regionen (SLR), aims to improve public services in rural areas through digital solutions. For this purpose, a digital ecosystem is being created that includes a digital platform at its core, called the SLR platform. The SLR platform follows the logic that the SLR platform operator (the asset broker) brokers digital solutions, e.g., mobility services (assets) provided by software companies (asset providers) to counties (asset consumers) and their citizens on its digital platform. The SLR platform was studied in an earlier work by Bartels & Schmitt (2022) as a use case for designing network effects for a platform business model. In this work, the SLR platform is used as a real-life object to test whether the taxonomy is suitable for N. Bartels, M. Koch, J. Gordijn: A Reconsideration of the Foundations of Identity Management 13 representing a platform revenue model. As shown in Table 1, the SLR platform's revenue model defines that software providers who want to offer their digital solutions on the SLR platform have to pay a fixed access fee to the SLR platform operator on a monthly basis. Table 1: Taxonomy applied to the SLR platform Description of the SLR platform revenue model DB1 The SLR platform operator generates revenue through an access model and monetizes the providing software companies. DB2 Revenues are generated through an access fee to the SLR platform. DB3 Software companies that provide solutions are monetized. DB4 Access fees accrue monthly. DB5 Access fees are set by the SLR platform operator. DB6 Access fees are fixed at 500€ and are not changeable. DB7 There is no price discrimination. DP1 The software companies generate revenue through the SLR platform by offering digital solutions based on a pay-per-use model and monetizing the counties. DP2 Revenues are generated through a usage fee for the digital solutions. DP3 Counties that request solutions from the SLR platform are monetized. DP4 Usage fees are incurred each time a digital solution is operated for a county. DP5 Usage fees are set by the providing software company. DP6 Usage fees are variable. DP7 Usage fees depend on the functionality of the digital solution and vary. In our view, the combination of “access fee” (in DB2) and “monthly” frequency (in DB4) is a subscription model, but we can express this more precisely through the taxonomy and consider it not as a standalone revenue model, but as a variant of the “access model” (in DB1). In this way, the digital solution listed on the SLR platform can be found by counties and booked for their citizens. Software companies generate revenue by offering counties their digital solutions through the SLR platform and customizing them to meet the needs of counties and citizens. 14 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 6 Discussion, Limitations, and Future Work The main contribution of this work is the creation of a meaningful taxonomy and metamodel in order to get a better understanding the revenue models used by platform business models. The research question of how to classify revenue models of platform business models is answered with a taxonomy of 14 dimensions and 68 characteristics. In their work, Täuscher & Laudien (2018) showed that 74% of platform business models use commission models as their core revenue model. Although this number is significant, it also indicates that much of the variation in revenue models is not ful y understood yet. In our view, there are variants such as commission per transaction (e.g., a fee per eBay product sold) or commission per unit of usage (e.g., a fee per Uber mile driven). Our taxonomy is a first step towards gaining a more nuanced understanding of revenue models of platform business models. The proposed taxonomy offers a more precise way of describing different revenue models compared to other taxonomies that use a single characteristic, such as 'subscription' (as seen in Täuscher & Laudien, 2018). As shown in our real-life case, we achieve this level of detail by combining multiple dimensions: “revenue stream” (DB2) with “access model” and “payment frequency” (DB4) with “monthly” frequency. Limitations. Our taxonomy focuses on revenue models as part of the value capture and does not address the value proposition and the value creation of a business model. Second, it focuses solely on platform business models with two-sided markets involving the asset broker and asset providers as actors with monetization intentions, and therefore cannot be used for one-sided or multi-sided platforms. Despite our transparent taxonomy development process (the research data can be found here: Bartels et al. (2023), there may still be important aspects that have gone unnoticed. An example can be seen in the payment frequency dimension, which is weakly backed in the existing literature and occurred only four times in our data (see concept matrix in Figure 2). However, recent work, such as the platform ontology of Derave et al. (2022), emphasizes the importance of frequency and shows that research on digital platforms and their business models is still evolving. Consequently, we may have overlooked other aspects in our taxonomy that need to be further elaborated in the future. N. Bartels, M. Koch, J. Gordijn: A Reconsideration of the Foundations of Identity Management 15 Future work on the proposed taxonomy should include the study of different “objects”, i.e., platform revenue models, to refine or extend the existing dimensions and characteristics, as suggested by Nickerson (2013) as an empirical-to-conceptual process. Our initial contribution of applying the taxonomy to the SLR platform is a first step. Now, the taxonomy needs to be tested on more real-life objects. The overall goal of this research is to provide this taxonomy as a design tool for practitioners to systematical y design revenue models, as proposed by Bartels & Gordijn (2022), who called this a “business model construction kit”. Acknowledgements This research was conducted as part of the Smarte.Land.Regionen project and funded by the German Federal Ministry of Food and Agriculture (BMEL) under grant number 2818SL001. Fraunhofer IESE is responsible for the implementation of the research project. We thank Patrick Mennig and Marcus Trapp for scientific discussions, Sonnhild Namingha for linguistic revision, and Anthony Woodard for his support with data preparation. References Bartels, N., Koch, M., & Gordijn, J. (2023). Reserach data_taxonomy for revenue models of platform business models (Version V1) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.7635726 Bartels, N., & Schmitt, A. (2022). Developing network effects for digital platforms in two-sided markets – The NfX construction guide. In Digital Business (Vol. 2, Issue 2, p. 100044). Elsevier BV. https://doi.org/10.1016/j.digbus.2022.100044 Bartels, N., & Gordijn, J. (2022). A Business model construction kit for platform business models - research preview. In V. Gervasi, & A. Vogelsang (Eds.), 13216. Requirements Engineering: Foundation for Software Quality. REFSQ 2022. Lecture Notes in Computer Science. Cham: Springer. https://doi.org/10.1007/978-3-030-98464-9_14. Curtis, S. K., & Mont, O. (2020). Sharing economy business models for sustainability. In Journal of Cleaner Production (Vol. 266, p. 121519). Elsevier BV. https://doi.org/10.1016/j.jclepro.2020.121519 Derave, T., Princes Sales, T., Gailly, F., & Poels, G. (2022). Sharing Platform Ontology Development: Proof-of-Concept. In Sustainability (Vol. 14, Issue 4, p. 2076). MDPI AG. https://doi.org/10.3390/su14042076 El Sawy, O. A., & Pereira, F. (2013). Business Modelling in the Dynamic Digital Space. In Springer, Briefs in Digital Spaces. Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-642-31765-1 Enders, A., Hungenberg, H., Denker, H.-P., & Mauch, S. (2008). The long tail of social networking. In European Management Journal (Vol. 26, Issue 3, pp. 199–211). Elsevier BV. https://doi.org/10.1016/j.emj.2008.02.002 Freichel, C., Fieger, J., & Winkelmann, A. (2021a). Developing a Taxonomy for Digital Platforms – A Conceptual Approach. In Proceedings of the Annual Hawai International Conference on System Sciences. Hawai International Conference on System Sciences. Hawai International Conference on System Sciences. https://doi.org/10.24251/hicss.2021.701 Freichel, C., Hofmann, A., Ernst, I., & Winkelmann, A. (2021b). A Platform Business Model for Col aborative Additive Manufacturing. In Proceedings of the Annual Hawai International 16 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Conference on System Sciences. Hawai International Conference on System Sciences. Hawai International Conference on System Sciences. https://doi.org/10.24251/hicss.2021.578 Ghezzi, A. (2012). Emerging business models and strategies for mobile platform providers: a reference framework. In info (Vol. 14, Issue 5, pp. 36–56). Emerald. https://doi.org/10.1108/14636691211256296 Giessmann, A., Kyas, P., Tyrvainen, P., & Stanoevska, K. (2014). Towards a Better Understanding of the Dynamics of Platform as a Service Business Models. In 2014 47th Hawaii International Conference on System Sciences. 2014 47th Hawai International Conference on System Sciences (HICSS). IEEE. https://doi.org/10.1109/hicss.2014.127 Gordijn, J., Wieringa, R.: e3value User Guide - Designing Your Ecosystem in a Digital World. The Value Engineers, 1st edn. (2021) Helfat, C. E., & Raubitschek, R. S. (2018). Dynamic and integrative capabilities for profiting from innovation in digital platform-based ecosystems. In Research Policy (Vol. 47, Issue 8, pp. 1391– 1399). Elsevier BV. https://doi.org/10.1016/j.respol.2018.01.019 Hoyer, V., & Stanoevska-Slabeva, K. (2009). Business models for digital business ecosystems: The case of the Open Negotiation Environment (ONE) platform. In 2009 3rd IEEE International Conference on Digital Ecosystems and Technologies (DEST). https://doi.org/10.1109/dest.2009.5276683 Hyrynsalmi, S., Suominen, A., Mäkilä, T., Järvi, A., & Knuutila, T. (2012). Revenue Models of Application Developers in Android Market Ecosystem. In Lecture Notes in Business Information Processing (pp. 209–222). Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-642-30746-1_17 Immonen, A., Palviainen, M., & Ovaska, E. (2014). Requirements of an Open Data Based Business Ecosystem. In IEEE Access (Vol. 2, pp. 88–103). Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/access.2014.2302872 Janssen, M., & Zuiderwijk, A. (2014). Infomediary Business Models for Connecting Open Data Providers and Users. In Social Science Computer Review (Vol. 32, Issue 5, pp. 694–711). SAGE Publications. https://doi.org/10.1177/0894439314525902 Kim, J. (2016). The platform business model and business ecosystem: quality management and revenue structures. In European Planning Studies (Vol. 24, Issue 12, pp. 2113–2132). Informa UK Limited. https://doi.org/10.1080/09654313.2016.1251882 Koch, M., Krohmer, D., Naab, M., Rost, D., & Trapp, M. (2022). A matter of definition: Criteria for digital ecosystem. Digital Business, 2(2). https://doi.org/10.1016/j. digbus.2022.100027 Kohler, T. (2015). Crowdsourcing-Based Business Models: How to Create and Capture Value. In California Management Review (Vol. 57, Issue 4, pp. 63–84). SAGE Publications. https://doi.org/10.1525/cmr.2015.57.4.63 Kübel, H., & Zarnekow, R. (2014). Evaluating Platform Business Models in the Telecommunications Industry via Framework-based Case Studies of Cloud and Smart Home Service Platforms. Americas Conference on Information Systems. Laczko, P., Hul ova, D., Needham, A., Rossiter, A.-M., & Battisti, M. (2019). The role of a central actor in increasing platform stickiness and stakeholder profitability: Bridging the gap between value creation and value capture in the sharing economy. In Industrial Marketing Management (Vol. 76, pp. 214–230). Elsevier BV. https://doi.org/10.1016/j.indmarman.2018.08.010 Lin, P., Zhang, X., Yan, S., & Jiang, Q. (2020). Dynamic Capabilities and Business Model Innovation of Platform Enterprise: A Case Study of DiDi Taxi. In Scientific Programming (Vol. 2020, pp. 1–12). Hindawi Limited. https://doi.org/10.1155/2020/8841368 Mancha, R., & Gordon, S. (2021). Multi-sided platform strategies for organizations: transforming the business model. In Journal of Business Strategy (Vol. 43, Issue 3, pp. 175–183). Emerald. https://doi.org/10.1108/jbs-09-2020-0203 Moore, J.F.: The Death of Competition: Leadership and Strategy in the Age of Business Ecosystems. Wiley Harper Business (1996) N. Bartels, M. Koch, J. Gordijn: A Reconsideration of the Foundations of Identity Management 17 Nickerson, R. C., Varshney, U., & Muntermann, J. (2013). A method for taxonomy development and its application in information systems. European Journal of Information Systems, 22(3), 336-359. Park, H., Kim, S., Jeong, Y., & Minshal , T. (2020). Customer entrepreneurship on digital platforms: Chal enges and solutions for platform business models. In Creativity and Innovation Management (Vol. 30, Issue 1, pp. 96–115). Wiley. https://doi.org/10.1111/caim.12404 Parker, G.G., Van Alstyne, M.W., Choudary, S.P.: Platform revolution: How networked markets are transforming the economy and how to make them work for you. WW Norton & Company (2016) Rohn, D., Bican, P. M., Brem, A., Kraus, S., & Clauss, T. (2021). Digital platform-based business models – An exploration of critical success factors. In Journal of Engineering and Technology Management (Vol. 60, p. 101625). Elsevier BV. https://doi.org/10.1016/j.jengtecman.2021.101625 Ruggieri, R., Savastano, M., Scalingi, A., Bala, D., & D’Ascenzo, F. (2018). The impact of Digital Platforms on Business Models: an empirical investigation on innovative start-ups. In Management & Marketing (Vol. 13, Issue 4, pp. 1210–1225). Walter de Gruyter GmbH. https://doi.org/10.2478/mmcks-2018-0032 Schreieck, M., Wiesche, M., & Krcmar, H. (2017). The Platform Owner's Challenge to Capture Value - Insights from a Business-to-Business IT Platform. International Conference on Interaction Sciences. Springer, V., & Petrik, D. (2021). Towards a Taxonomy of Impact Factors for Digital Platform Pricing. In Agile Processes in Software Engineering and Extreme Programming – Workshops (pp. 115– 124). Springer International Publishing. https://doi.org/10.1007/978-3-030-88583-0_11 Staub, N., Haki, K., Aier, S., & Winter, R. (2021). Taxonomy of Digital Platforms: A Business Model Perspective. In Proceedings of the Annual Hawai International Conference on System Sciences. Hawai International Conference on System Sciences. Hawai International Conference on System Sciences. https://doi.org/10.24251/hicss.2021.744 Stil , K., Seppanen, M., Korhonen, H., Valkokari, K., Suominen, A., & Kumpulainen, M. (2017). Business Model Innovation of Startups Developing Multisided Digital Platforms. In 2017 IEEE 19th Conference on Business Informatics (CBI). 2017 IEEE 19th Conference on Business Informatics (CBI). IEEE. https://doi.org/10.1109/cbi.2017.86 Täuscher, K., & Laudien, S. (2017). Uncovering the Nature of Platform-based Business Models: An Empirical Taxonomy. Proceedings of the 50th Hawai International Conference on System Sciences, Hawaii: USA. Täuscher, K., & Laudien, S. M. (2018). Understanding platform business models: A mixed methods study of marketplaces. In European Management Journal (Vol. 36, Issue 3, pp. 319–329). Elsevier BV. https://doi.org/10.1016/j.emj.2017.06.005 Teece, D. J., & Linden, G. (2017). Business models, value capture, and the digital enterprise. In Journal of Organization Design (Vol. 6, Issue 1). Springer Science and Business Media LLC. https://doi.org/10.1186/s41469-017-0018-x Teece, D. J. (2018). Profiting from innovation in the digital economy: Enabling technologies, standards, and licensing models in the wireless world. In Research Policy (Vol. 47, Issue 8, pp. 1367– 1387). Elsevier BV. https://doi.org/10.1016/j.respol.2017.01.015 van de Ven, M., Abbas, A. E., Kwee, Z., & de Reuver, M. (2021). Creating a Taxonomy of Business Models for Data Marketplaces. In 34th Bled eConference Digital Support from Crisis to Progressive Change: Conference Proceedings. Digital Support from Crisis to Progressive Change. University of Maribor Press. https://doi.org/10.18690/978-961-286-485-9.23 Verstegen, J. A., & Doorneweert, B. (2017). How to build a successful business model with big data platforms?. In CEUR Workshop Proceedings (Vol. 2025). Webster, J. and Watson, R.T. (2002). “Analyzing the past to prepare for the future: Writing aliterature review”, MIS Quarterly, vol. 26, no. 2, xiii –xxiii 18 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Weking, J., Stöcker, M., Kowalkiewicz, M., Böhm, M., & Krcmar, H. (2020). Leveraging industry 4.0 – A business model pattern framework. In International Journal of Production Economics (Vol. 225, p. 107588). Elsevier BV. https://doi.org/10.1016/j.ijpe.2019.107588 Weking, J., Hein, A., Böhm, M., & Krcmar, H. (2018). A hierarchical taxonomy of business model patterns. In Electronic Markets (Vol. 30, Issue 3, pp. 447–468). Springer Science and Business Media LLC. https://doi.org/10.1007/s12525-018-0322-5 HYPERTENSION SELF-MANAGEMENT SUCCESS IN 2 WEEKS; 3 PILOT STUDIES LUUK PA SIMONS,1 BAS WIELAARD,2 MARK NEERINCX1 1 Delft University of Technology, Faculty of Computer Science, Delft, Netherland l.p.a.simons@tudelft.nl, M.A.Neerincx@tudelft.nl 2 Health Coach Program, Hague, Netherland Bas.Wielaard@healthcoachprogram.nl Platform business models like Uber Ride or Airbnb Lodging enable innovative business models by operating digital platforms to connect providers and consumers of products and services in two-sided markets. A particular challenge with platform business models is designing an appropriate revenue model to capture value. This paper presents a taxonomy that classifies the different dimensions and characteristics of revenue models for platform business models. A proven taxonomy development method is used that includes a review of current literature related to Keywords: platform business models. The taxonomy provides a hy pertension, comprehensive classification of platform revenue models and is self-management support, applied to a real-life case. The results of this paper include a UML microlearning, class model and a final taxonomy with 14 dimensions and 64 social characteristics. The paper contributes to the design process of learning, eHealth, novel platform business models and expands the understanding employee of how digital platforms can generate revenues. health DOI https://doi.org/10.18690/um.fov.4.2023.2 ISBN 978-961-286-751-5 20 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Hypertension is the main risk factor for death worldwide, according to Lancet publications from the largest health study ever (Lozano, 2012, Lim, 2012). Though it is preventable and reversible, most people get hypertension before they retire (Ostchega, 2020, Zhou, 2021, Carey & Whelton, 2020). So, we need healthier lifestyles. But how do we effectively learn the required competences? Standard advice in health care hypertension is a bit simplistic when viewed from a competence building perspective. Plus, the feedback cycle takes too long. You might hear: “Try less salt and more sports, then come back in three months to check your blood pressure again.” This approach contrasts strongly with the lessons from SMS (Self-Management Support) literature for the need of individualized learning support, regular monitoring and follow up coaching (Dineen-Griffin, 2019). If we add to this the microlearning lessons on competence building (Emerson & Berge, 2018, Simons, 2015, 2020b), we can hypothesize why many people experience unsatisfactory results. Standard care provides virtual y no support for competence building. In a previous study we reported on a preliminary pilot (Simons, 2022a) showing feasibility and perceived usefulness of daily hypertension feedback. Still, a question remained: how robust are effects across cases (external validity)? Next, on a more detailed level of design analysis: which elements in the support intervention are valued most; how does this depend on the intervention context? Where is the room for improvement? In this paper we conduct an analysis across three different cases. We compare the results and user evaluations from three hypertension Self- Management Support (SMS) pilots of 2 weeks each, across different employee groups, organisations, and intervention settings. Our Research Question is: How do the dif erent support elements across the three cases relate to dif erences in health competences, -behaviors and blood pressure outcomes? L. PA. Simons, B. Wielaard, M. Neerincx: Hypertension Self-Management Success in 2 Weeks; 3 Pilot Studies 21 2 Theory and concepts Four areas of expertise form the basis for the interventions and design analyses of this paper. They are: lifestyle medicine for hypertension, persuasive technology for health, Self-Management Support (SMS), and microlearning in a multichannel support mix. It was researched elsewhere how the chal enges of persuasive technology (Fogg 2003, 2009) for health are not just located in the ICT (Information- & Communication Technology) design, but also in the design of the overall service scape, including health effectiveness and coaching performance (Starr 2008, Simons 2014b). It should generate positive, mutually reinforcing service experiences across communication channels, for effective health behaviours and - results. This is reflected in the following design evaluation framework for health improvement ICT solutions (Simons 2014), see Figure 1. Health effectiveness: - Health literacy - Health behaviors - Health outcomes - Quality of life and well-being Coaching performance: ICT value adding: - Promoting health actions - Quality of motivators, triggers, experiences - Supporting self-efficacy - Simplicity: familiar interfaces, ease of use - Activating intrinsic motivation - Embedded in and enhancing coach relation Figure 1: Design requirements for designing ICT-enabled healthy lifestyle support Figure 1 addresses three domains to evaluate the impact of ICT-enabled health interventions: health effectiveness, coaching performance and ICT value adding. We use it as analysis framework for section 4, Results. Lifestyle medicine for hypertension has a longstanding research tradition: overall (Roberts & Barnard, 2005) and regarding powerful short term effects on hypertension, inflammation and endothelial health of for example antioxidant foods (Franzini, 2012), flaxseed (Rodriguez-Leyva, 2013), beetroot and nitrates (Kapil, 2015), salt reduction (Dickinson, 2014) and healthy, low-fat food choices (Siervo, 2015), combined with exercise (Greger & Stone, 2016, Simons, 2022c) and stress 22 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY reduction (Pickering, 2001). We translated this research into lifestyle advice to generate short term, measurable improvements for hypertension. As discussed in more detail elsewhere (Simons, 2022a), the field of health Self- Management Support (SMS) has support process components, besides support for specific health behaviours (exercise, diet, sleep, smoking etc) and for tailoring the action plan to a participant’s own context and priorities (Demark-Wahnefried, 2007, Jonkman, 2016, Dineen-Griffin, 2019, Simons, 2013, 2017, 2020a, 2021). This set of SMS process components also forms the evaluation framework we used for user evaluation in section 4: 1. Monitoring of symptoms (regular, active self-monitoring) 2. Information transfer (throughout the learning process) 3. Competence building, including: a. Problem solving/decision making b. Plan making: self-treatment through use of an action plan c. Coping management: skills for handling challenges, frustrations etc d. Resource utilization: incl. social context or medication management Finally, microlearning concepts are highly relevant to our objectives of increasing health behaviour competence levels of participants. Especial y since our study took place in a (busy) work context, which creates a need for very efficient learning and rapid proof of effectiveness. “Business is about productivity, not learning. [. ] Inserting learning interventions into a busy employee’s schedule is a real chal enge” (Emerson & Berge, 2018). Giurgiu (2017) states that microlearning should focus on only what you need to know. And that it should fulfil the human craving for instant gratification: satisfying short term goals that support long term goals. Gabrielli et al (2017) stress the “contextual” learning in a “conversation with the world and oneself”. This conversation includes reflection, experimentation, and interpretation of results. Competence building is about embedded learning, where doing and achieving results are at least as important as learning (Emerson & Berge, 2018, Simons, 2010). Multiple studies show how self-management tools and ICT (Information and Communication Technology) are useful, in a multichannel service-scape, for: goal setting based on personal preferences, ICT supported tracking and progress feedback (Kari, 2017, Lehto, 2013, Lopez, 2011, Ricciardi, 2013, Wickramasinghe, 2010). Elements like individual coaching, eTools like microlearning for health, L. PA. Simons, B. Wielaard, M. Neerincx: Hypertension Self-Management Success in 2 Weeks; 3 Pilot Studies 23 Quantified Self (QS, Swan, 2012, 2013) progress tracking, WhatsApp groups and peer coaching in virtual support teams have al been shown to aid motivation and success (Simons, 2015, 2016, 2020b, 2022b). So does the power of group-based social learning (Bandura, 1971) 3 Methods and Materials In a design research approach, we follow the design cycle methodology of Vaishnavi & Kuechler, (2004) which goes from problem awareness and solution suggestion to development, evaluation and conclusion. After reporting our multiple-case study results in section 4, we discuss design lessons in section 5. The hybrid lifestyle intervention with twice-daily biofeedback consisted of: • Telephone intake & instructions for BP home measurements • Start- and final group sessions (2 weeks apart, face-to-face) • Daily MS Teams eCoaching in week 1 • (Case A: individual and group; Case B: group; Case C: only email tips) • Twice-daily BP measurements and logging email (Case A & B) • Feedback on group progress after 1st week (Case A & B) • Healthy recipe suggestions • Content (portal and/or email) on health, BP, and behaviour strategies From Nov 2021 to Feb 2023, a multiple-case study with three employee groups was conducted to evaluate real world impacts of the healthy lifestyle intervention for hypertension Self-Management in Dutch work settings. They were smal scale pilots (n=8 to n=4), for three reasons. First, because we saw previously (Simons, 2022a) how robust the BP effects were across users, which enables smal group sizes. Second, because we fol ow the design approach of multiple smal tests to col ect and test multiple improvement options, instead of conducting one big test (Cennamo, 2019). Third, we depended on employer organisations for volunteers. In total, n=20 volunteers participated. 24 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 1: Case description and start situation Aspects Case A (n=8)1 Case B (n=6) 2 Case C (n=4) Case start Nov ‘21 Nov ‘22 Jan ‘23 Participants 4 men, 4 women 3 men, 3 women 4 men Avg. start blood 145/92 161/112 155/95 pressure (mmHg) Intervention 11 days 11 days 17 days duration Final user 10 weeks after start 5 weeks after start 5 weeks after start evaluation Support format Extra App for In week 1: longer Light-weight: specifics healthy menus daily e-Sessions, * no coaching with more content * no daily BP log- & group mail interaction * info via mail instead of portal Cases A and B were conducted with mixed groups of university groups (mostly support staff, with some academics) and Case C with ICT professionals. Their SES (Socio-Economic Status) and education levels were on the Dutch average or above. They al had hypertensive BP at start and volunteered for these 2-week in-company BP interventions. There were some cross-case differences: in the intervention service mix, group make-up, and organisational context, which enabled some interesting cross-case observations, see next section. 4 Results & Cross-Case Comparisons A first question for our findings is: were there meaningful BP improvements across these cases? The short answer is: yes. 1 One of the participants had a user evaluation outlier pattern, see Table 3. 2 One of the participants had a user evaluation outlier pattern, see Table 3. L. PA. Simons, B. Wielaard, M. Neerincx: Hypertension Self-Management Success in 2 Weeks; 3 Pilot Studies 25 Figure 2: Average blood pressure drop in Case B (n=6) Previously, we reported on the large BP improvements of case A (Simons, 2022a): from 145/92 down to 126/86 mmHg in 11 days (from Monday morning at start, to Friday morning in the second week). Case C also showed BP improvements, see Table 2, from 155/95 to 139/85 mmHg. The largest improvements occurred in Case B, see the BP trend line in Figure 2. Since participants measure their BP every morning and evening, an 11-day (average) ‘spiky’ trend line was created per pilot. The line is ‘spiky’ since evening BP tends to be higher than morning BP. Based on our user evaluations, we can say that participants generally find it very motivation to see their individual and collective trends: “I was positively and strongly surprised how large the impacts of our behaviour changes were.” In case B, hypertension dropped from 161/112 down to 129/90 mmHg in 11 days. A second finding is that the intended outcome of this intervention on BP occurs quite robustly across individuals. (Which, on a methodological level, also enables us to work with small pilot groups and still observe robust effects per group.) The extend of BP effect robustness across individuals is indicated in Table 2, with the ‘High BP Responder’ percentage per case. We defined a ‘High BP Responder’ as a participant who had an average or above average BP improvement. This leads to a third finding: Case A and B have more 'High Responders' than Case C. Although this could be a coincidence, we propose that this is caused by the lesser degree of competence support given in Case C. This proposition is supported by our 26 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY qualitative user- and case evaluations below, plus it is explored further in section 5, discussion. Table 2: Cross-Case findings and authors’ design evaluation, on design requirements from Figure 1 (authors’ opinions, 5-point scale from - - to ++) Findings Case A (n=8) Case B (n=6) Case C (n=4) Final Blood 126/86 129/90 139/85 Pressure (BP, Avg. mmHg) BP drop, mmHg -19/-6 -32/-22 -16/-10 % of High BP 63% 66% 25% Responders3 Health + Healthier diet ++ Largest diet - No changes in behaviours + Avg. 10.000 improvements intensive exercise steps/day ++ Highest physical - Most are stil activity searching how to implement in daily patterns Coaching + Raise efficacy ++ Largest + Relevance of performance + Adoption information transfer content was valued (except by some: & impact from - Progress depends time constraints) assignments on user him/herself, ++ Largest social without daily learning coaching ICT value add + BP log mails + BP log mails daily; - Portal and daily daily; impact impact logging not used + Daily coaching + Daily digital ‘day- + Daily mail tips (indiv & group) start’ sessions were appreciated + Info in portal + Info in portal Table 2 shows a design evaluation across cases, based on the theory framework: health behaviours, coaching performance and ICT value add. First, regarding health, al cases showed improvements in BP and health behaviours. Case B showed the largest improvements and Case C the smal est. In our observation, this was a consequence of the second aspect: coaching. The coach assignments for behaviour 3 This the % of participants in a case that had average or above average BP improvments. This is an indicator of robustness of BP results across participants. L. PA. Simons, B. Wielaard, M. Neerincx: Hypertension Self-Management Success in 2 Weeks; 3 Pilot Studies 27 improvement were more explicit in case B than A (e.g. “no cheese or meat for two weeks and at least 8oo grams per day of fruits or vegetables”) and there was extensive daily group coaching and content on everyday chal enges. Case C had no coaching, just a start workshop to explain people what to do, plus daily content mails until the final workshop. Thus, the extent of behaviour progress largely depended on a person’s self-management. The third design evaluation aspect, ICT value add, was higher in Case A and B, than in C: twice-daily BP logging mails, portal information on health and BP, healthy recipes, daily e-coaching in week 1 (in Case A more on the individual, in Case B more on group level learning, content, exchange of experiments and follow-up assignments) and feedback on group level BP progress after week 1. In Case C this was replaced with daily emails tips on hypertension, health, and behaviour change tactics. From the user evaluations, we discuss the perceived usefulness of the various intervention components across cases, see Table 3. Scores were given on a 7-point Likert scale, ranging from ‘totally disagree’ to ‘totally agree’, in answer to the question: ‘Which components stimulated you to adopt healthier behaviours?’ The components are clustered in the SMS process framework, even though some components support more than one SMS process. 28 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 3: Components that stimulated healthier behaviours (7-point (dis)agree, Avg) Monitoring: Case A4 Case B5 Case C6 1. Mail triggers for blood pressure logging 4.9 6.2 n.a. 2. Daily management 5.4 5.6 6.0 3. Gaining more blood pressure control 6.3 6.4 6.5 Information transfer: 4. Start workshop 6.4 6.4 6.8 5. Healthy menu suggestions (App/portal) 4.4 4.8 n.a. 6. Health and blood pressure information in 5.4 5.2 6.3 portal/mails 7. More understanding of blood pressure & 6.1 6.2 6.0 health Competence building: 8. Follow-up workshops 6.3 6.2 7.0 9. Individual tips and answers to my questions 6.6 6.2 6.8 from the coaches 10. Doing this as a group 6.4 5.8 5.5 11. Tips in dealing with chal enges 6.0 6.2 6.0 In Table 3 we see that the main perceived benefit from Monitoring was the blood pressure control participants gained (3.). For the second SMS process element, Information transfer, we see that the start workshop (4.) and increased understanding of blood pressure and health (7.) were valued most. These two intervention components (4. & 7.) were not just about information transfer, but also about increasing competences for: effective plan making and prioritizing efforts on those lifestyle choices that have the best combination of short-term effectiveness and long-term perceived attractiveness/ feasibility for a participant. The element, Competence building, is key for training sustainable self-management skills and coping strategies. All four components (8. to 11.) receive relatively high scores, and for case B the perceived value of doing this as a group (10.), was explicitly stressed by participants 4 One of the participants had an outlier pattern of scoring (since she could not be present at several of the group coach sessions, due to il ness plus family logistics). Table 3 displays the average scores of the other 7. (Score 4=neutral) Her scores were resp.: 6; 6; 6/3; 3; 4; 5/3; 6; 3; 3. 5 One of the participants had an outlier pattern of scoring (she had hereditary hypertension since 18 years old and her BP values did not change, despite her best efforts). Table 3 displays the average scores of the other 5. (Score 4=neutral) Her scores were resp.: 4; 4; 4/4; 4; 4; 5/4; 4; 6; 5. 6 n.a. = not applicable L. PA. Simons, B. Wielaard, M. Neerincx: Hypertension Self-Management Success in 2 Weeks; 3 Pilot Studies 29 in the joint group evaluation after 5 weeks. So, support for competence building was general y valued by the participants. Previously, a more detailed presentation of the healthy behaviour chal enges was given, as wel as what helped participants to improve those behaviours (Simons, 2022a). In answer to our Research Question (How do the support elements relate to differences in health competences, -behaviours and blood pressure outcomes), several elements for promoting health improvements found previously, were confirmed in our cross-case analysis: a. Rapid feedback: twice-daily measurement of progress b. Achieving results and enhancing self-efficacy c. ‘Quick results’-tips & education: which large steps for large benefits d. Practical tips for every-day choices and practicing new behaviour patterns e. High quality coaches and coach sessions to increase health competence f. Doing this in a group and teaching each other g. Coaching on coping strategies Next, we highlight findings from the main cross-case differences. After Case A, we made three adaptions to the support components, based on user feedback. First, we had rented a (commercial) mobile App for the users, with many easy and adaptable ‘hypertension-friendly’ menu options, including a ‘home-delivery’ function. However, this was hardly used and thus discarded. Besides, we relied quite a bit on individual coaching. But given the large benefits we saw from group learning in case A, we changed two other things for Case B: (1) we replaced individual coaching with longer, information-intensive daily digital group workshops as day-starters in week 1, and (2) we gave more explicit daily assignments: what experiments and behaviours to practice today. We subsequently observed larger improvements in healthy eating and exercise in Case B, see Table 2, which we propose is a consequence of these changes that were made. In Case C support was lightweight, see also Table 2, given the different organisation context and participants’ preferences for a more light-weight (and less time-intensive) approach. So, there was no coaching in between the start and final workshop, of two hours each. Besides, daily logging and portal access were not used (even though the participants did monitor their blood pressure themselves during the two weeks). Instead, they received daily email tips on health behaviours, self-management, and blood pressure. 30 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY The main finding from case C is that health behaviour changes were smaler and that only one participant achieved above-average BP improvements. Cross-case differences confirm that most individuals find it hard to make large changes in their health choices solely based on information. Cases A and B (versus C) il ustrate that for most participants daily support with coaching and group learning processes appear to be required to generate sufficient behaviour-and Blood Pressure (BP) improvements , and even daily support is not enough for every participant in Cases A and B. Next, one element stands out from the user feedback which was very similar across al three cases: the quality of information/tips was highly valued by participants, because it was effective for generating rapid BP improvements. In al three cases, the information was deemed important (“these real y helped”) and valued (“most useful information in years”). 5 Discussion This cross-case design analysis has several limitations. Firstly, due to its small scale (in total n=20 participants across 3 cases) our research question is qualitatively answered, not quantitatively. Secondly, the BP trends are based on self-reported measures. Stil , they practiced these measurements during a wash-in period of three days before the start, which helps create robustness. Thirdly, we qualitatively assessed health competence growth, changes in lifestyle behaviours, and learning strategies. In future research, we would like to use more formal and validated surveys to generate health competence improvement scores for example. However, this is not straightforward, for two reasons: (1) BP related competence is not the same as general health competence and (2) health competence was already above average in these participants at the start. Fourthly, each intervention/case tested multiple intervention components together, without control group. So, cross-case intervention differences provide some insights, but interpretations are qualitative and (inter-)subjective. Based on the cross-case analysis, we propose four design lessons. • First, since many individuals indicate that they find it hard to make large health behaviour changes based solely on information, as stated in the L. PA. Simons, B. Wielaard, M. Neerincx: Hypertension Self-Management Success in 2 Weeks; 3 Pilot Studies 31 findings, extra support is needed: daily group coaching speeds up social learning processes and stimulates participants to do new health experiments or try new tactics to cope with difficult situations. • Second, light-weight support (like in Case C) only works if self- management competences of participants are high. Besides, participants struggle more with establishing longer term health patterns. • Third, the quality of information and tips must be high: it is valued (and applied) if participants recognize it as effective for generating rapid BP improvements. This in turn requires daily ‘proof’ from BP improvements. • Fourthly, the power of the ‘Chal enge Regime’ was mentioned in many evaluations. The two main elements: (1) making a commitment for large health changes and experiments, (2) knowing that it is for only two weeks. This combination enhances temporary attention, motivation, effort, and willingness to experiment. The result is: more improvement, more learning, a positive experience, plus a desire to continue using some of the lessons learned in the longer run. As a lesson on practical implications, we should not forget the important added value of technology in this intervention: daily home monitoring is now feasible thanks to affordable and reliable blood pressure consumer electronics; our mail/web-based coaching portal enables real-time progress tracking by participants and coaches alike; daily MS Teams meetings enabled high quality group and individual coaching without travel- or time constraints; our portal content database supports participants with multiple lessons on blood pressure and healthy lifestyle; the healthy menu App offered even included a button to directly order/deliver the ingredients to participants’ homes (even though this latter option was not used by the participants). In short, the intervention was highly dependent on these technologies and tools. By contrast, microlearning is sometimes framed as ‘a tool’ or technology. But we saw that it is much more. Its value as an embedded learning strategy to create daily, relevant, and ‘rich’ learning instances was key in our case implementations: creating multiple, daily competence-building microlearning opportunities, also face-to-face and in group discussions. In conclusion, this intervention was attractive and feasible for the participants. It was effective for achieving blood pressure improvements in two weeks. The value was confirmed of the Self-Management Support (SMS) and microlearning 32 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY approaches for competence building. Besides, our study illustrates the added value of: (a) group coaching/social learning; (b) a ‘Challenge Regime’ with high commitments for a short time; (c) self-efficacy growth for users from large health results within days; (d) multiple (technology-enabled) health competence building lessons each day. These options hold promise for future health Self-Management Support innovations. Acknowledgements This research was (partly) funded by the https://www.hybrid-intel igence-centre.nl/ a 10-year programme funded the Dutch Ministry of Education, Culture and Science through the Netherlands Organisation for Scientific Research, grant number 024.004.022 and by EU H2020 ICT48 project ``Humane AI Net'' under contract $\# $ 952026. References Bandura, A. (1971). Social learning theory (Vol. 1). Prentice Hall, Englewood cliffs. Carey, R. M., & Whelton, P. K. (2020). Evidence for the universal blood pressure goal of< 130/80 mm Hg is strong: controversies in hypertension-pro side of the argument. Hypertension, 76(5), 1384-1390. Cennamo, K., & Kalk, D. (2019). Real world instructional design: An iterative approach to designing learning experiences. Routledge, Oxfordshire, UK. Demark-Wahnefried, W., Clipp, E., Lipkus, I., Lobach, D. et al. (2007). Main Outcomes of the FRESH START Trial: A Sequentially Tailored, Diet and Exercise Mailed Print Intervention Among Breast and Prostate Cancer Survivors. J Clin Oncol, 25(19), pp. 2709-2718. Dickinson, K. M., Clifton, P. M., Burrel , L. M., Barrett, P. H. R., & Keogh, J. B. (2014). Postprandial effects of a high salt meal on serum sodium, arterial stiffness, markers of nitric oxide production and markers of endothelial function. Atherosclerosis, 232(1), 211-216. Dineen-Griffin, S., Garcia-Cardenas, V., Wil iams, K., & Benrimoj, S. I. (2019). Helping patients help themselves: a systematic review of self-management support strategies in primary health care practice. PloS one, 14(8), e0220116. Emerson, L. C., & Berge, Z. L. (2018). Microlearning: Knowledge management applications and competency-based training in the workplace. Knowledge Management & E-Learning, 10 (2), pp.125-132. Fogg, B. J., & Fogg, G. E. (2003). Persuasive Technology: Using Computers to Change What We Think and Do. Morgan Kaufmann, San Francisco. Fogg, B. J. (2009). A behavior model for persuasive design. Proceedings of the 4th international conference on persuasive technology. ACM, 2009. Franzini, L., Ardigo, D., Valtuena, S., Pel egrini, N., Del Rio, D., Bianchi, M. A., . . & Zavaroni, I. (2012). Food selection based on high total antioxidant capacity improves endothelial function in a low cardiovascular risk population. Nutrition, Metabolism and Cardiovascular Diseases, 22(1), 50-57. Gabrielli, S., Kimani, S., & Catarci, T. (2017). The design of microlearning experiences: A research agenda (on microlearning). Giurgiu, L. (2017). Microlearning an evolving elearning trend. Scientific Bulletin, 22(1), 18-23. Greger, M., & Stone, G. (2016). How not to die: discover the foods scientifical y proven to prevent and reverse disease. Pan Macmil an, New York City. L. PA. Simons, B. Wielaard, M. Neerincx: Hypertension Self-Management Success in 2 Weeks; 3 Pilot Studies 33 Jonkman, N. H., Schuurmans, M. J., Jaarsma, T., Shortridge-Baggett, L. M., Hoes, A. W., & Trappenburg, J. C. (2016). Self-management interventions: proposal and validation of a new operational definition. Journal of clinical epidemiology, 80 (December 2016), 34-42. DOI: https://doi.org/10.1016/j.jclinepi.2016.08.001 Kapil, V., Khambata, R. S., Robertson, A., Caulfield, M. J., & Ahluwalia, A. (2015). Dietary nitrate provides sustained blood pressure lowering in hypertensive patients: a randomized, phase 2, double-blind, placebo-control ed study. Hypertension, 65(2), 320-327. Kari, T., Koivunen, S., Frank, L., Makkonen, M., & Moilanen, P. (2017). The expected and perceived wel -being effects of short-term self-tracking technology use. International Journal of Networking and Virtual Organisations, 17(4), 354-370. Lozano, R., Naghavi, M., Foreman, K., Lim, S., Shibuya, K., Aboyans, V., . . & Remuzzi, G. (2012). Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. The lancet, 380(9859), 2095-2128. Lim, S. S., Vos, T., Flaxman, A. D., Danaei, G., Shibuya, K., Adair-Rohani, H., . . & Pelizzari, P. M. (2012). A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. The lancet, 380(9859), 2224-2260. Lehto, T., Oinas-Kukkonen, H., Pätiälä, T., & Saarelma, O. (2013). Virtual health coaching for consumers: a persuasive systems design perspective. International Journal of Networking and Virtual Organisations 4, 13(1), 24-41. Lopez, G., Shuzo, M., & Yamada, I. (2011). New healthcare society supported by wearable sensors and information mapping-based services. International Journal of Networking and Virtual Organisations, 9(3), 233-247. Ostchega, Y., Fryar, C. D., Nwankwo, T., & Nguyen, D. T. (2020). Hypertension prevalence among adults aged 18 and over: United States, 2017–2018, NCHS, National Health and Nutrition Examination Survey, 2017–2018: https://stacks.cdc.gov/view/cdc/87559. Pickering, T. G. (2001). Mental stress as a causal factor in the development of hypertension and cardiovascular disease. Current hypertension reports, 3(3), 249-254. Ricciardi, F., Rossignoli, C., & De Marco, M. (2013). Participatory networks for place safety and livability: organisational success factors. International Journal of Networking and Virtual Organisations 4, 13 (1), 42-65. Roberts, C. K., & Barnard, R. J. (2005). Effects of exercise and diet on chronic disease. Journal of applied physiology, 98(1), 3-30. Rodriguez-Leyva, D., Weighel , W., Edel, A. L., LaVal ee, R., Dibrov, E., Pinneker, R., . . & Pierce, G. N. (2013). Potent antihypertensive action of dietary flaxseed in hypertensive patients. Hypertension, 62(6), 1081-1089. Siervo, M., Lara, J., Chowdhury, S., Ashor, A., Oggioni, C., & Mathers, J. C. (2015). Effects of the Dietary Approach to Stop Hypertension (DASH) diet on cardiovascular risk factors: a systematic review and meta-analysis. British Journal of Nutrition, 113(1), 1-15. Simons, L. P. A., & Hampe, J. F. (2010). Service Experience Design for Healthy Living Support; Comparing an In-House with an eHealth Solution. The 23rd Bled eConference, pp. 423-440. Accessed 2010 from www.bledconference.org Simons LPA, Hampe JF, Guldemond NA. (2013). Designing Healthy Living Support: Mobile applications added to hybrid (e)Coach Solution. Health and Technology. 3 (1) pp.85-95. DOI 10.1007/s12553-013-0052-9 Simons LPA, Hampe JF, Guldemond NA. (2014). ICT supported healthy lifestyle interventions: Design Lessons. Electronic Markets. 24 pp. 179-192. DOI 10.1007/s12525-014-0157-7. Simons LPA, Foerster F., Bruck PA, Motiwalla L & Jonker CM. (2014b). Microlearning mApp to Improve Long Term Health Behaviours: Design and Test of Multi-Channel Service Mix. Paper presented at the 27th Bled eConference. Bled, Slovenia, Proceedings. Retrieval from www.bledconference.org and http://aisel.aisnet.org/bled2014/4 34 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Simons LPA, Foerster F., Bruck PA, Motiwala L & Jonker CM. (2015). Microlearning mApp Raises Health Competence: Hybrid Service Design. Health and Technology, 5 pp 35-43. Simons, L. P., Pijl, H., Verhoef, J., Lamb, H. J. et al (2016). Intensive Lifestyle (e) Support to Reverse Diabetes-2. In Bled eConference (p. 24), accessed Dec 20, 2016 www.bledconference.org. Simons LPA, Hafkamp MPJ, Bodegom D, Dumaij A, Jonker CM. (2017). Improving Employee Health; Lessons from an RCT. Int. J. Networking and Virtual Organisations, Vol. 17, No. 4, pp.341– 353. DOI https://doi.org/10.1504/IJNVO.2017.088485 Simons, LPA, Heuvel, AC van den, Jonker CM. (2018). eHealth WhatsApp Group for Social Support; Preliminary Results, pp. 225-237, presented at the 31st Bled eConference. Bled, Slovenia, Proceedings retrieval from www.bledconference.org. ISBN-13: 978-961-286-170-4, DOI: https://doi.org/10.18690/978-961-286-170-4 Simons, LPA., van den Heuvel, W. A., & Jonker, C. M. (2019). WhatsApp Peer Coaching Lessons for eHealth. In Handbook of Research on Optimizing Healthcare Management Techniques (pp. 16-32). IGI Global, ISBN 9781799813712. Simons, LPA, (2020a). Health 2050: Bioinformatics for Rapid Self-Repair; A Design Analysis for Future Quantified Self, pp. 247-261, 33rd Bled eConference. June 28-29, Bled, Slovenia, Proceedings retrieval from www.bledconference.org. ISBN-13: 978-961-286-362-3, DOI: https://doi.org/10.18690/978-961-286-362-3.17. Simons, LPA, Heuvel, AC van den, Jonker CM. (2020b). eHealth WhatsApp for social support: design lessons. International Journal of Networking and Virtual Organisations, 23(2), 112-127. DOI https://doi.org/10.1504/IJNVO.2020.108857. Simons, LPA, Neerincx MA, Jonker CM (2021). Health Literature Hybrid AI for Health Improvement; A Design Analysis for Diabetes & Hypertension, pp. 184-197, 34th Bled eConference. June 27-30, Bled, Slovenia, Proceedings retrieval from www.bledconference.org. ISBN-13: 978-961-286-385-9, DOI: https://doi.org/10.18690/978-961-286-385-9 . Simons, LPA, Gerritsen, B, Wielaard, B, Neerincx MA (2022a). Health Self-Management Support with Microlearning to Improve Hypertension, pp. 511-524, 35th Bled eConference. June 26-29, Bled, Slovenia, Proceedings retrieval from www.bledconference.org. ISBN-13: 978-961-286-616-7, DOI: 10.18690/um/fov.4.2022 Simons, LPA, Pijl M, Verhoef J, Lamb HJ, van Ommen B, Gerritsen B, Bizino MB, Snel M, Feenstra R, Jonker CM. (2022b). e-Health Diabetes; 50 Weeks Evaluation. Int. J. Biomedical Engineering and Technology, 38(1), 81-98. Simons, LPA, Neerincx MA, Jonker CM (2022c). Is Google Making us Smart? Health Self-Management for High Performance Employees & Organisations, International Journal of Networking and Virtual Organisations, Vol 27, No 3, pp.200-216. DOI: 10.1504/IJNVO.2022.10053605 Starr, J. (2008). The coaching manual: the definitive guide to the process, principles and skil s of personal coaching. Prentice Hal , New York. Swan, M. (2012). Health 2050: The realization of personalized medicine through crowdsourcing, the quantified self, and the participatory biocitizen. Journal of personalized medicine, 2(3), 93-118. Swan, M. (2013). The quantified self: Fundamental disruption in big data science and biological discovery. Big data, 1(2), 85-99. Vaishnavi, V and Kuechler, W. 2004. Design Research in Information Systems. Last updated August 16, 2009 from http://desrist.org/design-research-in-information-systems Wickramasinghe, N., & Goldberg, S. (2010). Transforming online communities into support environments for chronic disease management through cell phones and social networks. International journal of networking and virtual organisations, 7(6), 581-591. Zhou, B., Perel, P., Mensah, G. A., & Ezzati, M. (2021). Global epidemiology, health burden and effective interventions for elevated blood pressure and hypertension. Nature Reviews Cardiology, 18(11), 785-802. DATA GOVERNANCE CAPABILITIES; EMPIRICAL VALIDATION IN CASE STUDIES OF LARGE ORGANISATIONS JAN R. MERKUS, REMKO W. HELMS, ROB J. KUSTERS Open Universiteit, Faculty of Science, Dept of Information Science, Heerlen, Netherland jan.merkus@ou.nl, remko.helms@ou.nl, rob.kusters@ou.nl The exponential growth of data within organisations necessitates the implementation of effective data management practices, which in turn necessitates the establishment of data governance. The evaluation of the maturity of data governance can be carried out using maturity models. However, the existing data governance maturity models are limited in their consistency in terms of data governance capabilities used and lack empirical validation. To address this gap, this study aims to validate the set of data governance capabilities identified in prior research within large organisations. This study employs a case study research design, using semi-structured interviews with experts in data governance. As a basis for the semi-structured interviews, maturity models are designed as questionnaires to discuss the relevance of each data governance capability. The results of this study provide empirical validation of the set of data governance Keywords: data capabilities and contribute to the advancement of both data governance governance research and practice by providing a comprehensive, capabilities, validated set of data governance capabilities for maturity model empirical validation, design to advance data governance within and between maturity organisations. model DOI https://doi.org/10.18690/um.fov.4.2023.3 ISBN 978-961-286-751-5 36 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The growing amount of data in organisations, due to the increasing digitization of processes, necessitates the implementation of proper data management (Weber et al., 2008). Data governance (DG) must be established to ensure proper data management. DG is defined as “establishing management of data in an organisation assuring quality and access during its life-cycle to be accountable for data assets” (J. Merkus et al., 2019). Maturity models, which are used to measure and improve organisational performance in various application areas (Rosemann et al., 2004; a. van Looy et al., 2011), aid in implementing DG. Maturity models are based on organisational capabilities, which are the collective abilities of an organisation to carry out business processes that contribute to its performance (Brennan et al., 2018; J. R. Merkus et al., 2020). These capabilities are used to measure organisational maturity by evaluating organisational activities against staged levels of maturity (Becker et al., 2009; J. R. Merkus et al., 2020, 2021; Paulk et al., 1993). Maturity models can also focus on specific capabilities like for instance data governance. Hence, Data Governance Capabilities (DGCs) can be used to measure DG maturity (Merkus et al., 2021). Research on Data Governance Maturity Models (DGMMs) has identified several sets of DGCs for maturity model design (J. R. Merkus et al., 2021; Olaitan et al., 2019; Permana & Suroso, 2018). However, the DGCs used in existing DGMMs are inconsistent, resulting in different sets of DGCs for each DGMM in the literature (Heredia-Vizcaíno & Nieto, 2019; Permana & Suroso, 2018; Rivera et al., 2017). Furthermore, only a few DGMMs have been validated in practice, and, when validated, they only did so in a single or smal organisations (J. R. Merkus et al., 2020; Olaitan et al., 2019; Permana & Suroso, 2018; Rivera et al., 2017). Thus, there is a need for a more comprehensive DGC model validated in practice (J. R. Merkus et al., 2021). To address this gap, we aim to empirical y validate the set of 34 DGCs that we developed in our earlier, theoretical study. That study used a systematic literature review to identify DGCs from a broad range of literature (J. R. Merkus et al., 2020). However, the resulting set of DGCs was based solely on literature and requires proper validation in practice. Therefore our research question is: To what extent J. R. Merkus, R. W. Helms, R. J. Kusters: Data Governance Capabilities; Empirical Validation in Case Studies of Large Organisations 37 are the Data Governance Capabilities for Maturity Model design valid in practice? This paper provides new knowledge through the empirical validation of the previously proposed set of DGCs, resulting in a set of DGCs that can serve as a reference for other studies in the area of Data Governance. The practical significance of this research lies in the contribution of a more comprehensive and validated set of DGCs that can serve as to assess the status quo of DG in organisations more accurately e.g. as the basis for a DG maturity model. Validated DGCs also enable more precise benchmarking with other organisations. The remainder of this paper is structured as fol ows. The literature review on DGMM and DGC is presented in section 2. Then in section 3, we present our research methodology to validate our DGCs, followed by the validation results in section 4. Finally, the conclusions are presented in section 5. 2 Background This section provides a theoretical foundation for our research. First, we define the concept of DG capabilities and subcapabilities as the cornerstone of our study. Second, we identify a gap in the literature regarding the validation of existing DGCs Third, we present the set of DGCs we developed in an earlier study and wil be validated in this research. DG Capabilities (DGCs) encompass an organisation's collective abilities to govern data assets effectively (Brennan et al., 2018; J. Merkus et al., 2019). A DGC indicates what an organisation is capable of doing concerning specific DG activities. For example, the DGC establish data stewardship describes an organisations ability to set up data stewardship functions. Subsequently, each DGC can be broken down into subcapabilities, a term also used by other authors (Bandara et al., n.d.). The term subcapability refers to a set of subdimensional capabilities that specifies in more detail what makes up the capability dimension. DGCs are used in Data Governance Maturity Models (DGMM) to measure the maturity of data governance in an organisaton (J. R. Merkus et al., 2021; Rosemann et al., 2004; a. van Looy et al., 2011; A. van Looy et al., 2011). To date, only a few 38 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY peer-reviewed DGMMs have been developed, with widely differing sets of DGCs selected as the basis for each model (Dasgupta et al., 2019; Heredia-Vizcaíno & Nieto, 2019; Olaitan et al., 2019; Permana & Suroso, 2018; Rivera et al., 2017). Nevertheless, there are also common DGCs in these models and the most common DGCs are: establishing policies, principles, and procedures and managing metadata. The other DGCs used vary widely and some DGCs have been used in one DGMM only. Based on this, one can conclude that there is a lack of agreement amongst researchers on a single set of DGCs. This is further illustrated by comparing existing sets of DGCs from different DGMMs, as presented in Table 1. Despite recent progress in DG research, which has focused on DG mechanisms and DG activities between organisations, the validation of these mechanisms, activities and the capabilities to execute them is still lacking (Abraham et al., 2019; Jagals et al., 2021; Lis & Otto, 2021). A limited number of researchers have identified certain principles, activities, and critical success factors for Data Governance (DG) as Data Governance Capabilities (DGCs) (Alhassan et al., 2016, 2018, 2019). These DGCs have been validated in a few individual case studies. Other researchers have discovered DGCs as mechanisms from DG-related research area of information technology governance, aimed at planning and controlling data management activities (Abraham et al., 2019). However, these mechanisms are yet to be validated in practical settings. Recently, DG research has outlined DG activities that take place between organizations (Lis & Otto, 2021). However, the execution of these activities and the capabilities required to carry them out need to be empirical y validated in practice. So, although some DGCs have been identified by recent DG research, the empirical validation of these DGCs is still limited. In previous research, we have identified an extensive set of DGCs based on a systematic literature review (J. R. Merkus et al., 2021). Table 1 provides an overview of this set of DGCs and compares it with the DGCs that have been identified/used in other research. The vertical axis in Table 1 lists the 34 DGCs we identified in earlier research. On the horizontal axis, the existing DGMMs are mentioned. Hence, the cel s in Table 1 show the mapping of the DGMMs from the literature with our set of DGCs. Green cel s indicate that the DGCs in the literature are validated, and orange cel s indicate those DGCs that are not validated. An empty cel means that our DGC is not found in the DGMM found in the literature. The comparison demonstrates that existing DGMMs use different sets of DGCs and none of the J. R. Merkus, R. W. Helms, R. J. Kusters: Data Governance Capabilities; Empirical Validation in Case Studies of Large Organisations 39 DGMMs is as exhaustive as our list of DGCs. Furthermore, table 1 also illustrates the majority of these DGCs have yet to undergo empirical validation. This study aims to empirically validate the comprehensive set of 34 DGCs within large organisations. The research method applied for the validation of the DGCs is presented in the next section. Table 1: DGMM capability sets comparison (Merkus et al., 2021) Data Governance Capabilities a ia Merkus et al. an tan fae vera ed Ri Ri Perm Dasgupta Her Olai 2021 2009 2017 2018 2019 2019 2019 Establish Leadership Establish & manage Communicate Establish & manage Train Establish & manage culture Establish & manage awareness Quantify data value Align with the business Formulate data strategy Make business case Set goals & objectives Establish accountability Establish decision making authority Establish committees Establish roles & responsibilities Establish data stewardship Establish policies, principles & procedures Establish KPI's Establish performance management Establish Monitoring Establish Auditing Manage processes Manage organisation Manage data Manage metadata Manage risk Manage issues Establish & manage DG tools Establish & manage security & privacy Establish & manage Data Technology Organize people Align & integrate data Contract data sharing agreements Comply with regulations Government 5 3 6 10 6 8 Validated in practice Not validated in practice 40 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3 Method To empirically validate the DGCs for relevance, we have selected the case study strategy as our research approach. According to Yin (2014), a case study is "an empirical method that investigates a contemporary phenomenon (the case) in-depth and within its real-world context." Using the case study approach enabled us to identify DG activities in practice (practices) that make up each of the 34 DG capabilities we derived from the literature and hence to validate the capabilities. Furthermore, other researchers in our field recommend using case studies with expert interviews to design and evaluate capability models (Legner et al., 2020). Since validation of a single capability takes considerable time. It was decided to divide the DGCs among a group of researchers that used different organizations as case studies. The case organisations were found using convenience-based sampling, i.e. organisations in the networks of the researcher. However, each organisation should have at least 500 employees and there should be a need for data governance to participate in the research. Semi-structured expert interviews were used as the main method for data col ection. The interviewees were DG experts, which are individuals with at least five years of experience in DG, data management, or similar positions. To structure and compare the results of the interviews, pre-defined questionnaires were used, al owing open-ended questions to facilitate in-depth discussions. To further facilitate the interview process, Maturity Models (MMs) were developed for each DGC by means of a scoping literature review. The aim of the literature review was to identify the subcapabilities of each DGC, which wil form the basis of a DGC-specific MM (Munn et al., 2018). The resulting MMs have been used during the interviews to uncover DG practices, or reveal new DGCs, and thus provide experience-based information to support the validation of the DGCs. We validated each subcapability by an expert’s work experience with DG practices in large organisation. By validating each subcapability, we validate the overarching DGC of which subcabilities are part of. J. R. Merkus, R. W. Helms, R. J. Kusters: Data Governance Capabilities; Empirical Validation in Case Studies of Large Organisations 41 Finally, all empirically validated DG subcapabilities were categorised using a card sorting approach with applying the Metaplan technique (Howard, 1994; Spencer & Warfel, 2004). Using the original set of 34 DGCs as a reference, we eliminated any misconceptions that may have arisen during the scoping literature reviews. Furthermore, it helped to reveal new DGCs and hence to enrich the set of DGCs from literature with DGCs found in practice. 4 Results A total of 16 researchers each conducted a study of one or two DGCs in 19 large organisations over the course of a five-year period. Initial y, each researcher conducted a scoping literature review to identify the relevant subcapabilities for DGCs. Examples of subcapabilities found for DGC Align and integrate external data are (1) interaction and cooperation processes management, (2) Standardise data exchange, (3) Policies for data integration and use, (4) Data provider management, and (5) improve customer satisfaction. Next, each researcher designed an MM for their DGCs using the same design methodology and five stages of maturity levels, yielding 16 distinct MMs (Becker et al., 2009; Rosemann et al., 2004). Thirdly, three to five respondents were interviewed for the validation of each DGC, generating 70 expert interview reports, with an average interview duration of approximately two hours and up to four hours in some cases. These reports provide the basis for the validation of all DG capabilities. The interview reports are accessible from the author. The participating organisations al employ more than 500 employees, reflecting the need for governance awareness in such large organisations. Additional y, the participating organisations represent a diverse range of business sectors, al owing for the measurement of the DGCs and DG in various business activities, and operate at various geographical scales, as indicated in Table 2 Case organisations. 42 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 2: Case organisations Sector \ Scale International National Regional Total Bank 1 1 Bio industry 2 2 Education 2 2 Energy 1 1 2 Government 4 4 Insurance 1 1 2 Manufacturing 1 1 2 Retail 1 1 Wholesale 2 2 Union 1 1 Total 7 9 3 19 Additional y, DG jobs were scarce ten years ago as the field of DG only began to gain recognition in 2007 (Otto et al., 2007). As a result, some interviewed experts gained relevant experience in DG-related employment, such as data management. Despite these pragmatic adjustments, our research was carried out in accordance with the previously described methodology. In a peer-led card sorting exercise, we categorized al 231 subcapabilities derived from the 16 substudies using the Metaplan technique described in the methodology section. To eliminate misconceptions for more unambiguous language, we reclassified 47 subcapabilities to other existing DGCs as they fit better with those DGCs. Our analysis revealed no duplicates among the subcapabilities, but six subcapabilities had to be redefined to better differentiate them from each other. Additionally, the researchers identified ten suggestions for new subcapabilities and one suggestion for a new DGC during their substudies. Card sorting revealed that al eleven suggestions were addressed in one of the other substudies. So, no additional new (sub-)capabilities were identified. The validation of the DGCs in interviews, along with a subsequent hybrid card sorting exercise, resulted in the outcomes depicted in Table 3 DGC validation results. This table lists in column 1 al 34 DGCs in the DGC model, sorted according to Table 1. Column 2 lists the number of DG subcapabilities per DGC. Column 3 lists the number of case organisations where the subcapabilities have been validated. Column 4 lists the number of interviews in which a DG subcapability has been discussed. Column 5 lists the number of DG practices per DG subcapability noted during the interviews. J. R. Merkus, R. W. Helms, R. J. Kusters: Data Governance Capabilities; Empirical Validation in Case Studies of Large Organisations 43 We empirically validated each subcapability by (a) having the researchers registered a DG expert’s experience and reasoning of a DG practice happening in the organisation of his employer by the researchers during the interviews, (b) having the author analyse the transripts of the interviews for DG practices and reasoning for each subcapability after the interviews, and (c) finally, having the author discuss the findings with the peer authors. This resulted in a database of DG practices, including their reasoning when known, sorted per subcapability and the overarching DGC. So, when a DG practice occurred in a large case organisation and already one relevant DG practice was registered, the number of practices is irrelevant because we conducted qualitative research, the subcapability is empirically validated, hence empirically validating its overarching DGC. For example, for DGC Manage Metadata, we revealed three subcapabilities from literature; metadata, metadata management, and data standards. Next, we validated these three subcapabilities in the form of a maturity model in three different, large case organisations during 16 interviews with one DG expert each. During those interviews, the researchers noted eight different DG practices. Examples of the DG practices for each subcapability are a.o. (a) the presence of a data dictionary, or (b) business definitions being aligned with technical data definitions by data lineage, (c) or metadata management being administered in a central system. The reasoning we found for each subcapability is resp. (a) a data dictionairy is needed to integrate processes, mutual y understand what data means and align data between departments, legal requirements, and internal control reasons, and (b) integration of departments and divisions but also classification for privacy (c) to run queries. Consequently, these three DG practices validate the three subcapabilities, hence validating the DGC Managing metadata. Our overall research results show that each of the 34 DGCs has been validated by at least one practice, but most DGCs are validated by many more DG practices. All 34 DGCs are validated by a total 840 DG practices. 44 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 3: DGCs validation results DG Capabilities # DG validated in # discussed in # # DG practices Subcapabilitie Cases interviews Establish Leadership 1 1 3 3 Establish & manage culture 4 3 13 15 Establish & manage awareness 4 4 10 15 Establish & manage Train 8 3 14 36 Establish & manage Communicate 13 7 23 44 Specify data value 19 6 24 53 Set goals & objectives 6 3 11 20 Make business case 1 1 3 3 Formulate data strategy 4 4 12 13 Align with the business 14 8 24 44 Establish roles & responsibilities 13 8 24 53 Establish policies, principles, procedures 11 7 31 51 Establish performance management 3 1 5 15 Establish Monitoring 1 1 3 4 Establish KPI's 1 1 3 2 Establish decision making authority 10 8 24 35 Establish data stewardship 5 3 13 28 Establish committees 1 1 3 4 Establish Auditing 7 4 18 23 Establish accountability 1 1 5 3 Manage risk 8 6 26 36 Manage processes + lifecycle 14 6 24 43 Manage organisation 7 3 11 23 Manage metadata 3 3 16 8 Manage issues 1 1 3 4 Manage data 20 5 19 74 Setup security & privacy 11 4 18 45 Setup IT 8 5 21 24 Setup DG tools 1 1 3 1 Organize people 9 3 11 23 Contract data sharing agreements 7 2 5 34 Align & integrate data 5 5 16 21 Comply with regulations 9 6 28 36 Establish environmental response 1 1 3 4 231 840 Significantly, the majority of 26 out of 34 DGCs have been validated with sometimes many subcapabilities and even more DG practices. The much higher number of expert DG practices in column five #DG practices is caused by the number of three to five experts being interviewed per subcapability. And although not al experts recognized each subcapability in practice, each subcapability was validated (sometimes by many experts). The many different subcapabilities in column four #DG Subcapabilities are the result of the separately conducted literature reviews. However, nine DGCs were validated by only one relevant subcapability. Based on our overall research results, we conclude that all 34 DGCs found in literature are valid in practice. Furthermore, we did not identify any new DGCs, as the new DGCs suggested in interviews were already covered by the existing DGCs. This does not imply it can be used as a normative model that is fully applicable in all situations. Local context may impact this. But it can be used as a reference providing helpful suggestions for an individual organisation. J. R. Merkus, R. W. Helms, R. J. Kusters: Data Governance Capabilities; Empirical Validation in Case Studies of Large Organisations 45 5 Discussions, Limitations & Conclusions Our findings reveal the following remarkable outcomes, thereby adding new knowledge. First, the existence in the practice of large organisations of certain new DGCs which we predicted earlier based on our theoretical DGC model in Table 1 (J. R. Merkus et al., 2021). Moreover, the findings of this study align with the capability groups and clusters in that DGC model, and further validate several previously unknown capability groups such as Leadership, Culture, Communication, and Value Chain, with substantial evidence. Remarkably, none of the 17 substudies did identify any new DGCs in practice, confirming the comprehensiveness of our theoretical DGC model. Second, the DGCs from previously unvalidated DGMMs present in literature have been subject to validation too. The results of our research have validated these DG capabilities with empirical evidence from practice. So, our research confirms the validity of the few existing, empirically unvalidated DGMMs in literature too. Third, framing our research results against the DGC model from Table 1 results in the DGC T-model as reflected in Figure 1 Data Governance Capabilities T-model (J. R. Merkus et al., 2021). This model reflects the 34 DGC concepts which make up DG according to our findings. In addition, Figure 1 groups the DGC concepts in a T- shaped model according to the Generic Capabilities Reference model from our earlier research, and with a division of the DGCs into more strategic, tactical, and operational groups according to an organisational chart (J. Merkus et al., 2020; Mintzberg, 1980). Given the limitations of our study, we have identified some shortcomings. Firstly, the internal validity of our research was improved through the use of case study methodology, and construct validity was strengthened through the scrupulous administration of the results from semi-structured questionnaires. However, each DGC was validated separately and the entire set of DGCS was not validated as a whole. And although the DGCs were selected using a reference model, the underlying concepts of each DGC are based on literature, and each DGC has been validated individually with practical evidence, the internal validity of the set of 34 DGCs as a whole could be improved in further research. Secondly, the external validity of the results could be enhanced, even though the DGCs have been validated 46 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY in various organisations. Therefore, further research to validate the set of DGCs as a whole is necessary, with case studies as a suitable research strategy again. In conclusion, we can deduce from the results of our study that al the known DGCs from literature have been validated in practice. Furthermore, no other DGCs were discovered during the empirical validation, which suggests that, although a claim for completeness can never be proven, at least the most relevant capabilities have been identified. This set of DGCs can be used as a reference to construct a locally relevant measure or benchmark for DG e.g. a DGMM. The theoretical implication is that our research adds new knowledge with the empirically validated comprehensive set of 34 DGCs. Further research could focus on the empirical validation of this set of DGCs as a whole. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. J. R. Merkus, R. W. Helms, R. J. Kusters: Data Governance Capabilities; Empirical Validation in Case Studies of Large Organisations 47 Acknowledgments The case studies for this research were conducted by the following researchers (in alphabetical order): F.Alberts MSc, J.Amghar MSc, F.Badloe MSc, R. van den Bos MSc, J.G.Boomgaarden MSc, H.Breunissen MSc, S.Cerimovic MSc, Ö.Karaca MSc, P.Kroos-Kerpershoek MSc, S.Martin MSc, M.Meijer MSc, M.van Sambeek MSc, J.Soukhou MSc, J.Standhart MSc, M.Veenstra MSc, K.Wouters MSc. Paragraphs were submitted to AI-based translators Grammerly, Deepl and ChatGTP for language improvement only. No content was added by the tools. References Abraham, R., Schneider, J., & Brocke, J. vom. (2019). Data Governance : A conceptual framework , structured review , and research agenda. International Journal of Information Management (IJIM), 49(December 2019), 424–438. Alhassan, I., Sammon, D., & Daly, M. (2016). Data governance activities: an analysis of the literature. Journal of Decision Systems, 25(S1), 64–75. Alhassan, I., Sammon, D., & Daly, M. (2018). Data governance activities: a comparison between scientific and practice-oriented literature. Journal of Enterprise Information Management, Vol 31(2), 300–316. Alhassan, I., Sammon, D., & Daly, M. (2019). Critical Success Factors for Data Governance: A Theory Building Approach. Information Systems Management, 36(2), 98–110. Bandara, W., van Looy, A., Merideth, J. C., & Meyers, L. (2020). Holistic Guidelines for Selecting and Adapting BPM Maturity Models (BPM MMs). In D. Fahland, C. Ghidini, J. Becker, & M. Dumas (Eds.), Business Process Management Forum: BPM Forum 2020 Proceedings (pp. 263–278). Springer. Becker, J., Knackstedt, R., & Pöppelbuß, J. (2009). Developing Maturity Models for IT Management – A Procedure Model and its Application. Business & Information Systems Engineering, 1(3), 213–222. Brennan, R., Attard, J., & Helfert, M. (2018). Management of data value chains, a value monitoring capability maturity model. ICEIS 2018 - Proceedings of the 20th International Conference on Enterprise Information Systems, 2(January), 573–584. Dasgupta, A., Gil , A., & Hussain, F. (2019). A conceptual framework for data governance in IoT-enabled digital IS ecosystems. In Proceedings of the 8th International Conference on Data Science, Technology and Applications (DATA 2019), Data, 209–216. Heredia-Vizcaíno, D., & Nieto, W. (2019). A Governing Framework for Data-Driven Small Organizations in Colombia. In Á. Rocha, H. Adeli, L. Reis, & S. Costanzo (Eds.), New Knowledge in Information Systems and Technologies. WorldCIST’19 2019. Advances in Intelligent Systems and Computing (Vol. 930, Issue 1, pp. 622–629). Springer. Howard, M. S. (1994). Quality of Group Decision Support Systems: a comparison between GDSS and traditional group approaches for decision tasks. Technical University Eindhoven. Jagals, M., Karger, E., & Ahlemann, F. (2021). Already grown-up or stil in puberty? A bibliometric review of 16 years of data governance research. Corporate Ownership & Control, 19(1), 105– 120. Legner, C., Pentek, T., & Otto, B. (2020). Accumulating design knowledge with reference models: Insights from 12 years’ research into data management. Journal of the Association for Information Systems, 21(3), 735–770. Lis, D., & Otto, B. (2021). Towards a Taxonomy of Ecosystem Data Governance. In Proceedings of the 54th Hawaii International Conference on System Sciences (HICSS 2021), 6067–6076. 48 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Merkus, J., Helms, R. W., & Kusters, R. J. (2019). Data Governance and Information Governance : Set of Definitions in Relation to Data and Information as Part of DIKW. ICEIS 2019 - Proceedings of the 21th International Conference on Enterprise Information Systems, 12. Merkus, J. R., Helms, R. W., & Kusters, R. J. (2020). Reference Model For Generic Capabilities In Maturity Models. In Proceedings of the 2020 12th International Conference on Information Management and Engineering (ICIME 2020), 10–17. Merkus, J. R., Helms, R. W., & Kusters, R. J. (2021). Data Governance Capabilities; Maturity Model design with Generic Capabilities Reference Model. In Proceedings of the 13th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (KMIS 2021), 102–109. Mintzberg, H. (1980). Structure in 5’s : A Synthesis of the Research on Organization Design. Management Science, 26(3), 322–341. Munn, Z., Peters, M. D. J., Stern, C., Tufanaru, C., McArthur, A., & Aromataris, E. (2018). Systematic review or scoping review? Guidance for authors when choosing between a systematic or scoping review approach. BMC Medical Research Methodology, 18(1), 1–7. Olaitan, O., Herselman, M., & Wayi, N. (2019). A Data Governance Maturity Evaluation Model for government departments of the Eastern Cape province, South Africa. South African Journal of Information Management, 21(1), 1–12. Otto, B., Wende, K., Schmidt, A., & Osl, P. (2007). Towards a Framework for Corporate Data Quality Management. In Proceedings of 18th Australasian Conference on Information Systems (ACIS 2007), Dec 2007, 916–926. Paulk, M. C., Curtis, B., Chrissis, M. B., & Weber, C. V. (1993). wrong The capability maturity model for software. In Software Process Improvement. Permana, R. I., & Suroso, J. S. (2018). Data Governance Maturity Assessment at PT. XYZ. Case Study: Data Management Division. In Proceedings of 2018 International Conference on Information Management and Technology (ICIMTech 2018), 15–20. Rivera, S., Loarte, N., Raymundo, C., & Dominguez, F. (2017). Data Governance Maturity Model for Micro Financial Organizations in Peru. In Proceedings of the 19th International Conference on Enterprise Information Systems (ICEIS 2017), 203–214. Rosemann, M., de Bruin, T., & Hueffner, T. (2004). A Model for Business Process Management Maturity. In Proceedings of 15th Australasian Conference on Information Systems (ACIS 2004), Paper 6, 1–7. Spencer, D., & Warfel, T. (2004). Card sorting: a definitive guide. In Boxes and arrows (p. 2). http://www.iimagineservicedesign.com/wp-content/uploads/2015/07/Card-sorting-a- definitive-guide-«-Boxes-and-Arrows.pdf van Looy, a., de Backer, M., & Poels, G. (2011). Questioning the Design of Business Process Maturity Models. In Proceedings of the 6th SIKS Conference on Enterprise Information Systems, 51– 60. van Looy, A., de Backer, M., & Poels, G. (2011). Defining business process maturity. A journey towards excel ence. Total Quality Management and Business Excellence, 22(11), 1119–1137. Weber, K., Cheong, L., Otto, B., & Chang, V. (2008). Organising Accountabilities for Data Quality Management-A Data Governance Case Study. Data Warehousing, 347–362. THE ROLE OF DIGITALISATION IN CHANGING THE BUSINESS MODELS IN LOGISTICS: CASE OF ROPAX PORTS YIRAN CHEN,1 ANASTASIA TSVETKOVA,2 KRISTEL EDELMAN,2 IRINA WAHLSTRÖM,1 MARIKKA HEIKKILÄ,2 MAGNUS HELLSTRÖM1,3 1 Åbo Akademi University, Laboratory of Industrial Management, Turku, Finland yiran.chen@abo.fi, anastasia.tsvetkova@abo.fi, Irina.wahstrom@abo.fi 2 University of Turku, Centre for Collaborative Research, Turku, Finland kristel.edelman@utu.fi, marikka.heikkila@utu.fi 3 University of Agder, Department of Working Life and Innovation, Kristiansand, Norway magnus.hel strom@abo.fi This article explores digitalisation’s potential to change traditional business models in the context of RoPax (roll-on, roll- off passenger vessels) ports in four Northern European countries. The study examines digitalisation’s role in addressing business model change drivers, focusing on the perspective of port authorities (PAs). While digitalisation slowly affects operations at RoPax ports and PAs’ business models, the research data exhibits no radical business model innovation. Instead, the findings indicate that PAs introduce new digitalisation and business activities, potentially leading to business model renewal. However, the current digitalisation is inefficient, as technology providers lack an in-depth Keywords: understanding of the port business and its ecosystem, and PAs port have scarce knowledge of digitalisation’s business impact. The authority, article concludes that connecting digitalisation strongly to the business model, business model and strategic renewal is the way to overcome this maritime, chal enge. digitalisation DOI https://doi.org/10.18690/um.fov.4.2023.4 ISBN 978-961-286-751-5 50 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Sea logistics is known for being among the slowest to adopt digital technologies (Acciaro et al., 2018; Transport Intelligence, 2019) and relies heavily on old communication and data exchange methods. However, digitalisation is considered a disruptive driver in the industry due to the opportunities for optimising logistics and integrating actors involved in transportation chains (Tsvetkova et al., 2021). The potential benefits of aggregating and analysing data on maritime transportation are vast (Watson et al., 2021), and various digital solutions are being implemented or developed (Brümmerstedt et al., 2017; Tijan, Jović, Aksentijević, et al., 2021; Tsvetkova et al., 2021). As critical links in sea logistics and other transport modes, ports can renew their business models through digitalisation, reducing shipping emissions (Haraldson et al., 2021) and improving supply chain efficiency. Ports have already been conceptualised as ecosystems (de Langen, 2021; Haraldson et al., 2021; Watson et al., 2021), and port authorities (PAs) are increasingly seen as ecosystem integrators or orchestrators (Caballini et al., 2009), making their critical role in leading digital transformation undisputable (von Malmborg, 2004). Passenger ports also have high public interest and are often public enterprises that local municipalities fully or partly own, making them an interesting public–private collaboration platform. They play a key role in transport digitalization and sustainable transitioning in maritime transportation (Damman & Steen, 2021; Del Giudice et al., 2021). This article explores digitalisation’s potential to change incumbent business models in a highly institutionalised, asset-heavy industry. Six RoPax (roll-on, roll-off passenger vessels) ports in four Northern European countries, handling both vehicle and passenger transportation, were interviewed on their digitalisation efforts and goals. Due to low regional transport intermodality, these ports can be seen as a continuation of road networks and can benefit from digitalisation differently than mega container ports, which have been extensively studied in light of digitalisation (Brümmerstedt et al., 2017; Port of Rotterdam, 2023). The current study provides insight into the digitalisation process and business model innovation from PAs’ perspective. Thus, our research question is, “How do incumbent firms in logistics use digitalisation in their business model innovation?” Y. Chen, A. Tsvetkova, K. Edelman, I. Wahlström, M. Heikkilä, M. Hel ström: The Role of Digitalisation in Changing the Business Models in Logistics: Case of RoPax Ports 51 To answer this, we explore the drivers for business model innovation in the chosen context and understand the general directions for business model change in RoPax ports. Then, we pinpoint digitalisation’s role in addressing those drivers through business model innovation. 2 Literature Review 2.1 Digitalisation’s Role in Changing Business Models Business model innovation (BMI) has been defined as the designed, novel, and nontrivial changes to the key elements of a company’s business model and architecture linking these elements (Foss & Saebi, 2017). The widespread diffusion of digital technologies has become one of the key drivers and enablers of BMI at the firm level (Holmström et al., 2019). Exploiting digital technologies through innovative business models has been distinguished from innovation in which new technologies are developed within the product innovation process (Caputo et al., 2021; Cozzolino et al., 2018). Furthermore, digitalisation grants opportunities for BMI beyond firm-level changes to cross multiple industries and ecosystems (Kamalaldin et al., 2021; Leminen et al., 2020; Sjödin et al., 2020). The conceptual relationship between digitalisation and corresponding changes in business models has been extensively explored (Caputo et al., 2021); studies of digitalisation in different industrial contexts show that digitalisation and BMI is context-dependent (Benghozi & Salvador, 2015; Kamalaldin et al., 2021; Vendrell-Herrero et al., 2017). Caputo et al. (2021) note that the proliferation of new business models characterised by a high degree of digital innovation has concerned innovative and traditional sectors, which are not characterised by high degrees of technological investment. Furthermore, while BMI is often necessary to reap digitalisation’s benefits, most incumbent firms across industries are il -prepared to benefit from digital transformation (Parida et al., 2019). Digitalisation in the maritime industry is considered slow (Acciaro et al., 2018; Transport Intelligence, 2019). Nevertheless, indications show that digitalisation in the maritime sector can change the relationships among the supply chain actors, restructure the ecosystems, and create an opportunity for new business models or 52 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY changes in incumbent business models (Fruth & Teuteberg, 2017; Heikkilä et al., 2022; Tijan et al., 2021 a; Tsvetkova et al., 2021). 2.2 Business Model Innovation and Digitalisation in Ports Ports are areas with maritime and hinterland access that have developed into logistics hubs (Van der Lugt & De Langen, 2007). Ports are managed by PAs, that often operate as landlords (World Bank, 2007). In most cases, a PA is a publicly owned company by a local municipality or state and manages and develops the port area, with income mainly based on land rent and port dues. Tenants in a port are often companies involved in port operations or logistics activities (Van der Lugt & De Langen, 2007). Thus, landlord PAs aim to balance between public (PA, municipality) and private (port industry) interests (World bank, 2007). Regulation and business environment alternations drive changes in ownership of PAs (Notteboom et al., 2022; Rönty et al., 2011), pressures come from numerous actors, such as NGOs; governments; municipalities (Verhoeven, 2010); and business partners and customers, such as logistic companies, cargo owners, and passengers (Notteboom et al., 2022, Chapter 4.1). Furthermore, a PA’s business model depends on numerous other factors, such as its traffic profile, typologies of cargo and passenger, location, existing facilities, and infrastructural conditions (Burns, 2014, p. 22; Paixão Casaca & Lyridis, 2022). Environmental, technological, geographical, and demographic changes also drive today PAs’ business model evolution (Vonck et al., 2021). The mentioned factors influence a PA’s strategic decision-making (Haraldson et al., 2021). Given such a changing landscape, the strategic responses are, for example, becoming ful -fledged partners in the logistics chain, restricting a port’s role to supporting activities or entirely disappearing from the scene (Heaver et al., 2010). Digitalisation offers new opportunities for a PA’s BMI. Scholars proved the importance of PAs’ initiatives when it comes to digitalisation, considering different types of port governments models (Tijan, Jović, Panjako, et al., 2021). However, limited studies have addressed the relevance of digitalisation, the PA’s business model, and ecosystem changes (Henríquez et al., 2022; Hirata et al., 2022). A major focus on port digitalisation has been set in mega and large container ports (Haezendonck & Langenus, 2019; Henríquez et al., 2022). Ports with other traffic Y. Chen, A. Tsvetkova, K. Edelman, I. Wahlström, M. Heikkilä, M. Hel ström: The Role of Digitalisation in Changing the Business Models in Logistics: Case of RoPax Ports 53 profiles, such as RoPax ports, are scarcely reviewed. Compared to mega and large container ports, these ports are limited regarding resources and investments (Del Giudice et al., 2021; Inkinen et al., 2019; von Malmborg, 2004) and face chal enges on understanding digitalisation’s influence on strategic business development (Inkinen et al., 2019). Thus, our study complements the current research agenda’s gap and focus on digitalisation in RoPax ports. 3 Methodology 3.1 Research Design The theoretical gap and emerging status of port digitalisation led researchers to adapt qualitative research with a multi-site case study approach (Creswell, 2007), where the RoPax port is considered a case studied in the diverse sites where these ports are located. A multi-site case study enables the understanding of a specific phenomenon that is merged into the context, adding to the phenomenon’s complexity (Audet & d’Amboise, 2001; Gillham, 2000; Yin, 2018). Similarly, to a multiple case study, which enables comparison (Eisenhardt & Graebner, 2007; Gibbert et al., 2008; Gioia et al., 2013), the case is compared between sites with the possibility for cross-site generalisations. This comparison aims to enhance understanding digitalisation in relationship with business models in RoPax ports. RoPax ports gained scarce research attention (see section 2.2), our ultimate research interests are commonalities in use of digitalisation. Our research logic takes an abductive approach, pursuing the iterative matching and simultaneous evolution between theories and empirical observations (Dubois & Gadde, 2002). By applying cross-site comparison (Creswell, 2007), this research approach enables identifying variations within the same case. 3.2 Selection of Sites The study includes six RoPax ports in four Northern European countries. Each port offers regular liner traffic operated by at least two shipping lines to more than one international destination port. The ports and countries were chosen based on generic 54 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY similarities such as a common trading area, ownership base, high share of short shipping services, business model and culture, and high level of national digital infrastructure readiness, integration, and adaption. 3.3 Data Collection and Analysis The primary research data was col ected with semi-structured interviews conducted with PAs, with an average duration of one hour. Secondary data served as foundation for drafting interview topics regarding business model transitions from a PA’s perspective, which comprises press releases, publications, statistics, and strategic and project reports related to PAs’ development. Drivers for business model, ecosystem, and activity changes were discussed during the interviews. For each selected port, one or more managerial-level representatives joined a discussion of selected topics. The interviews were recorded, transcribed, and documented. Table 1 presents the case ports’ characteristics. Table 1: Ports’ Characteristics PA Traffic profile Interviewees’ roles Data format Alpha Cargo + Passenger Port Development Strategist Interview Beta Cargo + Passenger Technical Director Workshop and Operation Manager interview IT Specialist Gamma Cargo + Passenger Head of Business Interview Development Delta Cargo+ Passenger Development Manager Interview Epsilon Passenger- Business Development Interview dominated Zeta Cargo + Passenger Chief Operations Officer Interview Data analysis followed the deductive category application and further inductive category development (Mayring, 2004), which can be label ed a directed approach to content analysis aimed to support or extend the theory (Hsieh & Shannon, 2005). The labels were later compared to existing theories regarding BMI, digitalisation, and literature related to port business models. The abductive iteration involves identifying mismatched empirical-theoretical concepts, which were further studied by aligning with other theoretical explanations. Y. Chen, A. Tsvetkova, K. Edelman, I. Wahlström, M. Heikkilä, M. Hel ström: The Role of Digitalisation in Changing the Business Models in Logistics: Case of RoPax Ports 55 Three researchers independently analysed the interview transcripts and secondary data to attain reliable categorisation and identified content pertinent to one of the topics defined in the analysis framework. All researchers then compared notes and agreed on the categories within each topic by merging some of them and revisiting the original texts in case of dissimilarities in identified drivers or enablers for BMI. 4 Findings 4.1 Drivers for Business Model Change in RoPax Ports Over the recent decades, combining short-sea passenger and cargo transportation has become an established and common transport concept (Marcadon, 1998). RoPax vessel design is based on roll-on and roll-off features, enabling efficient loading and unloading of wheeled commercial vehicles (trucks, trailers) and passenger cars and the capability to accommodate passengers. The case ports have short vessel turnaround times, low cargo standardisation, and are further chal enged by additional services passenger transportation. The analysis of the interviews indicates that managing and combining the flows and services of cargo and passengers leads to several specific chal enges, al leading to digitalisation opportunities. On the passenger side, typical chal enges are related, for instance, to non-integrated information on transport connections and services as well as crowding and queuing in the passenger terminal. Conversely, the multimodal transport chain is hampered by uncoordinated road traffic pulses and congestion at ship arrivals and departures inside the port, associated urban area, and its main approach roads. This problem is less likely to diminish with expanding residential and recreational urban areas and the associated shrinkage of the port areas. Furthermore, the vessels are likely to grow in cargo capacity, further worsening the overcrowding of the road network, while traffic jams give rise to idling and unnecessary emissions (Wahlström et al., 2022). Further, passenger transportation forms a vital revenue source for the shipping company and the PA. Several PAs mentioned the detrimental effects of trade shocks, such as the one caused by the COVID-19 pandemic, taking passenger transportation to a total standstill, impacting shipping services availability and, in most cases, 56 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY dramatic plunges in PAs’ revenues with potentially long-term consequences and chal enges to retain the main customers’ regular fleet capacity. 4.2 Changes in the Business Models of RoPax Ports Traditional y, RoPax ports’ principal customer base comprises shipping lines and port operators. Other relevant actors encompass various authorities and land-based passenger and cargo transporting companies. Meanwhile, passengers and road haulage companies are shipping lines’ main customers. A RoPax PA’s primary income derives from vessel, passenger, and cargo fees and the rental income of facilities. Conversely, the main expenditures typically include human resources costs and land leases paid to the municipality. A RoPax PA’s value proposition is providing required infrastructure, facilities, and quay-side vessel services, enabling safe, efficient, and timely port calls and smooth cargo and passenger flows. A PA’s strategic renewal is largely driven by the drivers mentioned in 4.1, and manifested in changes in their business models to a varied degree. Following our inductive approach, we identified several recurring topics. The changes in the RoPax PAs’ business models (see Table 2) are common for two or more studied ports and concern changes in value-creating activities, earning logic, revenue streams, and key partnerships, using resources for value creation. As presented in the table, the most radical change concerns the business expansion and further integration in logistics and supply chain. With the construction of new passenger terminals, several PAs are taking over the ownership of passenger terminals. Combined with the increasing perception of passengers as PA’s customers, the ownership transition of the terminal premises brings many changes to the business models, such as new value propositions for passengers and new activities to ensure customer satisfaction. In addition, PAs are also searching for new values for existing facilities, such as using terminal buildings as a venue outside of traditional port operation use. Y. Chen, A. Tsvetkova, K. Edelman, I. Wahlström, M. Heikkilä, M. Hel ström: The Role of Digitalisation in Changing the Business Models in Logistics: Case of RoPax Ports 57 Table 2: Changes in PAs’ Business Models Alpha Beta Gamma Delta Epsilon Zeta Business New values of x x x expansion port facilities New offerings x x x for passengers Ownership x x and operation of passenger terminals Integration in Enhanced x x x x logistics and communicatio supply chain n Enabler of x x digitalisation by other port actors Environmental Green x x x x impact incentives for reduction port users Alternative x x x x fuel infrastructure Efficiency Cargo and x x improvement passenger flows separation Increased x x x x x x facility utilisation rate Social Decreased x x responsibility human work Expanding x x safety and security measures Further integration with other actors enables optimised supply and logistics chains. This entails enhanced communication of port activities and service offering for diverse logistics actors. It is worth highlighting that PAs could also become enablers and integrators of logistics chains digitalisation, as mentioned by several interviewees. Because ports are critical transport hubs that could affect the overal logistics performance. 58 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Increased integration with other logistics actors is also crucial for reducing shipping’s environmental impact. There was a discussion with PA Alpha that a port’s emission mitigation scope should be extended to achieve overall logistics emission reduction. Several PAs have implemented or plan to implement green incentives for its customers, such as pricing models and rebates based on the visiting vessel’s environmental performance. Green corridors are considered another important initiative for shipping emission reduction; PAs play a key role in establishing those. In this vein, several PAs are planning to develop alternative fuel infrastructure. PAs focus on efficiency improvement to retain competitive, enhance and maximise the utilisation rate of port infrastructure and assets, including facilities, warehouses, quays etc. Thus, automating various processes is high on the agenda. Many RoPax ports, located close to city centres, face increasing threat of space limitations, and hence a potential solution would be to geographical y separate and relocate potential non-wheeled cargo from RoPax associated passenger and cargo services. Automation in operations rationalises and increases the efficiency, safety and security of various manual activities and processes, hence optimising, the workers’ work conditions and safety.as part of PA’s social responsibility. 4.3 Digitalisation Efforts at RoPax Ports We see different efforts and digital tools applied in ports. Our analysis indicates that PAs who have or are about to establish their digitalisation strategies or roadmaps are actively contributing to the understanding of how digitalisation could support their business activities, enabling the avoidance of a “digitalised mess”, as one of the interviewees mentioned, and prioritising various ongoing and upcoming digitalisation projects. In general, we recognised four distinct strategic areas where digitalisation was implemented or planned in the RoPax ports: 1. Infrastructure management: This is the traditional focus of landlord PAs. Apart from regular maintenance work, the PAs growingly face the pressure of adjusting and upgrading infrastructure in response to growing trade volumes and vessel capacities. The case PAs had developed digital solutions to help monitor and manage security in the ports. Furthermore, they Y. Chen, A. Tsvetkova, K. Edelman, I. Wahlström, M. Heikkilä, M. Hel ström: The Role of Digitalisation in Changing the Business Models in Logistics: Case of RoPax Ports 59 increasingly utilise port infrastructure data to improve maintenance and facility services. PAs also showed interest in developing digital twins of the port infrastructure. 2. Traffic fluency: Space limitation and co-existence with expanding port cities is a growing concern, driving PAs to search for automated and digitalised solutions, ensuring smooth passenger and cargo flows, efficient operations, and timely port calls. The digitalisation degree of the case PAs varied markedly. However, digitalising logistics and supply chains has increasingly pressured PAs to adopt new technologies, such as automated check-in for passengers and vehicles. 3. Green transition: Today, the environmental regulation of the shipping industry together with national climate goals constitutes one of the key development areas in RoPax ports. PAs have progressively shifted to electrified equipment and machinery, greener fuels, and more energy-saving infrastructure to reduce the environmental impact. Berthing vessels have been the leading emitter of carbon dioxide and other air pollutants. Hence, instal ing and investing in automooring and an onshore power supply have occurred in most of the interviewed ports. 4. Data sharing between port-operation-related organisations: PAs are introducing new digital solutions to strengthen communication, electronic trade documents sharing, customs clearance, and integration with other supply chain actors. More ambitious solutions that span across entire supply chains were mentioned, however less often. These solutions aim at improved efficiency and sustainability, amongst others. The above digitalisation areas are summarized in Table 3. 60 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 3: Major Digitalisation Areas Alpha Beta Gamma Delta Epsilon Zeta 1) Improved infrastructure management Security monitoring and x x x x x x management Digitising infrastructure data x x x Infrastructure use management x 2) Enhanced traffic fluency Automated check-in (passengers) x x x x x x Automated check-in (vehicles) x x x Intel igent traffic management x x x Digital buoys/fairways x Autonomous vehicles x 3) Green transition Monitoring emissions and air x x quality Monitoring and optimising energy x x x use Automooring x x x x 4) Data sharing between port-operation-related organisations Communication and procedures x x x x x x among port actors Intel igent supply chain x 5 Discussion The findings affirm that digitalisation gradualy affects RoPax port’s operations and business models. At the beginning of the study, we expected digitalisation to drive PAs to innovate their business models. However, the data il ustrated that PAs are introducing new digitalisation and business activities, potentially leading to business model renewal. However, no radical business model innovations were identified. For instance, PA’s value proposition remains the same, as most of the case ports would remain landlord ports (World Bank, 2007) also in future. Thus, the discussion focuses on drivers for business model changes and how digitalisation supports these changes rather than business model innovations. Although some PAs implemented strategies or roadmaps for enabling the creation of a more holistic view of digitalisation. The current digitalisation is fragmented and may cause a “digitalised mess” (see section 4.3). If PAs digitalise without Y. Chen, A. Tsvetkova, K. Edelman, I. Wahlström, M. Heikkilä, M. Hel ström: The Role of Digitalisation in Changing the Business Models in Logistics: Case of RoPax Ports 61 contextualising strategically, the actual digitalisation benefit would be limited. For example, many solutions mentioned in section 4.3 are short-term project outcomes that are not part of a PA’s strategic plan. Conversely, PAs may have limited knowledge of how digitalisation may impact and improve their efficiency and business performance, potential y prolonging the decision-making process. For instance, several PAs mentioned the inability to comprehend the true advantages that solutions such as a digital twin would bring, especial y as the development and implementation are perceived to require significant resources, whilst the application’s benefits may remain undisclosed. This situation echoes previous studies on the maritime transport sector that emphasise there is a “lack of awareness of how digital transformation may af ect the business” (Tijan, Jović, Aksentijević, et al., 2021). As the interview analysis shows, for overcoming this challenge, job positions and responsibilities are established for developing digitalisation roadmaps and embedding digitalisation in strategic planning based on individual business development needs. We could observe that connecting digitalisation with business development needs could contribute for PA’s strategic business model renewal, such as acquiring relevant knowledge or planning digitalisation in a long-term context. Despite the “digitised mess” (see Section 4.3), digitalisation enables and drives several prominent business model changes for the case PAs, namely, supporting the formation of green logistic chains or becoming a digital infrastructure owner. Besides the more ‘physical’ (as opposed to digital) activities, such as developing alternative fuel infrastructure, ports provide awareness of emissions and pollution from port operations through monitoring solutions and communication with other logistic actors. Improving passenger and cargo flow efficiency through avoiding unnecessary idling, could also reduce emissions. PAs naturally achieve further integration with other actors alongside the logistic chain. Digitalisation aids this goal in several ways. Firstly, PA could facilitate digitalising documentation and information sharing between organisations. Secondly, several optimisation solutions, such as automated check-in and flow of passengers and vehicles and intelligent traffic management, support PAs in affecting the activities beyond a RoPax terminal to enhance port operations efficiency. 62 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY PAs could also be a central digital infrastructure provider for other organisations to improve port operations performance. In line with Hollen et al. (2015), we argue that RoPax ports seemingly explore complementary roles and activities in pursuing creating strategic value as landlord ports, i.e., the value for critical actors. Digitalisation integrates interorganisational logistic activities, increases traffic flow transparency, and trace emissions – the information basis for decision-making by PAs and other logistics actors. Digital solutions like autonomous roads and intelligent supply chains also contribute to expanding the role of ports in the hinterland of logistics chains (Gonzalez Aregall et al., 2018). In Table 4, we present the digital solutions implemented or planned to be implemented in studied RoPax ports. We also analysed digital solutions’ contribution (see section 4.3) to the business model changes discussed in section 4.2. Not surprisingly, PAs predominantly invest in digital solutions that allow to increase operation efficiency. These solutions mainly concern automation in the pursuit reduction of port operations’ time and costs, but also to decrease human work, thereby also addressing the changes related to social sustainability. 6 Conclusion We contribute to the literature on the business models of ports by explicating digitalisation’s role in aiding changes in the business models of RoPax ports. We also contribute empirical y with an account of digitalisation efforts in RoPax ports, which are less studied in the literature on digitalisation in ports and sea logistics. The study has several limitations. First, it focuses on a limited geographic context of Northern Europe. Further, although the change in business models is studied by discussing the current business models of ports and foreseeing changes in their business model, a more longitudinal study would be beneficial to uncover the business model evolution in this context. Y. Chen, A. Tsvetkova, K. Edelman, I. Wahlström, M. Heikkilä, M. Hel ström: The Role of Digitalisation in Changing the Business Models in Logistics: Case of RoPax Ports 63 Table 4: Types of Digital Solutions Planned and Implemented by RoPax PAs and Main Digitalisation Areas Contribution to business model change Digitalisation Digital Business Integrat- Environm- Efficiency Social areas solutions expansion ion in ental impact improvem- sustainabi- logistics reduction ent lity chains Improved Security x x infrastructure monitoring management and management Digitising port infrastructure data Infrastructure x use management Enhanced Automated x traffic fluency check-in for passengers and vehicles Intelligent x traffic management Digital buoys/ x x fairways Autonomous x x vehicles Monitoring x x emissions and air quality Green Monitoring x transition and optimising energy use Automooring x x Communica- x x Data sharing tion and between port- procedures operation- among port related actors organisations Intelligent x x x supply chain References Acciaro, M., Ferrari, C., Lam, J. S. L., Macario, R., Roumboutsos, A., Sys, C., Tei, A., & Vanelslander, T. (2018). Are the innovation processes in seaport terminal operations successful? 787–802. Audet, J., & d’Amboise, G. (2001). The Multi-Site Study: An Innovative Research Methodology. The Qualitative Report, 6(2), 1–18. Benghozi, P. J., & Salvador, E. (2015). How and where the R&D takes place in creative industries? Digital investment strategies of the book publishing sector*, 28(5), 568–582. 64 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Brümmerstedt, K., Fiedler, R., Flitsch, V., Jahn, C., Roreger, H., Sarpong, B., Saxe, S., & Scharfenberg, B. (2017). Digitalization of Seaports - Visions of the Future (C. Jahn, S. Saxe, & Fraunhofer CML, Eds.; p. 168). Fraunhofer Verlag. Burns, M. G. (2014). Port management and operations. CRC Press. Caballini, C., Carpaneto, L., & Parola, F. (2009). Italian port authorities approaching the post-reform: The Ligurian Case (pp. 191–207). Caputo, A., Pizzi, S., Pel egrini, M. M., & Dabić, M. (2021). Digitalization and business models: Where are we going? A science map of the field. Journal of Business Research, 123, 489–501. Cozzolino, A., Verona, G., & Rothaermel, F. T. (2018). Unpacking the Disruption Process: New Technology, Business Models, and Incumbent Adaptation. Journal of Management Studies, 55(7). Creswell, J. W. (2007). Qualitative inquiry and research design: Choosing among five approaches, 2nd ed. Sage Publications, Inc. Damman, S., & Steen, M. (2021). A socio-technical perspective on the scope for ports to enable energy transition. Transportation Research Part D: Transport and Environment, 91, 102691. de Langen, P. W. (2021). Seaports as Clusters of Economic Activities. In International Encyclopedia of Transportation (pp. 310–315). Elsevier. del Giudice, M., di Vaio, A., Hassan, R., & Palladino, R. (2021). Digitalization and new technologies for sustainable business models at the ship–port interface: a bibliometric analysis. 49(3), 410– 446. Dubois, A., & Gadde, L. E. (2002). Systematic combining: An abductive approach to case research. Journal of Business Research, 55(7), 553–560. Eisenhardt, K. M., & Graebner, M. E. (2007). Theory building from cases: Opportunities and chal enges. Academy of Management Journal, 50(1), 25–32. Foss, N. J., & Saebi, T. (2017). Fifteen Years of Research on Business Model Innovation: How Far Have We Come, and Where Should We Go? Journal of Management, 43(1). Fruth, M., & Teuteberg, F. (2017). Digitization in maritime logistics—What is there and what is missing? Cogent Business and Management, 4(1). Gibbert, M., Ruigrok, W., & Wicki, B. (2008). What passes as a rigorous case study? Strategic Management Journal, 29(13), 1465–1474. Gillham, B. (2000). Case study research methods. Bloomsbury Publishing. Gioia, D. A., Corley, K. G., & Hamilton, A. L. (2013). Seeking Qualitative Rigor in Inductive Research: Notes on the Gioia Methodology. Organizational Research Methods, 16(1), 15–31. Gonzalez Aregal , M., Bergqvist, R., & Monios, J. (2018). A global review of the hinterland dimension of green port strategies. Transportation Research Part D: Transport and Environment, 59, 23– 34. Haezendonck, E., & Langenus, M. (2019). Integrated ports clusters and competitive advantage in an extended resource pool for the Antwerp Seaport. Maritime Policy and Management, 46(1), 74– 91. Haraldson, S., Lind, M., Breitenbach, S., Croston, J. C., Karlsson, M., & Hirt, G. (2021). The Port as a Set of Socio-technical Systems: A Multi-organisational View. In Maritime Informatics (pp. 47– 63). Heaver, T., Meersman, H., Moglia, F., & van de Voorde, E. (2010). Do mergers and al iances influence European shipping and port competition? 363–373. Heikkilä, M., Saarni, J., & Saurama, A. (2022). Innovation in Smart Ports: Future Directions of Digitalization in Container Ports. Journal of Marine Science and Engineering, 10(12). Henríquez, R., Martínez de Osés, F. X., & Martínez Marín, J. E. (2022). Technological drivers of seaports’ business model innovation: An exploratory case study on the port of Barcelona. Research in Transportation Business & Management, 43, 100803. Hirata, E., Watanabe, D., & Lambrou, M. (2022). Shipping Digitalization and Automation for the Smart Port. Supply Chain - Recent Advances and New Perspectives in the Industry 4.0 Era [Working Title]. Y. Chen, A. Tsvetkova, K. Edelman, I. Wahlström, M. Heikkilä, M. Hel ström: The Role of Digitalisation in Changing the Business Models in Logistics: Case of RoPax Ports 65 Holen, R. M. A., van den Bosch, F. A. J., & Volberda, H. W. (2015). Strategic levers of port authorities for industrial ecosystem development. Maritime Economics and Logistics, 17(1), 79–96. Holmström, J., Holweg, M., Lawson, B., Pil, F. K., & Wagner, S. M. (2019). The digitalization of operations and supply chain management: Theoretical and methodological implications. Journal of Operations Management, 65(8), 728–734. Hsieh, H. F., & Shannon, S. E. (2005). Three approaches to qualitative content analysis. Qualitative Health Research, 15(9), 1277–1288. Inkinen, T., Helminen, R., & Saarikoski, J. (2019). Port digitalization with open data: Challenges, opportunities, and integrations. Journal of Open Innovation: Technology, Market, and Complexity, 5(2). Kamalaldin, A., Sjödin, D., Hul ova, D., & Parida, V. (2021). Configuring ecosystem strategies for digital y enabled process innovation: A framework for equipment suppliers in the process industries. Technovation, 105, 102250. Leminen, S., Rajahonka, M., Wendelin, R., & Westerlund, M. (2020). Industrial internet of things business models in the machine-to-machine context. Industrial Marketing Management, 84, 298–311. Marcadon, J. (1998). L’évolution récente de l’activité maritime et portuaire dans les pays de la Baltique (Recent evolution of the maritime and port activity in the Baltic states). Bul etin de l’Association de Géographes Français, 75(4), 428–436. Mayring, P. (2004). Qualitative content analysis. A Companion to Qualitative Research, 1(2), 159–176. Notteboom, T., Pal is, A., & Rodrigue, J.-P. (2022). Port Economics, Management and Policy. In Port Economics, Management and Policy. Routledge. Paixão Casaca, A. C., & Lyridis, D. v. (2022). Guest Editorial: Port Business and Green Innovation. Maritime Business Review, 7(1), 2–4. Parida, V., Sjödin, D., & Reim, W. (2019). Reviewing literature on digitalization, business model innovation, and sustainable industry: Past achievements and future promises. In Sustainability (Switzerland) (Vol. 11, Issue 2). MDPI. Port of Rotterdam. (2023, Feb 12). Smart shipping process. https://www.portofrotterdam.com/en/to-do-port/futureland/smart-shipping-process Rönty, J., Nokkala, M., & Finnilä, K. (2011). Port ownership and governance models in Finland Development needs & future chal enges. In VTT Technical Research Centre of Finland. Sjödin, D., Parida, V., Jovanovic, M., & Visnjic, I. (2020). Value creation and value capture alignment in business model innovation: A process view on outcome‐based business models. Journal of Product Innovation Management, 37(2), 158–183. Tijan, E., Jović, M., Aksentijević, S., & Pucihar, A. (2021). Digital transformation in the maritime transport sector. Technological Forecasting and Social Change, 170, 120879. Tijan, E., Jović, M., Panjako, A., & Žgaljić, D. (2021). The Role of Port Authority in Port Governance and Port Community System Implementation. Sustainability 2021, Vol. 13, Page 2795, 13(5), 2795. Transport Intelligence. (2019). Global Freight Forwarding 2019. Tsvetkova, A., Gustafsson, M., & Wikström, K. (2021). Digitalizing maritime transport: digital innovation as a catalyzer of sustainable transformation. A Modern Guide to the Digitalization of Infrastructure, 123–148. van der Lugt, L., & de Langen, P. W. (2007). Port authority strategy: beyond the landlord–a conceptual approach. Proceedings of the 2007 IAME Conference. Vendrel -Herrero, F., Myrthianos, V., Parry, G., & Bustinza, O. F. (2017). Digital dark matter within product service systems. Competitiveness Review, 27(1), 62–79. Verhoeven, P. (2010). A review of port authority functions: towards a renaissance? 37(3), 247–270. von Malmborg, F. (2004). Networking for knowledge transfer: Towards an understanding of local authority roles in regional industrial ecosystem management. Vonck, I., Camphuijsen, R., & Jongejan, J. (2021). Europe’s ports at the crossroads of transitions A Deloitte and ESPO Study. 66 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Wahlström, I., Helström, M., Tsvetkova, A., & Chen, Y. (2022). Greening Ferry Ports by Truck Flow Optimisation. Baltic Transport Journal. Watson, R. T., Lind, M., Liesa, F., & Delmeire, N. (2021). Shipping: A Self-Organising Ecosystem. In Maritime Informatics (Issue 2020, pp. 13–32). World Bank. (2007). Port Reform Tool Kit. Yin, R. K. (2018). Case study research and applications: design and methods (Sixth edit). Sage. THE DATA ANALYTIC CAPABILITY WHEEL: AN IMPLEMENTATION FRAMEWORK FOR DIGITALIZATION ALINDA KOKKINOU,1,2 TON VAN KOLLENBURG,2 ALBERT MANDEMAKERS,1 ERIC HOPSTAKEN,1 JAN VAN ELDEREN1 1 Breda University of Applied Sciences, Academy of Built Environment and Logistics, Breda, Netherland kokkinou.A@buas.nl, mandemakers.a@buas.nl, hopstaken.e@buas.nl, elderen2.j@buas.nl 2 AVANS University of Applied Sciences, Center of Expertise Well-Being Economy and New Entrepreneurship, Breda, Netherland ajc.vankollenburg@avans.nl For mature organizations to engage in digital transformation, they first must engage in digitization and digitalization. Digitalization requires the organizations to possess data analytic capability: the ability to transform data into useful insights in a way that creates or maintains competitive advantage. The purpose of this study was to formulate a practical framework for the implementation of digitalization. For this, a qualitative approach was used. Relevant aspects of data analytic capability were identified, based on a review of the literature supplemented with semi-structured interviews with organizations currently implementing digitalization. With these findings a preliminary implementation framework entitled the “Data Analytic Keywords: Capability Wheel” was formulated. The aspects encompassed by digitalization, data, this framework included data quality, data analytics, IT analytic infrastructure, processes, employee knowledge and skills, and capability, management. Future research should refine and validate this change management, framework and examine whether it leads to the successful data, implementation of DAC in organizations. implementation DOI https://doi.org/10.18690/um.fov.4.2023.5 ISBN 978-961-286-751-5 68 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction In an increasing complex and globalized world, digital transformation has been recognized as an important avenue for organizations to create or at least maintain competitive advantage (Hess et al., 2016). Digital transformation has been linked to improved firm performance (Popovič et al., 2018), through increased organizational agility (Ghasemaghaei et al., 2017; Gong & Ribiere, 2023). Digital transformation is defined as ” a process that aims to improve an entity by triggering significant changes to its properties through combinations of information, computing, communication, and connectivity technologies” (Vial, 2019, p. 121), where entity refers to organizations, industries, or more general y, society. Two related phenomena are confounded with digital transformation, namely digitization and digitalization, but for the purpose of this study they will be disentangled following Machado et al.’s (2019). Digitization is defined as “the technical process of converting analog signals into a digital form, and ultimately into binary digits” (Legner et al., 2017, p. 301). Digitalization is defined as “the use of the technologies and data to improve and transform the business processes” while digital transformation is broader, “encompassing changes in the business models, activities, processes, and competences to enable to have al benefits of the ful deployment of the new technologies” (Machado et al., 2019, p. 1114). The scope of the present study is on the process of digitalization. Organizations benefit from digitalization in two ways: (1) by increasing their internal efficiency (e.g. through more efficient product development and/or more efficient manufacturing) and (2) by adding value for customers and other stakeholders (e.g. through more sophisticated products and services) (Björkdahl, 2020). To achieve these outcomes, organizations need to develop their Data Analytic Capability (DAC), that is the ability to turn data into actionable insights by orchestrating data assets, IT infrastructure, and human talent in such a way that they create competitive advantage (Garmaki et al., 2016; Mikalef et al., 2018). A variety maturity models have been developed to support organizations pursuing the development of DAC, digitization and/or digital transformation (Cosic et al., 2012; E. Gökalp & Martinez, 2021; Hein-Pensel et al., 2023; Korsten et al., 2022) from the perspective of diverse disciplines. From the Information Systems (IS) perspective, several models have been developed that stem from the Capability A. Kokkinou, T. van Kol enburg, A. Mandemakers, E. Hopstaken, J. van Elderen: The Data Analytic Capability Wheel: An Implementation Framework for Digitalization 69 Maturity Model (CMM), and include the Business Analytics CMM (BACMM) (Cosic et al., 2012), the Data Science CMM (DSCMM) (M. O. Gökalp et al., 2022), the Advanced Data Analytics CMM (ADACMM) (Korsten et al., 2022) and the Advanced Analytics CMM (Social Security Administration Analytics Center of Excel ence, 2020). Other maturity models developed include those of Comuzzi and Patel (2016), and Grossman (2018). In the discipline of operations management, maturity models have been developed for digitalization and digital transformation (E. Gökalp & Martinez, 2022) and industry 4.0 (Ganzarain & Errasti, 2016; Mittal et al., 2018). These maturity models encompass a variety of aspects required for the organization to develop, including organization, infrastructure, data management, analytics, and governance (Comuzzi & Patel, 2016). These maturity models have one to three complementary purposes, namely describing the current situation, prescribing guidelines for development, and enabling organizations to benchmark their development internal y and external y (Hein-Pensel et al., 2023; Pöppelbuß & Röglinger, 2011). However, a common criticism of maturity models include the lack theoretical and empirical grounding of these models, having mostly been developed by consultants and software vendors (Comuzzi & Patel, 2016; Gupta & George, 2016; Mikalef et al., 2019) and their lack of actionability for organizations due to their descriptive/ comparative nature in combination with their complexity (Barton & Court, 2012; Hein-Pensel et al., 2023). Furthermore, the majority of maturity models examined do not consider al the aspects needed to develop capabilities related to digitalization (Hein-Pensel et al., 2023). Existing research frequently takes an information technology (IT) perspective and focuses on issues of data quality, IT infrastructure and systems (Mikalef et al., 2017). Yet, a successful development of DAC is much more complex, involving factors such as people’s knowledge and skil s, processes, and organizational change (Ghasemaghaei et al., 2018; Mikalef et al., 2018). Studies that have embraced a more holistic view have also recognized the importance of organizational context (Mikalef & Krogstie, 2020). Thus, much remains unexplored about how organizations can develop their DAC, given their organizational context. 70 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY The purpose of this study was to formulate a practical framework for the implementation of digitalization, taking a multi-disciplinary approach and focusing on how organizations can “configure, orchestrate and exploit competencies, assets, and data generated from digital technologies” (Björkdahl, 2020). As many models are perceived to be too complex (Hein-Pensel et al., 2023), the point of departure for this framework was that it needed to be relatable for practitioners. A multi-disciplinary review of the literature was used to identify relevant aspects of DAC and formulate a preliminary framework (section 2). The qualitative research approach to col ecting empirical data is presented in section 3 and the findings thereof in section 4. These led to the refinement of the preliminary framework and the formulation of an implementation framework: the Data Analytic Capability Wheel presented in section 5, followed by the discussion. Limitations and recommendations for future research are detailed in section 7. 2 Review of the Literature Digitalization is a widely used term, both in academic and professional publications, leading to ambiguity (Bloomberg, 2018). For the purpose of this study, digitalization denotes an organizations’ ability to improve its processes through the use of data and/or related technologies. However, this definition only reduces the ambiguity slightly, as digitalization has been defined and operationalized differently across disciplines. The present study therefore reviews literature from several disciplines, including management, manufacturing, supply chain, and business information systems to identify relevant aspects for the implementation of digitalization. 2.1 Digitalization, Data Analytic Capability, and Industry 4.0 In the context of manufacturing, digitalization has led to the concept of Industry 4.0 or Smart Industry (Rosin et al., 2020) and “represents the current trend of automation technologies in the manufacturing industry” (Shahin et al., 2020, p. 2928). From this perspective, digitalization is primarily concerned with improving the efficiency of processes, by improving cost, quality, lead time and flexibility (Khanchanapong et al., 2014). In the context of supply chain management, digitalization has led to the concepts of DAC and Big Data Analytic Capability (BDAC). DAC refers to an organization’s capability to deploy data, technology and people to quickly access and analyze information to support complex decision-making (Yu et al., 2021) with some A. Kokkinou, T. van Kol enburg, A. Mandemakers, E. Hopstaken, J. van Elderen: The Data Analytic Capability Wheel: An Implementation Framework for Digitalization 71 authors scoping this definition to big data (Dubey et al., 2019). While there are merits to distinguishing between the two, technological developments in this area are very rapid. Due to technological developments, what was considered big data and thus chal enging to deal with in the past may no longer be considered chal enging a few years later, thus broadening the applicability of DAC (Kokkinou et al., 2022a). While digitalization, Industry 4.0, and DAC are related concepts, Industry 4.0 focuses more on the application of the technologies associated with digitalization (IoT, robotics etc.) whereas DAC encompasses the pre-requisites thereof (Garmaki et al., 2016; Mikalef et al., 2018). Furthermore, the concept of DAC is more comprehensively defined and is coupled with a sounder theoretical foundation. Therefore, the remainder of this paper wil focus on DAC. 2.2 Theoretical Foundations DAC has strong theoretical foundations in the contingent Resource-Based View of the firm (RBV) (Brandon-Jones et al., 2014; Mikalef & Krogstie, 2020; Wu et al., 2006; Yu et al., 2018). According to RBV, organization’s resources can be a source of sustained competitive advantage if they are valuable, rare, cannot be imitated or substituted (Barney, 1991). However, according to the contingent perspective, the potential of capabilities to lead to competitive advantage wil depend on their alignment with contextual factors such as national context and culture, firm size, strategic context, and other organizational context variables (Aragón-Correa & Sharma, 2003). Thus DAC can only become a source of competitive advantage for an organization if the organization is able to configure, orchestrate and exploit the tangible, intangible, and human aspects necessary in a way that fits its unique context (Björkdahl, 2020; Ghasemaghaei et al., 2018; Mikalef & Krogstie, 2020). 2.3 Data Analytic Capability Development The topic of how DAC should be developed has been investigated from a variety of perspectives, including industry drivers and barriers, organizational enablers, organizational readiness, and organizational maturity (Nayernia et al., 2022). On an organizational level, which is the scope of our study, several ways exist to classify the aspects that make up DAC. Gupta and George (2016) distinguished between tangible, intangible, and human resources, where tangible resources included data, 72 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY technology, basic resources such as time and investment. Intangible resources included a data-driven culture and the intensity of organizational learning, and human resources included managerial skills and technical skills. Mikalef et al. (2017, 2018) further elaborated on this classification. Their review of resources needed to build DAC formed the basis of the list of relevant aspects shown in Table 1. Table 1: Relevant aspects for DAC implementation Access to relevant data: Organizations need to be able to identify, access, and if necessary, acquire relevant data (Behl et al., 2019) Data Quality : Data needs to be complete, accurate, timely, reliable and of value (Mikalef et al., 2017, 2018). Data Governance: Organizations need to put in place procedures to ensure that can create, capture, store, use, retrieve and delete data (Mikalef & Krogstie, ata D 2018; Tal on, 2013) also referred to as Data Management (Jha et al., 2020) Inductive vs. Deductive Approaches: inductive approaches to data can result in insights that are new to the organization but require large investments in data and the ability to analyze it. Conversely, a deductive approach where data are col ected, processed, and visualized for specific purposes can be more effective yet lead to tunnel vision (Günther et al., 2017) Data Analytic Tools are typical y classified in descriptive, predictive and a ics prescriptive tools (Ghasemaghaei et al., 2017) at Tool sophistication: more sophisticated analytical tools (e.g. machine learning and D Analyt artificial intelligence) enable organizations to conduct deeper analysis (Ghasemaghaei et al., 2018) nology IT Infrastructure: Organizations need to have at their disposal an infrastructure that can col ect, analyze, store and share data (Gupta & George, 2016; Mikalef et al., 2018). ech Technical support from vendor: organizations stil rely on technology providers to T support them (Behl et al., 2019) Centralized or Decentralized Structure: Centralization seems to facilitate the ure ses development of DAC by pooling scarce resources whereas decentralization and improved collaboration between domain experts and data scientists (Günther Struct Proces et al., 2017) t Organizational learning refers to the degree to which employees are open to ng extending their knowledge in the face of new emerging technologies. nal rganiza io Training and development of employees is an important mechanism for O Learni organizational learning (Behl et al., 2019; Kokkinou et al., 2021) A. Kokkinou, T. van Kol enburg, A. Mandemakers, E. Hopstaken, J. van Elderen: The Data Analytic Capability Wheel: An Implementation Framework for Digitalization 73 Commitment and Support: managers need to have a long-term orientation to investments and provide resources to data analytic teams (Tabesh et al., 2019) also referred to as Attitude of top management (Behl et al., 2019) ent Effective communication and coordination: managers should encourage cross-functional collaboration, disseminate data-driven insights, and create a anagem common understanding of big data goals (Tabesh et al., 2019) M Gaining managerial analytics acumen: managers need to gain relevant analytics knowledge and help and incentive their staff (Tabesh et al., 2019; Vidgen et al., 2017) Domain knowledge: employees need a deep understanding of the ge procedures, facts, and processes of the organization in order to be able led to solve business problems of interest to the firm (Ghasemaghaei et al., now kils 2018) and S Talent and skil s to analyze and interpret data: Employees need to be able to ployee Km generate business insights from the use of data analytics (Ghasemaghaei E et al., 2018), also referred to as technical skills (Behl et al., 2019) 3 Methodology Consistent with previous research, we used qualitative approach consisting of a combination of interviews with key informants, review of company documents and thematic analysis (Jha et al., 2020). The unit of analysis for our study was organizations’ implementation of digitalization, a complex phenomenon affected by internal and external factors. A qualitative research approach was therefore deemed appropriate as it allowed for an in-depth understanding of the phenomenon with the purpose of analytic generalization (Yin, 2013). We used purposive homogeneous sampling (Gray, 2014) to identify four Dutch organizations in our network that were actively engaged in digitalization projects. We conducted individual and group interviews with key informants from each organization (Jha et al., 2020), supplemented with archival research of internal company documents for one organization (see table 2) and reviewed the websites of al organizations to col ect additional information about relevant contextual variables. The use of semi-structured interview guide based on open question interview protocols al owed important topics to be addressed while giving interviewees the possibility to express their emerging insights and comments. All interviews were 74 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY recorded with participants’ permission, freeing up the researchers to observe and take notes and thus improving their understanding of each interview’s context (Ralston & Blackhurst, 2020). The semi-structured interview guide consisted of an introductory section, a general section about the organization and the respondent. The subsequent sections were about data and IT infrastructure, processes, employee skil s and training, and the role of management. Table 2: Overview of Data Collection Org Type of Data Col ection Methods Duration Organization 1 Production Project Manager (I) 70 min Company (W) 2 Retail (W) Transportation Manager (I) 80 min 3 SME in High Management Team (GI) 150 min Tech Production Project Reports (Docs) (W) 4 Production Process Engineer & 120 company in High Continuous Improvement Specialist (GI) minutes Tech Sector (W) (I: individual interview, GI: group interview, W: website review, Docs: internal document review) The transcripts and notes were analyzed using the software Atlas.ti by applying the steps of thematic analysis (Jha et al., 2020) as recommended by Braun and Clarke (2006), namely familiarization with the data, generation of deductive codes based on the literature review (based on table 1), search for themes, revision of themes and selection of illustrative codes. 4 Findings The purpose of the study was to formulate an implementation framework for digitalization, by exploring relevant aspects for the development of DAC. The numbers in brackets (e.g. [Org#1]) refer to the organizations listed in Table 2. Theme 1: Knowledge and skil s within the organization as a pre-requisite A. Kokkinou, T. van Kol enburg, A. Mandemakers, E. Hopstaken, J. van Elderen: The Data Analytic Capability Wheel: An Implementation Framework for Digitalization 75 The organizations we interviewed were al actively developing their DAC. However, three of the four organizations recognized that there were insufficient knowledge and skills within the organization to do so in a structured way. First, a lack of awareness about digitalization inhibited the urgency for the organization to pursue development in this area. Second, a lack of knowledge of data, data analytics and data management led to difficulties imagining how data could be used to improve decision-making. In the words of a respondent: “we fel behind [in digitalization] in the sense that the need had not been recognized in our department, and the knowledge was not there to dive into it. Since my col eague and I started working on it, the need has been recognized,. Before, no one was busy with the idea that we had to become more data-driven. ” [Org#2] To develop their DAC, the organizations first had to introduce knowledge and expertise about data and data analytics within the organization. For organizations 1 to 3 this happened by employing interns and/or recent graduates with an interest and affinity for data analytics and ensuring management gave them space to experiment. Organizations 1 and 2 also hired consultants to work for a longer period along their employees. Organizations 3 and 4 maintained close contacts with their software vendors who provided some of the knowledge. Theme 2: Role of management to provide leadership, support and resources In all four organizations, management’s knowledge and skills about digitalization were limited. although their interest in the topic was increasing. In organizations 1 and 2, interested and knowledgeable employees receiving the time and resources to demonstrate its added value fueled the desire to increase digitalization. Management was also willing to invest in the IT infrastructure. As an interviewee stated: “management sees that is real y important and are prepared to invest in good systems. Think of a new supply chain application, a centralized department, employees for it, and capacity to manage al of this” [#Org. 1] Theme 3: Evolution of structure and processes Al four organizations seemed to be undergoing extensive developments in terms of structure, responsibilities, and processes related to digitalization. Organizations 1 and 2 saw the development of a centralized department, separate from IT, consolidating knowledge and expertise of DAC. In both cases, this department was 76 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY in the process of taking ownership of the data management processes of the organization and supported departmental employees and teams in their choices of IT infrastructure. In organizations 1, 2 and 4 there was a concerted effort to inventory all the applications currently in use in the organization and formulate a plan to coordinate and manage them. For example, according to organization 4: “our senior management has appointed a task force to review the application landscape within the organization and come up with a comprehensive data management plan”[#Org 4]. In organization 3, the need for a data management plan was increasingly felt by senior management. However, due to a lack of knowledge and skills in this area within the organization, management was encountering difficulties in formulating a plan of action and deciding how to invest in IT infrastructure. In their words: “we keep making smal steps forwards with the best intentions, but we can’t say ‘this is where we are going’ and make big steps.” [#Org3] Theme 4: Missing link to strategy For organizations 3 and 4, the pressure to digitalize was external as it was imposed on them by customers and as a requirement to remain competitive. Despite a lack of skil s and expertise on the topic of DAC, management was very committed to digitalization. For example, both organizations were actively seeking cooperations with universities by participating in student projects, providing internships, and by participating in academic research activities. A striking finding was that none of the organizations involved formulated clear objectives for the implementation of digitalization were formulated beyond the departmental level. This translated to chal enges deciding what data were relevant, and what projects should be prioritized. So while management was supportive and committed, it was not able to communicate in such a way that digitalization efforts were channeled in ways that supported the strategic objectives of the firm. 5 The Data Analytic Capability Wheel as a Metaphor The above findings show that aspects of DAC do not operate independently. Instead, they are intertwined and thus require organizations to address them in a comprehensive manner. The findings of the literature review were combined with the empirical findings to formulate a preliminary implementation framework for A. Kokkinou, T. van Kol enburg, A. Mandemakers, E. Hopstaken, J. van Elderen: The Data Analytic Capability Wheel: An Implementation Framework for Digitalization 77 digitalization, using a bicycle wheel as a metaphor. Bicycle wheels consist of three main parts: the hub, the spokes and the rim. The hub of a wheel is the part at its center that gives the wheel its integrity and al ows it to rotate. The hub also attaches the wheel to the rest of the bicycle. The spokes of a wheel connect the hub to the rim and are meant to support the structure of the wheel. Their invention in 2000 BC was considered a revolution as they made wheels lighter and faster (Frithowulf, 2022). The rim and tire of the wheel make contact with the environment, absorbing shocks, keeping grip on the road, while transferring the wheel’s speed without slipping. Applying the wheel metaphor to DAC creates a practical and relatable framework for the implementation of DAC, as explained below. The hub of the DAC wheel is where an organization’s DAC links to its organizational strategy (the bicycle). By clearly identifying and communicating how the organization’s DAC contributes to its strategy, management can ensure that efforts to develop DAC are coherent. The hub is also the point around which the wheel revolves. This translates to decision-making about which resources to invest in and which projects to prioritize. The link to strategy is essential to ensure legitimacy and coherence to managerial decision-making, answering the “why pursue digitalization?” question. Figure 1: Implementation Framework 78 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY The spokes of the DAC Wheel are the six implementation aspects, namely Data, Data Analytics, Employee Knowledge and Skills, Management, Processes and Structure, and IT Infrastructure,. As the empirical findings show, these aspects are interrelated (see figure 1) and thus need to be considered holistically. Returning to the wheel metaphor, if spokes are of unequal length, the wheel wil not turn properly and will be structurally unsound. Similarly, organizations’ implementation of DAC wil not be smooth or even functional if one aspect receives too much attention at the expense of the other aspects. For example, organizations that spend an outsize budget and effort on their data quality at the expense of knowledge and skil s of employees wil not be successful. The rim and tire of the wheel are where the competitive advantage that DAC confers to the organization becomes apparent, by enabling the organization to sense its environment and respond quickly to changes in an effective way. The rim and tire of the DAC Wheel represent the organizational agility (Ghasemaghaei et al., 2017), defined as its “ability to quickly respond and proactively embrace unanticipated changes in dynamic environments through ef ective resource reconfiguration and rapid decision-making” (Gong & Ribiere, 2023, p. 5). Organizations that exhibit a high level of fit between different aspects such as analytical tools, data, and people wil be able to better use DAC to generate organizational agility (Ghasemaghaei et al., 2017). Just as different types of wheels are appropriate depending on the type and purpose of a bicycle, organizations wil develop DAC in a way consistent with their strategy (e.g. cost leadership, customer engagement) (Sebastian et al., 2020). 5 Discussion Despite the recognized importance of digitalization, there is a lack of understanding of how it should be implemented. Even when focusing on the better developed theoretical construct of DAC, there is a lack of empirical research focusing on implementation aspects. This is possibly due to the fact that academic research on the topic is typical y conducted from the narrow perspective of a single discipline, most often information systems research (Gupta & George, 2016). Our findings show that a comprehensive perspective is necessary and that al three categories of DAC aspects defined by Gupta and George (2016) are relevant when implementing A. Kokkinou, T. van Kol enburg, A. Mandemakers, E. Hopstaken, J. van Elderen: The Data Analytic Capability Wheel: An Implementation Framework for Digitalization 79 DAC. Furthermore, our findings show that DAC aspects are strongly inter-related and thus the development of DAC needs to take a holistic approach. Two aspects of DAC play a more prominent role, namely management and knowledge and skills. First, an initial seed of knowledge and skills within the organization is needed to create awareness and help the organization make the first steps towards developing DAC. Management plays an important role in acquiring this knowledge and expertise for the organization, and turning it into explicit and implicit organizational knowledge, either through hiring new employees, training existing employees by appealing to their intrinsic motivation (Pieters et al., 2022), or creating access to outside expertise (Behl et al., 2019; Kokkinou et al., 2021). Second, consistent with Tabesh et al. (2019), Mikalef et al. (2019) and Vidgen et al. (2017), we found that management plays an important role in orchestrating the necessary aspects of DAC. To be effective, management first need to acquire knowledge and skil s themselves. Management needs to showing commitment and give support by allocating the right resources to the right people (Kokkinou et al., 2023). Furthermore, management needs to communicate the importance of DAC by linking it to the strategic objectives of the firm, a finding that parallels literature on continuous improvement implementation (Kokkinou et al., 2022b). Consistent with IS research (Ghasemaghaei et al., 2017, 2018), we found that data and IT infrastructure received the most attention as organizations focused on collecting data of sufficient quality for data analytics projects. However, organizations were increasingly recognizing that processes and structure played an important role in ensuring that appropriate data were collected and shared across the organizations, leading to changes in the organizational structure and corresponding processes. Our study contributes to the notion that the implementation of digitalization, and specifical y the development of DAC concerns complex socio-technical processes, requiring a multi-disciplinary perspective (Legner et al., 2017; Mikalef & Krogstie, 2020). 80 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 7 Limitations and Further Research While the implementation framework presented in this paper is based on a multi- disciplinary review of the literature, this approach remains inferior to a structured review of the literature. It is therefore recommended to refine this framework through a structured review of the literature that encompasses more disciplines related to the use of data in decision-making. Similarly, four interviews are insufficient to validate the framework. Further research should adopt an action methodology to further test, validate, and refine the DAC Wheel. Finally, the study findings show similarities and overlap with literature on continuous improvement implementation. Further research should examine whether success factors of continuous improvement implementation could also apply to the implementation of digitalization. References Aragón-Correa, J. A., & Sharma, S. (2003). A Contingent Resource-Based View of Proactive Corporate Environmental Strategy. The Academy of Management Review, 28(1), 71. https://doi.org/10.2307/30040690 Barney, J. (1991). Firm Resources and Sustained Competitive Advantage. Journal of Management, 17(1), 99–120. https://doi.org/10.1177/014920639101700108 Barton, D., & Court, D. (2012). Making advanced analytics work for you. Harvard Business Review, 90(10), 78–83. Behl, A., Dutta, P., Lessmann, S., Dwivedi, Y. K., & Kar, S. (2019). A conceptual framework for the adoption of big data analytics by e-commerce startups: A case-based approach. Information Systems and E-Business Management, 17(2–4), 285–318. https://doi.org/10.1007/s10257-019-00452-5 Björkdahl, J. (2020). Strategies for Digitalization in Manufacturing Firms. California Management Review, 62(4), 17–36. https://doi.org/10.1177/0008125620920349 Bloomberg, J. (2018, April 29). Digitization, Digitalization, And Digital Transformation: Confuse Them At Your Peril. Forbes. https://www.forbes.com/sites/jasonbloomberg/ 2018/04/29/digitization-digitalization-and-digital-transformation-confuse-them-at-your-peril/?sh=62d90292f2c7 Brandon-Jones, E., Squire, B., Autry, C. W., & Petersen, K. J. (2014). A Contingent Resource-Based Perspective of Supply Chain Resilience and Robustness. Journal of Supply Chain Management, 50(3), 55–73. https://doi.org/10.1111/jscm.12050 Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa Comuzzi, M., & Patel, A. (2016). How organisations leverage Big Data: A maturity model. Industrial Management & Data Systems, 116(8), 1468–1492. https://doi.org/10.1108/IMDS-12-2015-0495 Cosic, R., Shanks, G., & Maynard, S. (2012). Towards a business analytics capability maturity model. Dubey, R., Gunasekaran, A., & Childe, S. J. (2019). Big data analytics capability in supply chain agility: The moderating effect of organizational flexibility. Management Decision, 57(8), 2092–2112. https://doi.org/10.1108/MD-01-2018-0119 A. Kokkinou, T. van Kol enburg, A. Mandemakers, E. Hopstaken, J. van Elderen: The Data Analytic Capability Wheel: An Implementation Framework for Digitalization 81 Frithowulf, H. (2022, September 4). Timeline of the Wheel: History and Invention. Malevus. https://malevus.com/how-the-wheel-was-invented-and-developed/ Ganzarain, J., & Errasti, N. (2016). Three stage maturity model in SME’s toward industry 4.0. Journal of Industrial Engineering and Management (JIEM), 9(5), 1119–1128. Garmaki, M., Boughzala, I., & Wamba, S. F. (2016). The ef ect of Big Data Analytics Capability on Firm Performance. 301. Ghasemaghaei, M., Ebrahimi, S., & Hassanein, K. (2018). Data analytics competency for improving firm decision making performance. The Journal of Strategic Information Systems, 27(1), 101–113. https://doi.org/10.1016/j.jsis.2017.10.001 Ghasemaghaei, M., Hassanein, K., & Turel, O. (2017). Increasing firm agility through the use of data analytics: The role of fit. Decision Support Systems, 101, 95–105. https://doi.org/10.1016/j.dss.2017.06.004 Gökalp, E., & Martinez, V. (2021). Digital transformation capability maturity model enabling the assessment of industrial manufacturers. Computers in Industry, 132, 103522. https://doi.org/10.1016/j.compind.2021.103522 Gökalp, E., & Martinez, V. (2022). Digital transformation maturity assessment: Development of the digital transformation capability maturity model. International Journal of Production Research, 60(20), 6282–6302. https://doi.org/10.1080/00207543.2021.1991020 Gökalp, M. O., Gökalp, E., Kayabay, K., Koçyiğit, A., & Eren, P. E. (2022). The development of the data science capability maturity model: A survey-based research. Online Information Review, 46(3), 547–567. https://doi.org/10.1108/OIR-10-2020-0469 Gong, C., & Ribiere, V. (2023). Understanding the role of organizational agility in the context of digital transformation: An integrative literature review. VINE Journal of Information and Knowledge Management Systems. https://doi.org/10.1108/VJIKMS-09-2022-0312 Gray, D. E. (2014). Doing research in the real world (3. ed). Sage. Grossman, R. L. (2018). A framework for evaluating the analytic maturity of an organization. International Journal of Information Management, 38(1), 45–51. https://doi.org/10.1016/j.ijinfomgt.2017.08.005 Günther, W. A., Rezazade Mehrizi, M. H., Huysman, M., & Feldberg, F. (2017). Debating big data: A literature review on realizing value from big data. The Journal of Strategic Information Systems, 26(3), 191–209. https://doi.org/10.1016/j.jsis.2017.07.003 Gupta, M., & George, J. F. (2016). Toward the development of a big data analytics capability. Information & Management, 53(8), 1049–1064. https://doi.org/10.1016/j.im.2016.07.004 Hein-Pensel, F., Winkler, H., Brückner, A., Wölke, M., Jabs, I., Mayan, I. J., Kirschenbaum, A., Friedrich, J., & Zinke-Wehlmann, C. (2023). Maturity assessment for Industry 5.0: A review of existing maturity models. Journal of Manufacturing Systems, 66, 200–210. https://doi.org/10.1016/j.jmsy.2022.12.009 Hess, T., Matt, C., Benlian, A., & Wiesböck, F. (2016). Options for formulating a digital transformation strategy. MIS Quarterly Executive, 15(2). Jha, A. K., Agi, M. A. N., & Ngai, E. W. T. (2020). A note on big data analytics capability development in supply chain. Decision Support Systems, 138, 113382. https://doi.org/10.1016/j.dss.2020.113382 Khanchanapong, T., Prajogo, D., Sohal, A. S., Cooper, B. K., Yeung, A. C. L., & Cheng, T. C. E. (2014). The unique and complementary effects of manufacturing technologies and lean practices on manufacturing operational performance. International Journal of Production Economics, 153, 191– 203. https://doi.org/10.1016/j.ijpe.2014.02.021 Kokkinou, A., Mandemakers, A., & Mitas, O. (2022a). Data-analytic capability as a path to a resilient supply chain: An empirical investigation. Presented at the 2022 EurOMA Conference. EUROMA, Berlin, Germany. Kokkinou, A., van Kollenburg, T., & Touw, P. (2022b). The road to operational excel ence is scenic: Tales of continuous improvement journeys. Presented at the 2022 EurOMA Conference. European Operations Management Association (EurOMA), Berlin. 82 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Kokkinou, A., Kollenburg, T. van, & Mandemakers, A. (2023). Development of data analytic capability in organizations. Presented at the 2023 EUROMA Conference. Leuven, Belgium Kokkinou, A., van Kol enburg, T., & Touw, P. (2021). The Role of Training in the Implementation of Lean Six Sigma. Presented at the 2021 EurOMA Conference. European Operations Management Association (EurOMA), Berlin. Korsten, G., Aysolmaz, B., Turetken, O., Edel, D., & Ozkan, B. (2022). ADA-CMM: A Capability Maturity Model for Advanced Data Analytics. Hawai International Conference on System Sciences. https://doi.org/10.24251/HICSS.2022.032 Legner, C., Eymann, T., Hess, T., Matt, C., Böhmann, T., Drews, P., Mädche, A., Urbach, N., & Ahlemann, F. (2017). Digitalization: Opportunity and Chal enge for the Business and Information Systems Engineering Community. Business & Information Systems Engineering, 59(4), 301–308. https://doi.org/10.1007/s12599-017-0484-2 Machado, C. G., Winroth, M., Carlsson, D., Almström, P., Centerholt, V., & Hal in, M. (2019). Industry 4.0 readiness in manufacturing companies: Chal enges and enablers towards increased digitalization. Procedia CIRP, 81, 1113–1118. https://doi.org/10.1016/j.procir.2019.03.262 Mikalef, P., Boura, M., Lekakos, G., & Krogstie, J. (2019). Big data analytics and firm performance: Findings from a mixed-method approach. Journal of Business Research, 98, 261–276. https://doi.org/10.1016/j.jbusres.2019.01.044 Mikalef, P., Framnes, V. A., Danielsen, F., Krogstie, J., & Olsen, D. (2017). Big data analytics capability: Antecedents and business value. 136. Mikalef, P., & Krogstie, J. (2018). Big Data Governance and Dynamic Capabilities: The Moderating ef ect of Environmental Uncertainty. 206. Mikalef, P., & Krogstie, J. (2020). Examining the interplay between big data analytics and contextual factors in driving process innovation capabilities. European Journal of Information Systems, 29(3), 260–287. https://doi.org/10.1080/0960085X.2020.1740618 Mikalef, P., Pappas, I. O., Krogstie, J., & Giannakos, M. (2018). Big data analytics capabilities: A systematic literature review and research agenda. Information Systems and E-Business Management, 16(3), 547–578. https://doi.org/10.1007/s10257-017-0362-y Mittal, S., Khan, M. A., Romero, D., & Wuest, T. (2018). A critical review of smart manufacturing & Industry 4.0 maturity models: Implications for small and medium-sized enterprises (SMEs). Journal of Manufacturing Systems, 49, 194–214. https://doi.org/10.1016/j.jmsy.2018.10.005 Nayernia, H., Bahemia, H., & Papagiannidis, S. (2022). A systematic review of the implementation of industry 4.0 from the organisational perspective. International Journal of Production Research, 60(14), 4365–4396. https://doi.org/10.1080/00207543.2021.2002964 Pieters, J. J., Kokkinou, A., & van Kol enburg, T. (2022). Understanding Blockchain Technology Adoption by Non-experts: An Application of the Unified Theory of Acceptance and Use of Technology (UTAUT). Operations Research Forum, 3(1), 1. https://doi.org/10.1007/s43069-021-00113-9 Popovič, A., Hackney, R., Tassabehji, R., & Castel i, M. (2018). The impact of big data analytics on firms’ high value business performance. Information Systems Frontiers, 20(2), 209–222. https://doi.org/10.1007/s10796-016-9720-4 Pöppelbuß, J., & Röglinger, M. (2011). What makes a useful maturity model? A framework of general design principles for maturity models and its demonstration in business process management. Ralston, P., & Blackhurst, J. (2020). Industry 4.0 and resilience in the supply chain: A driver of capability enhancement or capability loss? International Journal of Production Research, 58(16), 5006–5019. https://doi.org/10.1080/00207543.2020.1736724 Rosin, F., Forget, P., Lamouri, S., & Pellerin, R. (2020). Impacts of Industry 4.0 technologies on Lean principles. International Journal of Production Research, 58(6), 1644–1661. https://doi.org/10.1080/00207543.2019.1672902 Sebastian, I. M., Ross, J. W., Beath, C., Mocker, M., Moloney, K. G., & Fonstad, N. O. (2020). How big old companies navigate digital transformation. In Strategic information management (pp. 133– 150). Routledge. A. Kokkinou, T. van Kol enburg, A. Mandemakers, E. Hopstaken, J. van Elderen: The Data Analytic Capability Wheel: An Implementation Framework for Digitalization 83 Shahin, M., Chen, F. F., Bouzary, H., & Krishnaiyer, K. (2020). Integration of Lean practices and Industry 4.0 technologies: Smart manufacturing for next-generation enterprises. The International Journal of Advanced Manufacturing Technology, 107(5–6), 2927–2936. https://doi.org/10.1007/s00170-020-05124-0 Social Security Administration Analytics Center of Excel ence (2020). https://www.ssa.gov/ data/data_governance_board/ACE_A2CM2_for%20DGB.pptx.pdf Tabesh, P., Mousavidin, E., & Hasani, S. (2019). Implementing big data strategies: A managerial perspective. Business Horizons, 62(3), 347–358. https://doi.org/10.1016/j.bushor.2019.02.001 Tallon, P. P. (2013). Corporate Governance of Big Data: Perspectives on Value, Risk, and Cost. Computer, 46(6), 32–38. https://doi.org/10.1109/MC.2013.155 Vial, G. (2019). Understanding digital transformation: A review and a research agenda. The Journal of Strategic Information Systems, 28(2), 118–144. https://doi.org/10.1016/j.jsis.2019.01.003 Vidgen, R., Shaw, S., & Grant, D. B. (2017). Management chal enges in creating value from business analytics. European Journal of Operational Research, 261(2), 626–639. https://doi.org/10.1016/j.ejor.2017.02.023 Wu, F., Yeniyurt, S., Kim, D., & Cavusgil, S. T. (2006). The impact of information technology on supply chain capabilities and firm performance: A resource-based view. Industrial Marketing Management, 35(4), 493–504. https://doi.org/10.1016/j.indmarman.2005.05.003 Yin, R. K. (2013). Validity and generalization in future case study evaluations. Evaluation, 19(3), 321– 332. https://doi.org/10.1177/1356389013497081 Yu, W., Chavez, R., Jacobs, M. A., & Feng, M. (2018). Data-driven supply chain capabilities and performance: A resource-based view. Transportation Research Part E: Logistics and Transportation Review, 114, 371–385. https://doi.org/10.1016/j.tre.2017.04.002 Yu, W., Wong, C. Y., Chavez, R., & Jacobs, M. A. (2021). Integrating big data analytics into supply chain finance: The roles of information processing and data-driven culture. International Journal of Production Economics, 236, 108135. https://doi.org/10.1016/j.ijpe.2021.108135 84 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY UNLOCKING THE POTENTIAL OF DATA- DRIVEN BUSINESS MODELS: AN EMPIRICAL INVESTIGATION INTO THE ROLE OF ECOSYSTEMS AND FAIR DATA USE MARIKKA HEIKKILÄ, FARHAM AHMAD, JUKKA HEIKKILÄ University of Turku, Turku School of Economics, Finland marikka.heikkila@utu.fi, farhan.ahmad@utu.fi, jups@utu.fi Data-driven business models are expected to stimulate new economic growth by promoting innovation and value creation through data. However, in addition to concerns about privacy and security, there are ongoing discussions about fair data usage at both the EU and global levels. This research analyses how business model change is influenced by the expected economic benefits as wel as how involvement in data ecosystems and the Keywords: adoption of fair data practices can encourage data-driven fair innovation. We develop and test a structural equation model with data a sample of 1,200 European companies. The findings suggest practices, data-driven that organizations recognize the potential for new business and innovation, innovation opportunities with data-driven business models. business Nevertheless, it is essential to engage in data ecosystems and model, structural implement fair and sustainable data usage practices in order to equation realize these benefits. model DOI https://doi.org/10.18690/um.fov.4.2023.6 ISBN 978-961-286-751-5 86 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Data has become an essential, multi-faceted asset of the modern global economy (Brynjolfsson et al., 2011; Jetzek et al., 2014; Saaristo & Heikkilä, 2022). New business models (BMs) for novel products and services are expected to be developed through creating, sharing, and using data. But regulations are also being put in place to safeguard against the misuse of information and potential threats to privacy and business secrets (e.g., Data Act, 2022; Krämer & Schnurr, 2021). Too strict protection may enforce monopolistic market positions, hinder innovation, increase the costs of heightened surveil ance, and limit freedom of expression in the worst case (Gawer & Srnicek, 2021; Zuboff, 2019; Gandhi et al., 2011). These are pertinent issues for individuals, businesses, and governments who are constantly seeking a fair balance between restrictions and freely flowing open data. Despite the differences in viewpoints and associated risks, the data economy is projected to drive future BMs and innovations, offering ample opportunities for al stakeholders, as noted by the European Commission in 2020. While the literature suggests that firms can leverage digitalization and their big data analytics capabilities to innovate their BMs (Bouwman et al., 2019; Ciampi et al., 2021), the experiences and outcomes of seizing these opportunities vary. Several studies have witnessed challenges in innovating from data due to poor data quality, unclear ownership, or usage rights, among other reasons (Ermakova et al., 2021; Lange et al., 2021; Rantanen et al., 2019; Eriksson & Heikkilä, 2023). Therefore, it can be concluded that experimenting with data does not automatical y translate into innovation. Recent literature, particularly in relation to Industry 4.0, suggests that to overcome these chal enges, it is advisable to focus on creating data ecosystems that adopt fair and sustainable data usage practices (Azkan et al., 2022; Hubaux & Juels, 2016). Through such data ecosystems, companies can gain access to additional data sources, increase their opportunities for collaboration, and stay up-to-date on the latest trends and technologies in the industry (Oliveira et al., 2019). Moreover, to ensure that data is processed and used ethical y and lawful y, practices of fair and sustainable data usage are becoming increasingly necessary (Bennett, 2019). Few empirical studies have investigated the impact of fair and sustainable data use practices and data ecosystem participation on BMs building on data-driven M. Heikkilä, F. Ahmad, J. Heikkilä: The Data Analytic Capability Wheel: Unlocking the Potential of Data-Driven Business Models: An Empirical Investigation into the Role of Ecosystems and Fair Data 87 Use innovation. We fill this gap by developing a model to analyse if expected business benefits from the data economy lead to BM changes, and whether participation in data ecosystems and adoption of fair data practices can increase data-driven innovation. The empirical data are collected from 1,200 diverse European companies. This study contributes to the literature on BMs, especially on data-driven BM, where data is a core ingredient of a company's BM (Trabucci & Buganza, 2019). Our objective is to deepen the knowledge on how changes in BMs eventually lead to increased data-driven innovation. Data-driven innovation is a representation of how BMs are implemented to enhance the overal performance of the ecosystem and individual companies within it (e.g., Jetzek et al., 2014). The rest of this paper includes a review of the literature, research hypotheses, methodology, findings, discussion, and conclusions. 2 Literature review and hypothesis Several companies have realized the potential advantages of creating new BMs that utilize data as a crucial resource (Hartman et al., 2014). Ciampi et al. (2021) assert that a company's competence in big data analytics positively affects its ability to innovate its BM. Bouwman et al. (2019) revealed that companies that actively utilize social media, big data, and information technology can increase their performance by transforming their BMs and strengthening their capacity for innovation. A BM describes how a company creates, delivers, and captures value (Teece, 2010). Companies respond to continuously evolving environments by modifying their BMs (Marolt et al., 2018; Pucihar et al., 2019). These BM changes can range from modest refinements to some BM elements to a complete overhaul of the entire BM (Saebi et al., 2017; Eriksson et al., 2022). Companies often make these changes to achieve strategic goals such as increasing profitability, expanding their business, or entering new markets (Heikkilä et al., 2018). Thus, On the basis of research literature, we could assume if a company sees potential benefits from data, it is motivated to change its BM (Bucherer et al., 2012; Hartmann et al., 2013; Lindgardt et al., 2009; Pohle and Chapman, 2006; Lafiti et al., 2021a,b). However, to date, there is no clear 88 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY empirical evidences showing this relationship. Thus, we propose the following hypothesis (see also Fig.1): H1. Potential benefits from data has a significant positive ef ect on BM change Generating economic value via data-driven innovation is seldom achievable by a lone organization. Rather, this process is said to necessitate involvement in data ecosystems (Hein et al., 2019), which refer to networks of entities that interact to exchange, create, and utilize data (Oliveira and Lóscio, 2018). To unlock the potential of data-driven business, companies need to share data and collaborate with other parties such as companies, government agencies, users, and other stakeholders. By doing so, they gain access to a broader range of data, which can be used to create new products and services for markets. Furthermore, collaborating with other companies enables them to leverage complementary skil s and expertise, accelerating innovation. Therefore, it is not surprising that a recent survey found that approximately one third of organizations are collaborating with partners to exchange data (MIT, 2021). There are several types of data ecosystems, which differ on their control, openness, participant interdependence and purpose (Gelhaar et al., 2021; Curry & Ojo, 2020). For instance, the data ecosystem may be centered around a keystone actor, or it can be a platform or marketplace (Gelhaar et al., 2021). They may also be collaboratively developed and influenced by various actors (De Reuver et al., 2018), such as European data ecosystem is International Data Spaces (Otto & Jarke, 2019), but a data ecosystem could also be formed between just a few companies. We expect that changes in BM would increase a company's activity in data ecosystems: H2. BM change has a significant positive ef ect on participation to data ecosystems Companies face various chal enges when leveraging data-driven opportunities, such as ensuring secure data sharing, complying with privacy regulations, and controlling personal and private data (MIT, 2021). Additionally, companies must respect copyright, intel ectual property, and non-disclosure requirements. To address concerns about potential data misuse, companies can combine their new BMs with practices that promote fair and sustainable data usage. These practices ensure that the company creates services and data-based products in an ethical, fair manner. In M. Heikkilä, F. Ahmad, J. Heikkilä: The Data Analytic Capability Wheel: Unlocking the Potential of Data-Driven Business Models: An Empirical Investigation into the Role of Ecosystems and Fair Data 89 Use a data economy, fairness requires protecting individuals' rights, respecting businesses' rights and contracts, and taking into account the needs of al stakeholders (Parikka et al., 2021). Practices that promote fair and sustainable data usage create trust, which is often a prerequisite for accessing customer data and building long-term relationships. Researchers note that simply complying with legal requirements is insufficient, and explicit approaches and processes must be adopted to promote fair data usage (Vermanen et al., 2022). Furthermore, we anticipate that participating in data ecosystems could enhance the adoption of fair and sustainable data usage practices (Koskinen et al., 2019)." H3. BM changes are positively related to fair & sustainable data usage practices. H4. Participation in data ecosystems has a significant positive ef ect on fair & sustainable data usage practices. Data-driven innovation refers to innovation that utilizes data as a core ingredient. The literature describes it as business innovation that is based on exploiting data and is capable of generating positive economic and social impacts (Jetzek et al., 2014). Companies utilize data to inform decision-making, improve organizational processes, or create customer value (Brynjolfsson et al., 2011). Empirical studies evidence how, for example, by col ecting and analyzing data from users (Trabuchi & Buganza, 2019) or other sources (Jetzek et al., 2014), a company can gain insights into how its customers value its products and services and how the company could add even more value to the market. We aim to analyze whether changes to the BMl increase data-driven innovation directly or whether participation in data ecosystems or practicing fair and sustainable use of data leads to increased data-driven innovation: H5. BM changes have a significant positive ef ect on data-driven innovation H6. Participation in data ecosystems has a significant positive ef ect on data-driven innovation H7. Fair & sustainable data usage practices have a significant positive ef ect on data-driven innovation 90 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Figure 1: Research model 3 Methodology 3.1 The measures Researchers have developed various measurement scales for assessing changes or innovations in BMs. Some scales focus on the scope of BM change, which evaluates architectural and modular changes in the BM (Spieth & Schneider, 2016; Foss & Saebi, 2017; Lafiti et al., 2021a, b; Eurostat, 2022; Bashir et al., 2023). Novelty scales, in turn, measure the degree of change in an enterprise or industry (Clauss, 2017; Marolt et al., 2018; Pucihar et al., 2019). Also, there are scales that measure disruptiveness of changes (Karimi & Walter, 2016), the novelty of digital innovation (Soluk et al., 2021), or the sustainability of changes (Bashir et al., 2022). Many researchers link changes in a BM to changes in its components. The widely adopted BM Canvas (Osterwalder et al., 2005; Osterwalder & Pigneur, 2010) depicts customer relationships and segmentation, channels, value proposition, key activities, key resources, partners, and the revenue model as components of BM. Both researchers and practitioners use the BM Canvas to analyse BMs and changes in BMs (Lafiti et al., 2021b). Even Eurostat (2022) introduced a scale for BM change in the community innovation survey (CIS) for the first time in 2022. Therefore, this study measures BM changes (i.e., the scope of BM innovation) using a subset of M. Heikkilä, F. Ahmad, J. Heikkilä: The Data Analytic Capability Wheel: Unlocking the Potential of Data-Driven Business Models: An Empirical Investigation into the Role of Ecosystems and Fair Data 91 Use seven BM components that follow the structure of the BM Canvas (Osterwalder & Pigneur, 2010). In this study, potential benefits from data-driven innovations were measured by three items (Heikkilä et al., 2018): First, the respondents were to consider how much potential they see in data economy to increase revenue from the current BM; secondly, whether they can find new revenue streams from innovations; or, thirdly, to save costs. Participation in data ecosystems was inquired with questions if the company is a partner or a facilitator of data ecosystem(s). Fair and sustainable data usage practices consists of a set of eight questions measuring e.g. ethical code of conduct; transparency; privacy; data sovereignty (Parikka et al., 2021; Vermanen et al., 2022). Data-driven innovation was measured with two questions on continuous improvement and innovation of products and services (e.g. Boer & Gertsen, 2003), and with questions on the use of data to enhance customer experience, and creating value for society, people, and environment (Sitra, 2021). The constructs and questions are provided in Appendix. 3.2 Survey administration, sample and data col ection The survey was conducted in four countries - Finland, France, Germany, and the Netherlands - in 2021, and was commissioned by the Finnish Innovation Fund Sitra1. The research data was col ected from a B2B decision-maker panel, consisting of professionals holding key decision-making roles, such as CxOs of data, digitalization, information systems, strategy, marketing, and business development. Table 1 provides background characteristics of the sample, with equal strata for the countries, and with close to equal number of respondents from large, medium, small, and micro-sized companies (excluding sole proprietors) in each country. No quotas were defined on industry or activity in the data economy. The raw data set is openly available from Sitra (2021). Our interest in the potential of data economy to change BMs seems reasonable: More than 80% of the sampled companies foresaw the potential of data economy in creating a competitive edge (Saaristo & Heikkilä, 2022; Ulander et al. 2021). 1 Sitra is a fund, which has a special national level role in Finland, because it is accountable and reports directly to the Finnish Parliament. Its capital was granted by the Finnish Parliament. 92 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 1: Profile of respondents’ companies Variable Category % Country Finland 25 France 25 Germany 25 The Netherlands 25 Firm size (turnover €) Micro (under 2 million) 28.1 Small (2 – 10 million) 26.6 Medium (10 – 50 million) 22.1 Large (over 50 million) 23.2 Industry sector Service activities 15.97 Manufacturing 11.34 Information and communication 9.83 Human health and social work 8.07 Financial and insurance activities 7.98 Wholesale and retail trade 6.64 Other 40.17 The survey data contained some missing values, which were less than 5 % of the total values in the dataset. Therefore, the mean imputation method was used to deal with missing values (Hair et al., 2021). Cross sectional surveys measuring both dependent and independent variables simultaneously are prone to common method variance (Kock et al., 2021), so in our study we use Harman’s single factor test to assess it. The principal axis factoring based first factor accounted for the 35.73 percent of the overal variance. The small size (i.e. below 70%, Kock et al., 2021) of single factor accounted variance confirms the absence of a common method variance problem in our dataset. 4 Data analysis and findings Partial least squares structural equation modeling (PLS-SEM) was used to test the hypotheses. The guidelines provided by Hair et al. (2021) were followed to test the hypotheses. We used SmartPLS 4.0 software for data analysis. M. Heikkilä, F. Ahmad, J. Heikkilä: The Data Analytic Capability Wheel: Unlocking the Potential of Data-Driven Business Models: An Empirical Investigation into the Role of Ecosystems and Fair Data 93 Use 4.1 Reliability and validity We tested first model’s reliability (indicator reliability, Cronbach’s alpha and composite reliability) and then convergent validity in Table 2 and discriminant validity in Table 3. The factor loading for an item is recommended to be at least 0.60 (Hair et al., 2021). As shown in Table 1, all items have a higher value than the recommended threshold. Moreover, both the values of Cronbach’s alpha and composite reliability are above the recommended threshold value of minimum 0.70. Thus, internal consistency and reliability are confirmed. Convergent validity was examined by computing average variance extracted (AVE). The value of at least 0.50 is suggested as a threshold (Hair et al., 2021). As Table 2 shows, all constructs passed the threshold. For discriminant validity, the Fornell and Larcker criterion was used, which requires the square root of the AVE of each construct to be higher than its correlation with other constructs (Wong, 2013). Table 3 shows the fulfilment of this criterion and thus the establishment of discriminant validity. Last, HTMT is used to compare the correlations between indicators measuring different constructs (heterotrait correlations) to the correlations between indicators measuring the same construct (monotrait correlations). If the heterotrait correlations are significantly higher than the monotrait correlations, it suggests that the measures are not distinct and that there may be issues with discriminant validity. HTMT values closer to 0 indicate better discriminant validity and 0.85 is often used as a conservative cut-off point, while a value of 0.90 is considered more liberal (Henseler et al., 2015). The HTMT values in Table 4 were all lower than 0.73, indicating that there were no issues with discriminant validity. 94 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 2: Measurement statistics of first-order constructs Construct Indicator loadings Cronbach's Composite Average alpha reliability variance extracted (AVE) Potential Benefit from data 0.86 0.92 0.79 Item 1 0.90 Item 2 0.91 Item 3 0.85 BM change 0.89 0.91 0.60 Item 1 0.74 Item 2 0.75 Item 3 0.80 Item 4 0.79 Item 5 0.76 Item 6 0.78 Item 7 0.81 Data ecosystem participation 0.90 0.94 0.83 Item 1 0.90 Item 2 0.91 Item 3 0.92 Fair & sustainable data usage practices 0.90 0.92 0.58 Item 1 0.78 Item 2 0.82 Item 3 0.74 Item 4 0.75 Item 5 0.72 Item 6 0.73 Item 7 0.78 Item 8 0.73 Data-driven innovation 0.74 0.83 0.55 Item 1 0.71 Item 2 0.76 Item 3 0.71 Item 4 0.77 M. Heikkilä, F. Ahmad, J. Heikkilä: The Data Analytic Capability Wheel: Unlocking the Potential of Data-Driven Business Models: An Empirical Investigation into the Role of Ecosystems and Fair Data 95 Use Table 3: Discriminant validity of the constructs - Fornel -Larcker Criterion (a) (b) (c) (d) (e) Potential benefit from data (a) 0.887 BM change (b) 0.387 0.775 Data ecosystem participation (c) 0.611 0.428 0.910 Fair & sustainable data usage practices (d) 0.614 0.326 0.583 0.758 Data-driven innovation (e) 0.594 0.349 0.515 0.651 0.739 Bold numbers represent the square roots of the AVEs Table 4: Heterotrait–monotrait ratio of correlations (HTMT) (a) (b) (c) (d) Potential benefit from data (a) BM change (b) 0.440 Data ecosystem participation (c) 0.694 0.479 Fair & sustainable data usage practices (d) 0.681 0.352 0.626 Data-driven innovation (e) 0.693 0.394 0.557 0.729 4.2 Structural model findings Next, relationships between constructs were analysed using path coefficients and significance levels. We examined the effect size using Cohen's f2, which reflects the effect of an exogenous variable on an endogenous variable (Cohen, 1988). f2 values greater than 0.02, 0.15, and 0.35 indicate small, moderate, and large effect sizes, respectively. However, as noted in previous studies, finding large effect sizes is rather uncommon in management research (Mazen et al., 1987, p. 406; Strauch, 2019). Results presented in Figure 2 confirm that al the hypotheses tested in this study are statistically significant. Firstly, we found that potential benefits from data have a clear positive correlation with BM change (H1: β = 0.39, f2 = 0.18, p < 0.001). Secondly, BM change was found to significantly increase participation in data ecosystems (H2: β=0.43, f2 = 0.23, p < 0.001), which, in turn, is strongly and positively correlated with fair and sustainable data usage practices (H4: β=0.54, f2 = 0.37, p < 0.001). Thirdly, business model change is found to enhance data-driven innovation, although the effect is small (H5: β=0.11, f2 = 0.02, p < 0.001). Finally, we observed that while data ecosystem participation has a small effect, fair and sustainable data 96 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY usage practices have a large positive impact on data-driven innovation (H6: β=0.17, f2 = 0.03, p < 0.001; H7: β=0.52, f2 = 0.33, p < 0.001). Figure 2: PLS-SEM analysis results. Notes: *p < 0.001 (two-sided test) 5 Discussion 5.1 Theoretical contributions This article is among the first to present a comprehensive perspective on data-driven innovation in business, arguing that for companies to benefit from the data, they must first modify their BMs. Empirical evidence presented in our study corroborates prior research (MIT, 2021; Azkan et al., 2022; Hubaux & Juels, 2016; Bennett, 2019), emphasizing the critical role of BM change in driving data-driven innovation. Moreover, our study found that the more European companies anticipate financial benefits from usage of data, the more likely they are to change their BMs. Specifical y, companies expect to gain by introducing new revenue streams from innovative products or services, enhancing their revenue from existing business, or reducing costs. Thus, our empirical results confirm previous literature (Heikkilä et al., 2018) that companies' expected benefits positively relate to BM change, here in the context of the data economy. M. Heikkilä, F. Ahmad, J. Heikkilä: The Data Analytic Capability Wheel: Unlocking the Potential of Data-Driven Business Models: An Empirical Investigation into the Role of Ecosystems and Fair Data 97 Use However, making changes to BMs alone is not enough. This study underscores the importance of fair and sustainable data usage practices, as well as active engagement in data ecosystems, as key enablers of data-driven innovation in European companies. This highlights the urgent need for more research and development of IS methods, tools, and data usage that are based on principles of fair and sustainable data practices, as previously proposed in studies such as Nahr & Heikkilä (2022), Alt & Klein (2011), and Pucihar (2020). This need is becoming increasingly urgent in the context of artificial intelligence and machine learning, as noted by Seppälä et al. (2021). 5.2 Managerial implications The findings suggest that the companies which aim at seizing the benefits from data economy by data-driven innovations should, 1) change their BM, 2) be prepared to take part in data ecosystems and 3) take measures to ensure the fair and sustainable data use. Appropriate BM, a suitable network of partners, and ethically sound practices for handling data make it possible to engage and increase innovation from data. BM tools - such as https://businessmakeover.eu/ and data-driven services card game (Breitfuss at al., 2023) - are available to define the changes in BM. Regarding data ecosystems, there are several local, regional, national and European industry specific innovation ecosystems and data spaces that welcome companies to join and co-create data driven new services. These ecosystems provide opportunities not only to share experiences from the data economy but also to share open data among partners or provide reference architectures and interfaces to data assets, as suggested by the Digital Single Market regulatory package within the limits of privacy and intel ectual property protection. In the future, sandboxed development of datadriven BMs and services wil be required for conformity assessment and accreditation of the ethical and rightful treatment of data. Therefore, data ecosystems are increasingly necessary in promoting practices for fair and sustainable data usage. 98 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 6 Conclusions This study provides evidence that the perceived economic potential from data economy serves as a motivation for a company to change its BM, which, in turn, leads to data-driven innovation through increased participation in data ecosystems and the implementation of fair and sustainable data practices. Overall, our study suggests that the data economy can provide a competitive advantage to companies, but it must be implemented in an ethical and responsible manner. These findings have important implications for businesses seeking to leverage data-driven innovation to achieve business benefits. Specifical y, companies must engage in data ecosystems and adopt responsible data usage practices to comply with data protection regulations and maintain ethical use of data. This study has some limitations that need to be taken into consideration when interpreting the results. Firstly, the data was col ected from four European countries only, and therefore, the results may not be generalizable to companies in other regions, such as the United States and China, which have different regulatory environments and attitudes towards fair and sustainable data practices. Further research is required to explore how multinational companies can effectively promote ethical data usage across diverse regulatory environments. Also, it is crucial to examine the validity of ethical data practices and European data sharing ecosystems outside the scope of European legislation. Secondly, the cross-sectional nature of the survey could introduce response bias, as firms at different stages of their innovation trajectories may hold varying views on data-driven innovation. However, the inclusion of a large sample size is expected to mitigate such variance. It is important to note that this study analysed companies of all sizes. Nevertheless, it is worth mentioning that smal and medium-sized enterprises (SMEs) often lag behind in digitalization and digital transformation efforts, resulting in missed opportunities for innovation. Therefore, we recommend that future research places particular emphasis on exploring data-driven innovation and business models within SMEs. Thirdly, future research could examine in more detail how participation in data ecosystems enables the development of data-driven services by companies. Although ecosystems play a significant role in generating collective intelligence and supporting the introduction of new business innovations, the priorities of such ecosystems vary (Elia et al., 2020). Therefore, it would be interesting to do qualitative M. Heikkilä, F. Ahmad, J. Heikkilä: The Data Analytic Capability Wheel: Unlocking the Potential of Data-Driven Business Models: An Empirical Investigation into the Role of Ecosystems and Fair Data 99 Use studies on the reasons behind a company's participation in an ecosystem and how they can leverage col ective intel igence to their benefit. Last, large data-driven organisations may utilize data governance for al ocating authority and control over data and making data-related decisions (Janssen et al., 2020). This points to need for further research of data governance in the context of data-driven business, and its impact on strategies for data ecosystems and ethical data practices. Acknowledgements We express our gratitude to Sitra for sharing their data, which was invaluable to our research. Their commitment to open data has greatly contributed to advancing knowledge in this area and has al owed us to draw meaningful insights from the data. References Alt, R., & Klein, S. (2011). Twenty years of electronic markets research – Looking backwards towards the future. Electronic Markets, 21(1), 41–51. https://doi.org/10.1007/s12525-011-0057-z Azkan, C., Möl er, F., Ebel, M., Iqbal, T., Otto, B., & Poeppelbuss, J. (2022). Hunting the Treasure: Modeling Data Ecosystem Value Co-Creation, Forty-Third International Conference on Information Systems, Copenhagen 2022. Bashir, M., Alfalih, A., & Pradhan, S. (2023). Managerial ties, business model innovation & SME performance: Moderating role of environmental turbulence. Journal of Innovation & Knowledge, 8(1), 100329. Bashir, M., Alfalih, A., & Pradhan, S. (2022). Sustainable business model innovation: Scale development, validation and proof of performance. Journal of Innovation & Knowledge, 7(4), 100243. Bennett, S. (2019). Governance in practice: Big data, privacy and information governance-incorporating an ethical-based assessment. Governance Directions, 71(5), 244-254. Boer, H., & Gertsen, F. (2003). From continuous improvement to continuous innovation: a (retro)(per)spective. International Journal of Technology Management, 26(8), 805-827. Bouwman, H., Nikou, S., & de Reuver, M. (2019). Digitalization, business models, and SMEs: How do business model innovation practices improve performance of digitalizing SMEs?. Telecommunications Policy, 43(9), 101828. Breitfuss, G., Santa-Maria, T., Fruhwirth, M., & Disch, L. (2023). Use Your Data: Design and Evaluation of a Card-Based Ideation Tool for Data-Driven Services. HICSS 2023, Retrieved from: https://scholarspace.manoa.hawaii.edu/items/b3df1410-42f5-4136-ac28-95f934a1a67c Brynjolfsson, E., Hitt, L. M., & Kim, H. H. (2011). Strength in numbers: How does data-driven decision making affect firm performance? SSRN Electronic Journal, doi:10.2139/ssrn.1819486 Bucherer, E., Eisert, U. & Gassmann, O. (2012). Towards systematic business model innovation: Lessons from product innovation management. Creativity and Innovation Management, 21(2), pp. 183–198. Ciampi, F., Demi, S., Magrini, A., Marzi, G., & Papa, A. (2021). Exploring the impact of big data analytics capabilities on business model innovation: The mediating role of entrepreneurial orientation. Journal of Business Research, 123, 1-13. 100 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Clauss, T. (2017). Measuring business model innovation: Conceptualization, scale development, and proof of performance. R&D Management, 47(3), 385-403. Cohen J. E. (1988). Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Curry, E. & Ojo, A. (2020). Enabling Knowledge Flows in an Intel igent Systems Data Ecosystem, in Real-time Linked Dataspaces: Enabling Data Ecosystems for Intel igent Systems, E. Curry (Ed). Springer: Cham, Switzerland. Data Act (2022). Data Act: Proposal for a Regulation on harmonised rules on fair access to and use of data. Retrieved from: https://digital-strategy.ec.europa.eu/en/library/data-act-proposal- regulation-harmonised-rules-fair-access-and-use-data De Reuver, M., Sørensen, C., & Basole, R. C. (2018). The digital platform: a research agenda. Journal of Information Technology, 33(2), 124–135. https://doi.org/10.1057/s41265-016-0033-3 Elia, G., Margherita, A., & Passiante, G. (2020). Digital entrepreneurship ecosystem: How digital technologies and col ective intelligence are reshaping the entrepreneurial process. Technological Forecasting and Social Change, 150, 119791. Eriksson, T. & Heikkilä, M. (2023). Capabilities for data-driven innovation in B2B industrial companies, Industrial Marketing Management, 111, 158-172. https://doi.org/10.1016/j.indmarman.2023.04.005 Eriksson, T., Heikkilä, M. & Nummela, N. (2022). Crafting SME Business Model for International Expansion with Data-Driven Services, in Business Models and Firm Internationalisation, Christian Nielsen, Svetla T. Marinova, Marin A. Marinov (eds.) in the series of Routledge Frontiers in the Development of International Business, Management and Marketing Ermakova, T., Blume, J., Fabian, B., Fomenko, E., Berlin, M., & Hauswirth, M. (2021). Beyond the hype: why do data-driven projects fail?. Proceedings of the 54th Hawaii International Conference on System Sciences. European Commission. (2020). The European Data Strategy. Brussels: European Commission. Retrieved from https://ec.europa.eu/commission/presscorner/detail/en/fs_20_283 Eurostat (2022). Community Innovation Survey (CIS): Questionnaire library, QUESTION - Business model changes, Retrieved from https://ec.europa.eu/eurostat/cache/website/cis/Questions/Question_BUSSMOD.html Foss, N. J., & Saebi, T. (2017). Fifteen years of research on business model innovation: How far have we come, and where should we go?. Journal of management, 43(1), 200-227. Gandhi, R., Sharma, A., Mahoney, W., Sousan, W., Zhu, Q., & Laplante, P. (2011). Dimensions of cyber-attacks: Cultural, social, economic, and political. IEEE Technology and Society Magazine, 30(1), 28-38. Gawer, A. & Srnicek, N. (2021). Online platforms: economic and societal effects, EPRS European Parliamentary Research Service: Brussels, Belgium Retrieved from: https://www.europarl.europa.eu/RegData/etudes/STUD/2021/656336/EPRS_STU(2021) 656336_EN.pdf Gelhaar, J., Groß, T., & Otto, B. (2021). A taxonomy for data ecosystems. Proceedings of the 54th Hawaii International Conference on System Sciences Hair Jr, J. F., Hult, G. T. M., Ringle, C. M., & Sarstedt, M. (2021). A primer on partial least squares structural equation modeling (PLS-SEM). Sage publications. Hartmann, M., Oriani, R. & Bateman, H., (2013). The performance effect of business model innovation: An empirical analysis of pension funds. Proceedings of the 35th DRUID Celebration Conference, pp. 17-19. Hartmann, P. M., Zaki, M., Feldmann, N., & Neely, A. (2014). Big data for big business? A taxonomy of data-driven business models used by start-up firms. Cambridge service al iance, 1-29. Heikkilä, M., Bouwman, H., & Heikkilä, J., (2018). From strategic goals to business model innovation paths: an exploratory study. Journal of Smal Business and Enterprise Development, 25 (1), 107–128. https://doi.org/10.1108/JSBED-03-2017-0097. M. Heikkilä, F. Ahmad, J. Heikkilä: The Data Analytic Capability Wheel: Unlocking the Potential of Data-Driven Business Models: An Empirical Investigation into the Role of Ecosystems and Fair Data 101 Use Hein, A., Weking, J., Schreieck, M., Wiesche, M., Böhm, M., & Krcmar, H. (2019). Value co-creation practices in business-to-business platform ecosystems. Electronic Markets, 29, 503-518. Henseler, J., Ringle, C.M., Sarstedt, M., 2015. A new criterion for assessing discriminant validity in variance-based structural equation modelling. Journal of the Academy of Marketing Science, 43, 115– 135. https://doi.org/10.1007/s11747-014-0403-8 Hubaux, J. P., & Juels, A. (2016). Privacy is dead, long live privacy. Communications of the ACM, 59(6), 39-41. Janssen, M., Brous, P., Estevez, E., Barbosa, L. S., & Janowski, T. (2020). Data governance: Organizing data for trustworthy Artificial Intel igence. Government Information Quarterly, 37(3), 101493. Jetzek, T., Avital, M., & Bjorn-Andersen, N. (2014). Data-driven innovation through open government data. Journal of Theoretical and Applied Electronic Commerce Research, 9(2), 15-16. doi:10.4067/S0718-18762014000200008 Karimi, J., & Walter, Z. (2016). Corporate entrepreneurship, disruptive business model innovation adoption, and its performance: The case of the newspaper industry. Long range planning, 49(3), 342-360. Kock, F., Berbekova, A., & Assaf, A. G. (2021). Understanding and managing the threat of common method bias: Detection, prevention and control. Tourism Management, 86, 104330. Koskinen, J., Knaapi-Junnila, S. & Rantanen, M.M. (2019). What if we Had Fair, People-Centred Data Economy Ecosystems?, 2019 IEEE SmartWorld, Ubiquitous Intelligence & Computing, Advanced & Trusted Computing, Scalable Computing & Communications, Cloud & Big Data Computing, Internet of People and Smart City Innovation (SmartWorld/SCALCOM/UIC/ATC/CBDCom/IOP/SCI), Leicester, UK, 2019, pp. 329- 334, doi: 10.1109/SmartWorld-UIC-ATC-SCALCOM-IOP-SCI.2019.00100. Krämer, J. & Schnurr, D. (2021). Big data and digital markets contestability: Theory of harm and data access remedies, Journal of Competition Law & Economics, 18(2) 255–322, https://doi.org/10.1093/joclec/nhab015 Latifi, M. A., Bouwman, H., & Nikou, S. (2021a). ICT-Driven Business Model Innovation in SMEs: The Role of Organizational Capabilities, Firm Size and Age. BLED 2021 Proceedings Latifi, M. A., Nikou, S., & Bouwman, H. (2021b). Business model innovation and firm performance: Exploring causal mechanisms in SMEs. Technovation, 107, 102274. Lange, H. E., Drews, P., & Höft, M. (2021). Realization of Data-Driven Business Models in Incumbent Companies: An Exploratory Study Based on the Resource-Based View. ICIS 2021 Proceedings. Lindgardt, Z., Reeves, M., Stalk, G & Deimler, M. S. (2009). Business Model Innovation, When the Game Gets Tough, Change the Game, Boston: The Boston Consulting Group. Marolt, M., Lenart, G., Borstnar, M. K., Vidmar, D., & Pucihar, A. (2018). SMEs Perspective on Business Model Innovation. 31th Bled eConference Digital Transformation – Meeting the Challenges (June 17 - 20, 2018, Bled, Slovenia). Mazen, A. M. M., Hemmasi, M., & Lewis, M. F. (1987). Assessment of statistical power in contemporary strategy research. Strategic Management Journal, 8(4), 403-410. MIT (2021). Capitalizing on the data economy. MIT Technology Review Insights, 10 p. Retrieved from https://wp.technologyreview.com/wp-content/uploads/2021/12/Capitalizing-on-the-data- economy_121521.pdf Nahr, N., & Heikkilä, M. (2022). Uncovering the identity of Electronic Markets research through text mining techniques. Electronic Markets, 32(3), 1257-1277. Oliveira, M. I. S., & Lóscio, B. F. (2018). What is a data ecosystem?. In Proceedings of the 19th Annual International Conference on Digital Government Research: Governance in the Data Age, pp.1-9. Oliveira, M. I S., Barros Lima, G. D. F., & Farias Lóscio, B. (2019). Investigations into data ecosystems: a systematic mapping study. Knowledge and Information Systems, 61, 589-630. Osterwalder, A., Pigneur, Y, & Tucci, C. (2005). Clarifying Business Models: Origins, Present, and Future of the Concept, Communications of the Association for Information Systems. 16(1), DOI: 10.17705/1CAIS.01601 Available at: https://aisel.aisnet.org/cais/vol16/iss1/1 102 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Osterwalder, A., & Pigneur, Y. (2010). Business Model Generation: A Handbook for Visionaries, Game Changers, and Chal engers. Wiley Otto, B. & Jarke, M. (2019). Designing a multi-sided data platform: findings from the International Data Spaces case. Electronic Markets, 29, 561–580. https://doi.org/10.1007/s12525-019-00362- x Parikka, H., Härkönen, T. & Sinipuro, J. (2021). A fair data economy is built upon col aboration, Sitra studies 191, June 2021, Retrieved from: https://www.sitra.fi/en/publications/a-fair-data- economy-is-built-upon-col aboration Pohle, G. & Chapman, M. (2006). IBM's global CEO report 2006: business model innovation matters, Strategy & Leadership, 34(5), pp. 34-40. Pucihar, A. (2020). The digital transformation journey: Content analysis of Electronic Markets articles and Bled eConference proceedings from 2012 to 2019. Electronic Markets, 30(1), 29–37. https://doi.org/10.1007/s12525-020-00406-7 Pucihar, A., Lenart, G., Kljajić Borštnar, M., Vidmar, D., & Marolt, M. (2019). Drivers and outcomes of business model innovation—Micro, small and medium-sized enterprises perspective. Sustainability, 11(2), 344. Rantanen, M. M., Hyrynsalmi, S., & Hyrynsalmi, S. M. (2019, June). Towards ethical data ecosystems: A literature study. In 2019 IEEE International Conference on Engineering, Technology and Innovation (ICE/ITMC) June, pp. 1-9. Saaristo, A., & Heikkilä, M. (2022). A fair data-driven economy - The future of Europe, CCR Insights 2/2022, Turku School of Economics, Retrieved from: https://www.utu.fi/sites/default/files/public%3A//media/file/ccr_insights_2_2022_afaird atadriveneconomy_thefutureofeurope_v3_saavutettava%281%29.pdf Saebi, T., Lien, L. and Foss, N. J. (2017) What drives business model adaptation? The impact of opportunities, threats and strategic orientation. Long Range Planning, 50 (5), pp. 567-581. Seppälä, A., Birkstedt, T., & Mäntymäki, M. (2021). From ethical AI principles to governed AI. In Proceedings of the 42nd International Conference on Information Systems (ICIS2021). Sitra (2021). Raw data (excel). Retrieved from : https://www.sitra.fi/en/publications/the-future-of- european-companies-in-data-economy-2021/ Soluk, J., Miroshnychenko, I., Kammerlander, N., & De Massis, A. (2021). Family influence and digital business model innovation: the enabling role of dynamic capabilities. Entrepreneurship Theory and Practice, 45(4), 867-905. Spieth, P., & Schneider, S. (2016). Business model innovativeness: designing a formative measure for business model innovation. Journal of business Economics, 86, 671-696. Strauch, M., Pidun, U., & zu Knyphausen-Aufseß, D. (2019). Process matters–How strategic decision-making process characteristics impact capital al ocation efficiency. Long Range Planning, 52(2), 202-220. Teece, D.J., (2010). Business models, business strategy and innovation. Long Range Plan ning. 43 (2–3), 172–194. https://doi.org/10.1016/j.lrp.2009.07.003. Trabucchi, D. and Buganza, T. (2019). Data-driven innovation: switching the perspective on Big Data. European Journal of Innovation Management, 22(1), 23-40. https://doi.org/10.1108/EJIM-01-2018-0017 Ulander, M.; Vierula, T.; Ahomäki, M.; Kultanen, H. & Polifke, A. (2021). The Future of European Companies in Data Economy. Sitra. https://www.sitra.fi/en/publications/the-future-of- european-companies-in-data-economy-2021/ Vermanen, M., Rantanen, M.M., Koskinen, J. (2022). Privacy in Internet of Things Ecosystems – Prerequisite for the Ethical Data Col ection and Use by Companies. In: Kreps, D., Davison, R., Komukai, T., Ishi , K. (eds) Human Choice and Digital by Default: Autonomy vs Digital Determination. HCC 2022. IFIP Advances in Information and Communication Technology, vol 656. Springer, Cham. https://doi.org/10.1007/978-3-031-15688-5_2 Wong, K. K. K. (2013). Partial least squares structural equation modeling (PLS-SEM) techniques using SmartPLS. Marketing Bul etin, 24(1), 1-32. M. Heikkilä, F. Ahmad, J. Heikkilä: The Data Analytic Capability Wheel: Unlocking the Potential of Data-Driven Business Models: An Empirical Investigation into the Role of Ecosystems and Fair Data 103 Use Zuboff, S. (2019). Surveillance capitalism and the challenge of collective action. In New labor forum, 28(1), pp. 10-29). Sage CA: Los Angeles, CA: SAGE Publications. 104 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Appendix Constructs and items Construct Item Potential How much potential you see in the data economy to… Benefit from (Scale: 1 = Strongly disagree. .5 = Strongly agree) data Item 1 Create additional revenue from current business model? Item 2 Create new revenue streams from innovations Item 3 Saving costs BM change Has your company made significant changes in the business model (for example the following elements) during the last two years? (Scale: 1= No, none, 2= Yes, some, 3= Yes, significant) Item 1 Customers (Example: Traditionally we have been B2B company only but now we also/only have consumer customers.) Item 2 Channels (Example: Traditionally we have sold our products through wholesalers but we have now opened up e-commerce site for direct access.) Item 3 Value proposition (Example: Previously we have sold products but now we offer services where our own and others´ products are part of.) Item 4 Activities (Example: We have stopped or started new activities - previously we have been purely a hardware company but now we have software coding also.) Item 5 Resources (Example: We have acquired new or discarded old resources (factories, people, skills).) Item 6 Partners (Example: We have new partners helping us in creation and production of our value proposition to our customers, e.g. offshoring, near-shoring.) Item 7 Revenue models (Example: From one-off transactions to time- based subscription.) Data How well do the following statements describe the current ecosystem status of your company? participation (Scale: 1 = does not describe at all .. 5 = describes very well) Item 1 Our company is part of one or more data ecosystems M. Heikkilä, F. Ahmad, J. Heikkilä: The Data Analytic Capability Wheel: Unlocking the Potential of Data-Driven Business Models: An Empirical Investigation into the Role of Ecosystems and Fair Data 105 Use Construct Item Item 2 We are planning to facilitate data ecosystems Item 3 Our company is the facilitator of one or more data ecosystems Fair & Please indicate whether the following objectives of data use sustainable are taken into practice in your company (Scale: 1 = totally data usage disagree .. 5 = totally agree) practices Item 1 We offer our customers easy tools to access and manage their personal and/or business data Item 2 We communicate about the use of data in our corporate social responsibility reporting Item 3 Data we gather from consumers is available for them to use in other services outside our company Item 4 We have defined ethical rules for our organization for using, collecting and sharing data Item 5 We strive to create trust by acting and behaving transparently Item 6 Our digital services are designed to respect privacy Item 7 Our digital services are designed to respect control over personal and business data Item 8 Company’s consideration of the rights of individuals and/or organizations exceeds statutory requirements Data-driven From the following data economy related statements, choose innovation how well they describe your company's current business. (Scale 1 = does not describe at all .. 5 = describes very well) Item 1 In product development, we primarily invest in continuous and gradual improvements in our products. Item 2 We wish to complement our data from several different data sources and interactive situations so that we can create the best possible experience for our customers. Item 3 We continuously invest in the innovation of new products / services Item 4 We create value from data, not just for our operations, but also for society, people, and environment. 106 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY DIGITALISATION IN ROPAX PORTS: THE TYPOLOGY OF AVAILABLE SOLUTIONS ANASTASIA TSVETKOVA,1 YIRAN CHEN,1 IRINA WAHLSTRÖM,1 ANDREI-RAOUL MORARIU,1 BOGDAN IANCU,1 MAGNUS HELLSTRÖM1,2 1 Åbo Akademi University, Faculty of Science and Engineering, Turku, Finland anastasia.tsvetkova@abo.fi, yiran.chen@abo.fi, irina.wahlstrom@abo.fi, andrei- raoul.morariu@abo.fi, bogdan.iancu@abo.fi 2 University of Agder, School of Business and Law at the, Department of Working Life and Innovation, Norway magnus.hel strom@abo.fi Sea logistics is slow to adopt digitalisation technologies and still relies heavily on old communication and data exchange methods. However, digitalising activities in RoPax ports can improve logistic chains. Managing and combining flows and services of cargo and passengers leads to several specific chal enges, creating Keywords: digitalisation opportunities. Data was gathered through desktop port, studies, interviews, site visits, and workshops to identify available digitalisation, and potential digital solutions. The research resulted in a logistics, RoPax, typology of digital solutions classified based on their functional Bled value and how much they impact logistics chains. eConference DOI https://doi.org/10.18690/um.fov.4.2023.7 ISBN 978-961-286-751-5 108 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Digitalisation can al ow companies to gain a competitive advantage, enabling new value creation (Kamalaldin et al., 2020) and opportunities for business model innovation. Business model innovation can stem from exploiting digital technologies and digital technology development at the firm (Cozzolino et al., 2018). Digitalisation’s impact on business models has been covered in extensive literature reviews (Caputo et al., 2021). While many studies review the impact of the firms’ capability to innovate business models with digitalisation’s help or benefit from digital technologies, in this research, the perspective of digital solutions providers and developing their value propositions towards users of digital solutions interests us. While industries face ubiquitous digitalisation in the sense it has concerned innovative and traditional sectors (Teece, 2010; Warner & Wäger, 2019), the specifics of digitalisation and emerging business models are context-specific. Studies have shown the effect of increasing digitalisation in the manufacturing industry regarding the solutions that manufacturing firms offer (Kohtamäki et al., 2019) and the specific value process the industry can gain from digitalisation (Kamalaldin et al., 2021). However, as Caputo et al. (2021) propose, continuing research is needed on different forms and archetypes of business models developed in recent years (Caputo et al., 2021). Maritime logistics, which can be considered a network industry, faces a different set of chal enges and thus can gain unique benefits from digitalisation, which brings forward the specific value digitalisation can create for business actors in the logistics chain. In particular, port authorities (PAs) are increasingly implementing digital solutions to improve coordination among many port stakeholders to make operations safer and more sustainable and efficient along logistics chains (Tijan et al., 2021; Tsvetkova, Gustafsson et al., 2021). This article explores the case of RoPax ports, referring to ports handling vessels capable of carrying wheeled cargo and passengers. Significant research has been devoted to digital solutions that large container ports have introduced (Haezendonck & Langenus, 2019; Henríquez et al., 2022). However, small to medium-sized RoPax ports face unique digitalisation chal enges and opportunities that make it chal enging to compare their efforts. In particular, smal and medium-sized ports are limited regarding resources and A. Tsvetkova, Y. Chen, I. Wahlström, A.-Raoul Morariu, B. Iancu, M. Hel ström: Digitalisation in RoPax ports: the Typology of Available Solutions 109 investments (Del Giudice et al., 2021; Inkinen et al., 2019; von Malmborg, 2004) and have challenges in understanding digitalisation’s influence on strategic business development (Inkinen et al., 2019). We are interested in the new value propositions emerging in the context of the RoPax port ecosystem and set out to explore how value is created through digitalisation in the logistics sector. In the study of digital solutions that have been implemented or are under development in RoPax ports, we explore the various types of digital solutions regarding their functionalities and capabilities, thus value creation potential for the PAs who manage and develop the port area. Given the multitude of such solutions, understanding the differences and interrelations between these solutions is relevant, thus focusing on the digital business ecosystem that is forming and evolving (Hanelt & Schneider, 2020). This article presents the typology for relevant digital solutions for RoPax ports, forming the basis for understanding new value propositions in this underexplored empirical context. In addition to the solutions already implemented by ports and confirmed as relevant, this paper considers solutions under development through collaboration with port actors as part of the relevant digital solutions for RoPax ports. 2 Literature Review 2.1 Digitalisation and its Effect on Business Models and Ecosystems Digitalisation facilitated transitioning the business model innovation within the same industry towards cross-industrial and ecosystem business model innovation (Kamalaldin et al., 2021; Leminen et al., 2020; Sjödin, Parida, Jovanovic et al., 2020). However, materialising digitalisation demands developed digital infrastructure, suitable technologies, and interoperability of data format. Business model innovation and ecosystem transition go hand-by-hand. Business model innovation enabled by introducing new digital solutions often requires redefining industry structures and business ecosystems due to the change in information or operation flows (Huikkola et al., 2020; Linde et al., 2021). Hence, incumbent companies must consider ecosystem transition dynamics to stay alert for new value creation and capturing opportunities. 110 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Regarding maritime logistics, the advent of ‘smart ports’ are opportunities to create more value through data-based services and data-driven business models. The general trend of ‘infrastructure as a service’ will likely affect ports’ business models as the information about infrastructure use becomes more valuable than possessing that infrastructure (Tsvetkova et al., 2020). While coordinating activities among port stakeholders requires data sharing, more complex solutions like providing optimisation require data analysis, such as predictive and prescriptive analytics, to predict events and plan the optimal resource al ocations (Haraldson et al., 2021). This brings the need for aligning value creation in multiple firms and affects incumbent actors in the port ecosystems and digital solutions providers. 2.2 Typologies for Digital Solutions in Industrial Sectors Digitalisation-related literature focused on numerous perspectives on the types of business models and offerings stemming from digitalisation (Caputo et al., 2021; Kohtamäki et al., 2019). In particular, this literature has explicated the capabilities of digital solutions and smart products and how these capabilities enable new value creation: digital servitisation-based business models and value propositions that digitalisation can enable (Caputo et al., 2021; Coreynen et al., 2017; Porter & Heppelmann, 2014; Sjödin, Parida, Kohtamäki et al., 2020). Digitalisation’s impact on business models has been described in diverse industrial contexts. Kamaladin et al. (2021) studied digitalisation in the process industry and how it enables process innovation; Tsvetkova et al. (2021) studied maritime infrastructure digitalisation, mentioning that a digital solution’s outcome differs based on the layers each solution addresses: infrastructure, service, and system layers. Digitalisation in the maritime sector is changing the overal business and operational processes (Tsvetkova, Hellström et al., 2021). Like other areas in maritime logistics, port digitalisation’s implications on the business models of relevant actors remain understudied. The current research on digitalisation in ports has been focused on phases of port digitalisation (Inkinen et al., 2021), influences of digitalisation on ports’ business performance (Holmström et al., 2019), barriers for implementing digitalisation (Brunila et al., 2021), and the comparison of technological configurations (Hirata et al., 2022). Few studies approached the topic with empirical A. Tsvetkova, Y. Chen, I. Wahlström, A.-Raoul Morariu, B. Iancu, M. Hel ström: Digitalisation in RoPax ports: the Typology of Available Solutions 111 examples, mostly among large container ports, such as Amsterdam and Barcelona (Anwar et al., 2019; Heikkilä et al., 2022; Henríquez et al., 2022). Compared to large container ports, smal er-sized ports face different difficulties concerning digitalisation (Inkinen et al., 2019). PAs have fewer resources for implementing digitalisation. Conversely, the traffic flow has a different pattern in RoPax ports due to the simultaneous involvement of passenger and cargo traffic. Thus, the value digital solutions create is critical but unapparent regarding RoPax ports. We aim to contribute to understanding how value can be created through digitalisation in RoPax ports with our study. 3 Methodology 3.1 Research Design The current research results from a three-year project devoted to developing digital solutions for smart RoPax ports. The project aimed to improve operation efficiency through collectively developed digital solutions for RoPax ports. Different types of actors were present: some are incumbent in the port operation ecosystem, such as PAs, shipping companies, and a set of digital solution providers, while others are new and aim to join the ecosystem based on their successful experience in another ecosystem. We did an explorative case study (Mills et al., 2012) to develop a typology of digital solutions relevant to RoPax ports based on a systematic combining approach grounded in abductive logic. The research process was continuous and manifested in the evolution of a theoretical framework, empirical fieldwork, and case analysis (Dubois & Gadde, 2002). As we collected data on the available and planned digital solutions for RoPax ports and studied the empirical context, we iteratively improved the theoretical framework, constituting the basis for the typology of digital solutions in RoPax ports. As new data appeared, we reconsidered the typology and searched for categories al owing us to accommodate the empirical evidence. This iterative process involved six researchers and the col ection of feedback from the project participants. 112 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3.2 Data Collection and Analysis We col ected the data through desktop study, interviews (including site visits to ports), and group workshops. Data collection occurred inside and outside the focal project. We explored the available and potential digital solutions developed in the focal project. Conversely, we benchmarked six Northern European RoPax ports to identify digital solutions that PAs implemented or planned. We interviewed many relevant organisations regarding port digitalisation to support our understanding of the function and value of digital solutions. In particular, we interviewed municipality departments and digital solution providers in the urban and transportation sectors. The interviews lasted about an hour each; one or more managerial-level representatives participated in a semi-structured interview. The interviews were recorded and transcribed, and we made observation notes. Altogether, 40 interviews were conducted during the project and were used in this paper’s research. Table 1 presents a list of the primary research data. Based on the interview outcomes, we gathered additional secondary that interviewees mentioned as relevant documents. Apart from interviews with individual actors, six field visits to ports and two workshops on identifying the types of digital solutions relevant to RoPax ports were organised. Field visits were made to the two focal project ports. One workshop focused on passenger traffic and another on cargo flow, where the participants identified the bottlenecks, goals, potential digital solutions, and collaboration steps for improving passenger and cargo traffic flows. The workshops were approximately two hours each, recorded and transcribed. Through these interactions, we better understood the current operations’ chal enges and development bottlenecks. A. Tsvetkova, Y. Chen, I. Wahlström, A.-Raoul Morariu, B. Iancu, M. Hel ström: Digitalisation in RoPax ports: the Typology of Available Solutions 113 Table 1: Overview of Primary Data Sources Companies Informants' roles Number of Form of participants interaction (number of interactions) Port PA: technical director, COO, 5 (2) Interviews authorities sales and deputy managing director, and IT manager Benchmarking RoPax ports 5 (5) Interviews Shipping Head of digitalisation 1 (2) Interview; company interview and site visit Head of IT group, captain, 4 (1) Interview and and cargo planner site visit Terminal manager 1 (1) Interview Sales manager 1 (1) Interview Operative manager 1 (1) Interview and site visit Incumbent Research leaders 2 (1) Interview digital CEO and project manager 2 (1) Interview solution providers CEO 1 (1) Interview CEO and research leader 2 (1) Interview New digital CPO and CFO 2 (1) Interview solution providers Director and research leader 2 (1) Interview CEO 1 (1) Interview CEO 1 (1) Interview Research leader 1 (1) Interview Project Project update workshops 36 (9) Workshop participants Various Workshops devoted to 36 (2) Workshop stakeholder identifying digital solutions s Municipalit Land planning department 4 (1) Interview y, and traffic planner authorities, Regional logistic coordinator 1 (1) Interview and external stakeholder Carbon neutral development 1 (1) Interview s Urban mobility developer 1 (1) Interview City infostructure builder 1 (1) Interview 114 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Traffic management 1 (1) Interview department Road digital solution 1 (1) Interview provider, CEO Total 78 (40) Note: The numbers of unique participants were reported. The numbers of interviews and workshops we conducted or participated in are in parentheses. It is important to note that a relatively bigger volume of data was col ected in relation to the two ports that participated in the research project described above. The primary data regarding the other four ports included one-hour interviews with the representative of each PA. However, the secondary data such as annual reports and strategy documents have been studied to enrich the study. We acknowledge the possibility of other biases influencing the data collection process. We made efforts to minimize biases in the data col ection process by including diverse perspectives and roles, using a variety of data col ection methods, and continuously revising our theoretical framework. The typology was also validated with the project participants in one of the workshops and through other regular interactions, and its relevance and applicability was confirmed. 4 Findings 4.1 Challenges Faced by RoPax Ports The coexistence of cargo, vehicle, and passenger flows in RoPax ports leads to several specific challenges, creating digitalisation opportunities. During the research project described in section 3, we have identified several recurring chal enges that RoPax ports face. Passenger flow improvement faces chal enges regarding information sharing and safe travel. For instance, the unavailability of real-time intermodal traffic information causes delays along the transit journeys to and from the terminal; relevant information comprises transport connections, schedules, services, and notifications on traffic disturbances. During check-in, security, and embarkation procedures cause unnecessary crowding and queuing in the passenger terminal, which the uptake and implementation of digital solutions could alleviate. The technology could take safety A. Tsvetkova, Y. Chen, I. Wahlström, A.-Raoul Morariu, B. Iancu, M. Hel ström: Digitalisation in RoPax ports: the Typology of Available Solutions 115 and security matters to another level as tools for managing crowds and identifying passengers’ unusual behaviour or unattended luggage. The multimodal transport chain is hampered by uncoordinated road traffic pulses and congestion at ship arrivals and departures inside the port, its associated urban area, and its main approach roads – a problem less likely to diminish with residential and recreational urban areas expanding and the associated shrinkage of the port facilities and stowing areas. Furthermore, the vessels’ cargo capacity will likely grow, further worsening the overcrowding of the road network. Traffic jams, in their turn, wil increase idling and unnecessary emissions. Due to relatively short distances over the sea and ship turnaround times in ports, the RoPax short sea segment is more time-sensitive concerning timely and efficient loading and unloading procedures, as delays may be difficult to compensate with speed out at sea. Accumulating trucks in the port area and using this area as a waiting or parking area long before the actual departure is another unwanted phenomenon. This is principally due to a limited number of dedicated resting and parking areas for trucks in the port’s vicinity. Introducing an integral just-in-time (JIT) solution with an associated pre-parking concept offered to road haulage customers would benefit al stakeholders’ operations and performance. Truck drivers would be informed and cal ed in when their vehicle could drive to the port area through a dedicated application or another interface. To a predetermined degree, the given slot times would enhance specific trucks’ orderly and timely arrival in certain batches, benefiting the actual optimal onboard stowing plan and order. 4.2 Typology Development During the research, it became apparent that digital solutions designed for or implemented in RoPax ports provide different functions – adding value to port actors – and involve smaller or bigger constellations of actors and respective data. Considering their effect, the solutions ranged from those addressing limited activities within one port actor’s boundaries to those affecting the efficiency of whole logistics chains. Certain solutions, such as those for situational awareness (SA) in ports, could provide data for other solutions that offer, for example, cargo or passenger flow optimisation. 116 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY The solutions were categorised into two dimensions. The first dimension includes six capabilities and functions of digital solutions: Communication, Visualisation, Monitoring, Control & Automation, Prediction, and Optimisation. The capabilities were inspired by the four capabilities proposed by Porter and Heppelmann (2014) and further developed with three additional capabilities. Communication refers to the ability of digital solutions to exchange information between different actors, whereas Visualisation concerns the presentation of data in an easily understandable way. Monitoring and Control & Automation capabilities al ow for real-time monitoring and control of port operations, respectively. Prediction involves the use of data to anticipate events and potential problems, and Optimisation refers to the ability to improve efficiency and performance. Examples of solutions that exhibit these capabilities are digital twins of port infrastructure for Visualisation or Monitoring, traffic and cargo management systems for Control & Automation, and predictive algorithms for Prediction. The typology development includes these capabilities and functions as part of the framework for analysing and classifying digital solutions in RoPax ports. The second dimension included three layers of digitalising port operations, following the layers of digitalisation in maritime logistics (Tsvetkova, Gustafsson et al., 2021): infrastructure, service, and system layers. The infrastructure layer includes solutions directly related to efficiently operating and maintaining maritime infrastructure, such as digital twins of port infrastructure and smart buoys, that help identify maintenance needs through predictive algorithms. The service layer includes solutions to improve how the users utilise this infrastructure. These solutions include managing the throughput of traffic and cargo and vessel flows through ports, so the maritime infrastructure is utilised efficiently, and the service quality is maximised for cargo shippers, ship operators, and other parties relying on port infrastructure. Combining the data related to maritime infrastructure and cargo flows makes it possible to achieve efficiencies on a larger scale at a system level, ensuring a smart cargo and passenger flow through digital corridors, end-to-end journeys, and optimising whole supply chains. A. Tsvetkova, Y. Chen, I. Wahlström, A.-Raoul Morariu, B. Iancu, M. Hel ström: Digitalisation in RoPax ports: the Typology of Available Solutions 117 4.3 Analysis of Digital Solutions for RoPax Ports We identified various solutions implemented or planned for implementation in the RoPax ports we studied (marked in blue in Figure 1) and the solutions developed or discussed within the case project (marked in yel ow in Figure 1). One finding is that RoPax ports strongly focus on digitising and improving communication among the many actors. Without digitising documentation flow and notification processes, more profound digitalisation and integration efforts across organisations is infeasible. The solutions the studied ports mention mostly focus on documentation exchange and notifications regarding cargo and vessel arrival. Several solutions relate to communication for passenger flow, creating an intriguing opportunity for improving customer satisfaction concerning passengers as customers of RoPax terminals. Moreover, using digital solutions, such as mobile applications or other interfaces, could help streamline security and embarkation procedures, reducing crowding and queuing in passenger terminals. Regarding monitoring, ports have implemented solutions for monitoring processes directly related to the infrastructure layer, such as energy use or air quality in the port area. Naturally, security monitoring exists in any RoPax port. However, new solutions were proposed for more automated security monitoring, including, for example, automated identification of abandoned luggage or identification of crowding in the terminal. These solutions can help to address safety and security matters as discussed in section 3.1. The project also developed solutions for identifying and counting passengers and vehicles, which can provide input for several automation or optimisation solutions. Control and automation are seemingly high on the agenda for many ports, as passenger and vehicle flow automation can save costs, improve customer satisfaction, and reduce the workload of port workers. Many overlaps between the solutions developed within the case project and solutions were implemented (often partly) or planned by RoPax ports, confirming that the value of these solutions is indisputable. 118 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Figure 1: The typology of digital solutions in RoPax ports Similar to visualisation, solutions falling under prediction capability are crucial for human decision-making in RoPax terminals or optimisation solutions. There are opportunities for predicting critical data points concerning passenger, vehicle, and vessel flow, which can ultimately help improve the timely and efficient loading and unloading procedures, reducing delays that may be difficult to compensate with speed out at sea. Regarding optimisation, ports have implemented several solutions addressing the infrastructure layer and are thus limited to the scope of a RoPax port (heat and power optimisation in the terminal area, smart security). Other solutions aim at integrating several processes going beyond actual port operations. However, the system layer is addressed much less. While RoPax ports have mentioned intel igent supply chains, it has yet to be developed into a working concept. The case project has produced A. Tsvetkova, Y. Chen, I. Wahlström, A.-Raoul Morariu, B. Iancu, M. Hel ström: Digitalisation in RoPax ports: the Typology of Available Solutions 119 concepts for several optimisation solutions. For instance, passengers’ end-to-end journey planner can reduce the passengers' need for private vehicles when arriving at the port, thereby al eviating traffic congestion in the port area. Another solution for integrating truck traffic flow with port and vessel operations can enhance specific trucks' orderly and timely arrival in certain batches, benefiting the actual optimal onboard stowing plan and order. 5 Discussion and Conclusions The digitalisation of end-to-end supply chains can fundamental y change logistics and thus has implications for the future management of maritime infrastructure. Specifical y, increased transparency and a better understanding of cargo and vessel flow through ports are inputs for smarter decisions regarding port infrastructure investments going forward. Combining the data related to maritime infrastructure and cargo flows makes it possible to achieve efficiencies on a larger scale at a system level, ensuring a smart cargo flow through digital corridors and optimising whole supply chains. Such synchronous modality can allow significant transport cost reductions and optimum utilisation of transport infrastructure while adhering to the respective delivery conditions (Tsvetkova, Gustafsson et al., 2021). We observed potential differences between the technology push and market pul regarding digital solutions for the RoPax port. In particular, solutions for visualisation or prediction developed in the case project can provide valuable input for other functions, such as optimisation or control. It is necessary to understand, for example, how predictions or visualisations of terminal flows can be used in, e.g. security management or passenger and cargo flow optimisation to develop a viable business offering. Besides the unclear value proposition of such solutions, we identified another chal enge related to the digitalisation in RoPax ports, namely, the difficulty of justifying investment in digital infrastructure that may not directly benefit the port authorities but require resources from them. While earlier studies (Kohtamäki et al., 2019) focused on making typologies of different digital BMs, this study specifically addresses the value propositions of digital solutions and their place in a functioning digital business ecosystem. The typology of digital solutions for RoPax ports can be a benchmark for different solution providers to identify their solution’s role and the value it can create for port 120 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY actors. Simultaneously, it identifies common trends in digitalising RoPax ports, which can be a reference for PAs and port actors to discover new technological opportunities. For example, a port operator could use this typology to identify areas where they could improve their operations using digital solutions, such as optimizing cargo flows or improving situational awareness in the port. A digital solution provider, on the other hand, could use this typology to identify gaps in the market and develop new solutions that address specific capabilities or functions that are currently lacking in the market. By using this typology, both port actors and solution providers can make more informed decisions and identify new opportunities for innovation. Further research would be needed to better understand how the structure of port business ecosystems shapes value creation. In particular, the interplay between the value captured by PAs and shipping companies from their customers is interesting regarding digitalising port operations. Further, this article's typology is based on earlier studies' proposed capabilities but adjusted with empirical evidence. While we believe it captures the current digitalisation state in RoPax ports, the evolving nature of digital solutions may necessitate future revisions. An example is the 'autonomy' capability, initial y part of Porter and Heppelman's (2014) four capabilities, but omitted in this typology as no solutions provided it. Autonomous shipping may require RoPax ports to incorporate autonomous capabilities in the future. We recognize geographical limitations as our study focused solely on ports in Northern Europe. Furthermore, this typology is tailored to the RoPax shipping context, which enhances its empirical significance, but also restricts its applicability to other ports and shipping modes. A comparative study of the types of digital solutions in different shipping segments is a promising avenue for further research. References Anwar, M., Henesey, L., & Casalicchio, E. (2019). Digitalization in Container Terminal Logistics : A Literature Review. 1–25. Brunila, O. P., Kunnaala-Hyrkki, V., & Inkinen, T. (2021). Hindrances in port digitalization? Identifying problems in adoption and implementation. European Transport Research Review, 13(1). Caputo, A., Pizzi, S., Pel egrini, M. M., & Dabić, M. (2021). Digitalization and business models: Where are we going? A science map of the field. Journal of Business Research, 123, 489–501. Coreynen, W., Matthyssens, P., & van Bockhaven, W. (2017). Boosting servitization through digitization: Pathways and dynamic resource configurations for manufacturers. Industrial Marketing Management, 60, 42–53. A. Tsvetkova, Y. Chen, I. Wahlström, A.-Raoul Morariu, B. Iancu, M. Hel ström: Digitalisation in RoPax ports: the Typology of Available Solutions 121 Cozzolino, A., Verona, G., & Rothaermel, F. T. (2018). Unpacking the Disruption Process: New Technology, Business Models, and Incumbent Adaptation. Journal of Management Studies, 55(7), 1166–1202. Dubois, A., & Gadde, L. E. (2002). Systematic combining: An abductive approach to case research. Journal of Business Research, 55(7), 553–560 Hanelt, A., & Schneider, S. (2020). Building Platform Ecosystems for IoT: Exploring the Impact on Industrial-Age Firms Digital Transformation View project Management of Social and Sustainable Business Models View project. Haraldson, S., Lind, M., Breitenbach, S., Croston, J. C., Karlsson, M., & Hirt, G. (2021). The Port as a Set of Socio-technical Systems: A Multi-organisational View. In Maritime Informatics. Heikkilä, M., Saarni, J., & Saurama, A. (2022). Innovation in Smart Ports: Future Directions of Digitalization in Container Ports. Journal of Marine Science and Engineering, 10(12). Henríquez, R., Martínez de Osés, F. X., & Martínez Marín, J. E. (2022). Technological drivers of seaports’ business model innovation: An exploratory case study on the port of Barcelona. Research in Transportation Business & Management, 43, 100803. Hirata, E., Watanabe, D., & Lambrou, M. (2022). Shipping Digitalization and Automation for the Smart Port. Supply Chain - Recent Advances and New Perspectives in the Industry 4.0 Era [Working Title]. Holmström, J., Holweg, M., Lawson, B., Pil, F. K., & Wagner, S. M. (2019). The digitalization of operations and supply chain management: Theoretical and methodological implications. Journal of Operations Management, 65(8), 728–734. Huikkola, T., Rabetino, R., Kohtamäki, M., & Gebauer, H. (2020). Firm boundaries in servitization: Interplay and repositioning practices. Industrial Marketing Management, 90, 90–105. Inkinen, T., Helminen, R., & Saarikoski, J. (2019). Port digitalization with open data: Challenges, opportunities, and integrations. Journal of Open Innovation: Technology, Market, and Complexity, 5(2). Inkinen, T., Helminen, R., & Saarikoski, J. (2021). Technological trajectories and scenarios in seaport digitalization. Research in Transportation Business and Management, 41, 2210–5395. Kamalaldin, A., Linde, L., Sjödin, D., & Parida, V. (2020). Transforming provider-customer relationships in digital servitization: A relational view on digitalization. Industrial Marketing Management, 89, 306–325. Kamalaldin, A., Sjödin, D., Hul ova, D., & Parida, V. (2021). Configuring ecosystem strategies for digital y enabled process innovation: A framework for equipment suppliers in the process industries. Technovation, 105, 102250. Kohtamäki, M., Parida, V., Oghazi, P., Gebauer, H., & Baines, T. (2019). Digital servitization business models in ecosystems: A theory of the firm. Journal of Business Research, 104, 380–392. Leminen, S., Rajahonka, M., Wendelin, R., & Westerlund, M. (2020). Industrial internet of things business models in the machine-to-machine context. Industrial Marketing Management, 84, 298–311. Linde, L., Sjödin, D., Parida, V., & Wincent, J. (2021). Dynamic capabilities for ecosystem orchestration A capability-based framework for smart city innovation initiatives. Technological Forecasting and Social Change, 166, 120614. Mil s, Al. J., Durepos, G., & Wiebe, E. (2012). Exploratory Case Study. In Encyclopedia of Case Study Research. SAGE Publications. Porter, M. E., & Heppelmann, J. E. (2014). How Smart, Connected Products Are Transforming Competition. Harvard Business Review, 92(11), 64. Sjödin, D., Parida, V., Jovanovic, M., & Visnjic, I. (2020). Value creation and value capture alignment in business model innovation: A process view on outcome‐based business models. Journal of Product Innovation Management, 37(2), 158–183. Sjödin, D., Parida, V., Kohtamäki, M., & Wincent, J. (2020). An agile co-creation process for digital servitization: A micro-service innovation approach. Journal of Business Research, 112, 478– 491. 122 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Teece, D. J. (2010). Business models, business strategy and innovation. Long Range Planning, 43(2–3), 172–194. Tijan, E., Jović, M., Aksentijević, S., & Pucihar, A. (2021). Digital transformation in the maritime transport sector. Technological Forecasting and Social Change, 170, 120879. Tsvetkova, A., Gustafsson, M., & Wikström, K. (2020). The digitalization of port infrastructure. Tsvetkova, A., Gustafsson, M., & Wikström, K. (2021). Digitalizing maritime transport: digital innovation as a catalyzer of sustainable transformation. A Modern Guide to the Digitalization of Infrastructure, 123–148. Tsvetkova, A., Hellström, M., & Ringbom, H. (2021). Creating value through product-service-software systems in institutionalized ecosystems – The case of autonomous ships. Industrial Marketing Management, 99, 16–27. Warner, K. S. R., & Wäger, M. (2019). Building dynamic capabilities for digital transformation: An ongoing process of strategic renewal. Long Range Planning, 52(3), 326–349. MALICIOUS INSIDER THREAT TYPES – AN EMPIRICAL ANALYSIS MANFRED HOFMEIER, ISABELLE HAUNSCHILD, ULRIKE LECHNER Universität der Bundeswehr München, Department of Computer Science, Neubiberg, Germany manfred.hofmeier@unibw.de, isabel e.haunschild@unibw.de, ulrike.lechner@unibw.de Malicious insider threats represent a particular challenge not only for defense, but also for research, as it is estimated there is a high number of unreported cases. Current taxonomies and typologies usually focus on specific aspects, such as goal or motivation, and tend to have tight boundaries. A number of malicious insider threat attack scenarios were identified in our research through qualitative interviews, enhanced with a game-based creative approach. The resulting data was used to develop a malicious Keywords: insider threat typology in an empirical bottom-up approach. We insider developed an analysis scheme from existing taxonomies and threats, typology, typologies and used it in an empirical analysis of malicious insider information roles and attack scenarios. We were able to identify eleven security, archetypes of malicious insider threats considering multiple Bled eConference, facettes. This paper describes the analysis and the identified research types. paper DOI https://doi.org/10.18690/um.fov.4.2023.8 ISBN 978-961-286-751-5 124 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The European Union Agency for Cybersecurity (ENISA) defines an insider threat as "an action that may result in an incident, performed by someone or a group of people affiliated with or working for the potential victim" (ENISA, 2020). Insider threats can also be distinguished as accidental or malicious (intentional). In this work we take a close look at malicious insider threats. These represent a particular chal enge not only for actual protection, but also for research, as it is estimated there is a high number of unreported cases. Organizations tend not to disclose information on malicious insider threats (e.g., not to lose the trust of customers or partners), and typically only cases with news value become public. Thus, the cases examined in the research are not necessarily representing the current insider threat landscape. In general, information systems literature such as the UMISPC (Moody et al., 2018) addresses the field of insider threats primarily regarding accidental insider threats while the malicious type is usual y not considered. As we col ected plenty of malicious insider threat attack scenarios in our research through an interview survey and a creative approach (using the serious game "Operation Digital Butterfly" (Hofmeier, 2021), which is designed to col ect plausible insider roles and attacks), we use this data to develop a malicious insider threat typology. Current taxonomies and typologies usual y focus on specific aspects, such as goal or motivation, and tend to have tight boundaries. For example, the ENISA Threat Landscape Report distinguishes five types of insider threats by rationales and objectives (ENISA, 2020): careless workers mishandling data, violating policies or installing unauthorized applications; inside agents who steal information on behalf of outsiders; disgruntled employees who seek to harm their organization; malicious insiders who use existing privileges to steal data for personal gain; third-parties who compromise security through intel igence, misuse or malicious access to or use of an asset. Cappelli et al. distinguish three types of insider threats by objectives in the “CERT Guide to Insider Threats” (Cappelli et al., 2012): theft (e.g., of intellectual property or other data), sabotage (malicious manipulation of data or processes or causing reputational damage) and fraud (e.g., stealing financial goods). The German Insurance Association (GDV) defines four perpetrator types (Bundeskriminalamt, 2020): The crisis perpetrator who is triggered by crisis events in private or professional M. Hofmeier, I. Haunschild, U. Lechner: Malicious Insider Threat Types – An Empirical Analysis 125 life that threaten status and lifestyle; the inconspicuous who takes advantage of an emerging opportunity; the perpetrator with an economic-criminological disorder who actively seeks or creates opportunities to commit a crime, and the dependent who is usually hierarchical y subordinate to a main offender or owes the main offender a favor and fears repression in case of refusal to cooperate. Some taxonomies and typologies distinguish by insider position (Cole & Ring, 2005; Magklaras & Furnell, 2001; Bundesamt für Sicherheit in der Informationstechnik, 2018) or attack vector (Phyo & Furnell, 2004). Homoliak et al. (2019) provide a comprehensive taxonomy with multiple aspects. However, there are only a few empirical y based typologies of malicious insider threats in terms of comprehensive archetypes. Therefore, the aim of this research is to develop archetypes of malicious insider threats using our col ected data in an empirical bottom-up approach. 2 Analytical Approach We analyzed malicious insider threat roles and attack scenarios using scheme-based content analysis to develop the typology. 2.1 Material for Analysis The data used in this analysis is interviews and serious game results. Interviews and games were part of LIONS and NutriSafe research projects. The interviews were conducted from December 2021 to August 2022 with 13 experts from various organizations (table 1). The experts were selected from the LIONS project research network, while the selection criterion was the possibility of points of contact with the subject of malicious insiders. The interviews were structured using an interview guide, and topics were real-world cases, plausible scenarios, and assessments of malicious insider threats. The interviews were recorded, and summarized. Summaries were then sent to the respective interview partner(s) for a review regarding correctness (and anonymization). The interview study col ected 21 instances or groups of attack scenarios, subject to the analysis presented in this paper. 126 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 1: Interviews used in the analysis (anonymized) Organization Interview partner(s) Consulting (logistics) Director Energy Senior Expert Penetration Testing Energy Senior Expert Cyber Forensics Senior Expert Cyber Forensics Compliance Investigations Energy Information Security Manager Civil engineering Manager Software Development Security technology Information Security Officer n.a. Information Security Expert n.a. People Manager Industrial software Managing director Public authority Information Technology Manager Security software CEO Table 2: Serious game iterations Date Game board Players Attack Scenarios May 2020 Meat Production 6 2 Jul 2020 Meat Production 15 4 Oct 2020 Logistics Hub 7 3 Nov 2020 Logistics Hub 15 4 Mar 2021 Logistics Hub 12 3 Sep 2021 Travel Management 12 3 Feb 2022 Travel Management 10 6 Feb 2022 Travel Management 9 6 Feb 2022 Travel Management 9 6 Jul 2022 Logistics Hub 13 3 Because interviews have limited capabilities regarding the field of malicious insiders (e.g., typically, there is no knowledge of intentions or motivations of perpetrators), a creative game-based approach was also used. The serious game "Operation Digital Butterfly" (Hofmeier, 2021) – in an earlier version "Operation Digital Ant" (Hofmeier, 2020; Hofmeier & Lechner, 2021) – was developed – using a design science approach according to Hevner et al. (2004) – to generate fictional but realistic M. Hofmeier, I. Haunschild, U. Lechner: Malicious Insider Threat Types – An Empirical Analysis 127 attack scenarios by malicious insiders. "Operation Digital Butterfly" is a tabletop game with a game board in the center, depicting an infrastructure of an organization. Three to four teams compete against each other in creating roles and attacks of malicious insiders, using role cards and scene cards. The game board describes the environment wherein insider threats take place. During our research, the game has been played with three different game boards: slaughterhouse and cutting plant, logistics hub with warehouse, and travel management in a public authority. Each team develops an insider role, an attack, and a security measure using a card deck. The cards structure the team discussion and the presentation of the attack vector and security measure developed in the team discussion. The teams are instructed to answer four questions on the role card to guide the creative design of attack measures: • Who is the insider (position in the organization)? • What does the insider want to achieve (intention)? • Why does the insider want that (motivation)? • How does the insider justify this to himself/herself (neutralization)? The attack is developed using the scene cards. The filmmaking metaphor is used to make development and descriptions of attacks easy – also for players not used to formal notations. A threat is a sequence of scenes. This way, each team can tell their fictional insider attack by using a sequence of scenes. To make the game more fun and to raise awareness about countermeasures to insider attacks, each team fil s out a security measure card. Teams are instructed to anticipate possible attacks from the other teams (the roles are known) and develop an adequate countermeasure. This measure is valid for the attack plans of al teams and is then taken into account when rating the attacks. The winning team is determined through a rating system, in which the teams rate each other by three given categories: (1) Plausibility of role, (2) plausibility of the attack story, and (3) damage potential. Each team can give up to ten points for each category to the other teams. Note that the most important category for later analyses is "plausibility". This incentivizes that the developed attacks and roles are – to some extent – realistic and fit the profile of the role. The "damage potential" category makes the teams more likely to develop attacks that cause significant damage and therefore are of particular interest in 128 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY security research. The game is also accompanied by a closing discussion, focusing on possible countermeasures regarding the attack scenarios developed in the game. The game is scientifically validated (Hofmeier & Lechner, 2021) and designed to create plausible insider roles with intention, motivation, and neutralization (Sykes & Matza, 1957) combined with according attack paths. It also features plausibility checks (e.g., through team rating criteria and following external validation) of roles and scenarios. In ten game performances from May 2020 to July 2022 with participants from research institutions, companies, and public authorities, 40 attack scenarios – with at least medium plausibility each – were developed (table 2). 2.2 Analysis Scheme To develop types that are as complete as possible, our aim was to look at as many aspects of malicious insider threats as possible. Therefore, we prepared an analysis scheme based on various existing taxonomies (see table 3), which partly follows the comprehensive taxonomy by Homoliak et al. (2019), who examined and combined various existing taxonomies. We grouped the different type aspects in six groups of characteristics: intention(s) or outcome, motivation, insider position, attack vector, timing, and psychosocial characteristics. 2.3 Analytical Procedure Using this scheme as a two-dimensional category system following the qualitative content analysis methodology according to Mayring (2008), we analyzed the material in two iterations. In the first iteration, the cases (from the interviews as wel as from the game results) were examined one by one. Each considered case was compared with the already extracted types. In case of similarity, it was assigned to the respective type, which is sometimes also accompanied by slight changes to the type (e.g. adding information to a category); in case of contrast with the existing types, a new type with the characteristics of the given case was created. In the second iteration, the cases were examined again by another researcher and in a different order to test whether each case had an adequate counterpart in the types. Among other things, this should prevent a sequence effect as wel as minimize the effect of subjectivity. M. Hofmeier, I. Haunschild, U. Lechner: Malicious Insider Threat Types – An Empirical Analysis 129 Table 3: Taxonomy-based analysis scheme Type Aspects Source(s) characteristics Intentions / theft (of IP), sabotage, fraud, miscel aneous Cappel i et al. (2012) outcome espionage, IP theft, unauthorized disclosure, MITRE (2022) sabotage, fraud, workplace violence information theft, harm to orgnization, ENISA (2020) personal gain Motivation levels: self-motivated, planted, recruited Cole & Ring (2005) motivation: financial, political, personal Cole & Ring (2005) economic-criminological disorder, crisis, BKA (2020) opportunity, dependency Insider topology: pure insider, inside associate, inside Cole & Ring (2005) position affiliate, outside affiliate system role: system masters, advanced users, Magklaras & Furnell application users, none (2001) physical access to control systems, privileged BSI (2018) users, third-party employees Attack vector manifestation level: physical world, network, Phyo & Furnell operating system, application, data (2004) Timing duration: long-term, short-term, one-time Homoliak et al. (2019) point in time: before job termination, after Homoliak et al. job termination (2019) Psychosocial levels: psychological, social, socioeconomic - characteristics indicators: disregard for authority, Greitzer et al. (2013) disgruntlement, anger management issues, confrontational behavior, disengagement, not accepting criticism, absenteism, self- centeredness, performance, lack of dependability, personal issues, stress Personal vulnerabilities: introversion, social Shaw et al. (1998) and personal frustrations, computer dependency, ethical "flexibility", reduced loyalty, entitlement, lack of empathy 130 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3 Malicious Insider Threat Types Based on the analysis of the interview results as wel as the attack scenarios from the game, we identified eleven insider threat types. Below, the individual types are described according to the taxonomy-based analysis scheme. 3.1 Disgruntled Employee / Disgruntled Leaver The disgruntled employee is motivated primarily by revenge and seeking “justice”. This type is discuss in may insider threat studies – e.g., in the Insider Threat Study (Keeney et al., 2005). The wish for revenge can be based on different causes, such as being fired, having disputes with supervisors or management, a lack of recognition, unequal salary structures, or missed salary increases, which cause a - subjective - feeling of injustice. Here, we also discuss the disgruntled leaver as a subtype, which has the same characteristics except that it related to a job termination (either by the organization or the employee resulting from a perception of injustice). This insider type usually seeks to harm the organization, typical y by sabotaging or causing reputational damage through physical sabotage (e.g., of products or machines) or data manipulation. A disgruntled employee is natural y an insider who has a contract with the organization, while he or she can have any position in any department within the organization. An incident is typically a one-time event at any time during employment or after job termination (disgruntled leaver). 3.2 Data Transfer to Competition In this threat type, an insider takes data such as intel ectual property (e.g., source code, blueprints) or other valuable information (e.g., customer data) with him or her when transferring to a competitor. The act is naturally a one-time event during job termination and typical y based on an aim for personal gain. Still, it can also be additionally motivated by revenge or curiosity. Insiders of this type can be pure insiders or external contractors but are limited to persons with access to valuable data (e.g., files, applications, databases). 3.3 Industry Espionage M. Hofmeier, I. Haunschild, U. Lechner: Malicious Insider Threat Types – An Empirical Analysis 131 Insiders engaging in industrial espionage use given privileges to steal valuable information such as intel ectual property while having any position with access to data (internal or external positions). They are typical y recruited by external actors, typical y using bribes. The wil ingness to accept bribes can be mediated by factors such as job satisfaction or financial situation. The act is usually a one-time event at any time during association with the organization. 3.4 State Espionage Similar to the industry espionage type, insiders engaging in state espionage use given privileges to steal critical information while having any position with access to data. They are recruited by external agents, are bribed or blackmailed (e.g., after being compromised privately). The wil ingness to accept bribes or give in to blackmail can be mediated by additional factors such as job satisfaction, financial situation, or family circumstances. As with the industry espionage type, the act is usual y a one-time event at any time. 3.5 Taking Advantage of Privileges for Personal Gain When insiders take advantage of privileges to achieve personal gain, the motivations are manifold: in some cases, there are financial problems (e.g., through having too high lifestyle standards or depts), and others subjectively think they deserve the profit (which is not granted to them by the organization), and some are envious (e.g., of customers when it comes to expensive products or services). To achieve personal gain, these insiders either simply steal property (e.g., IT equipment) or use their access to IT systems to commit fraud. The actions (theft or fraud) often start smal and then become more in case of success until the point where the actions become conspicuous. So, it is typical y a series of events. 3.6 Unauthorized Inspection of Personal Data Individuals who unauthorizedly inspect personal data are users or admins - typically pure insiders - with insights into databases (either directly or via applications). The act is usually short-term and motivated either by personal reasons (such as curiosity or having a crush on a person residing in the database) or by political agenda (e.g., looking up personal data of perceived enemies). 132 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3.7 Intel ectual Property Sale When intel ectual property is sold on black marketplaces (e.g., on the darknet), the perpetrators are usually people with technical skills, such as administrators or developers, and thus typical y pure insiders. In this case, intel ectual property can mean blueprints of products or source code of software products. When it comes to leaked source code, the risk of further cybersecurity incidents is also enhanced. The motivation resides in personal (financial) gain and opportunity through access to valuable data (e.g., repositories), supported by low commitment and ethical flexibility. 3.8 Whistleblowers Whistleblowers follow their political or moral understanding and seek to publish critical data of the organization (e.g., data that are evidence immoral in the eyes of the perpetrator). These individuals identify themselves as whistleblowers - even if they aren't (according to the perceptions of the general society). They feel the organization deserves the harm and to act for the greater good. These views correspond with the neutralization techniques "denial of the victim" and "appeal to higher loyalties" (Sykes & Matza, 1957). In principle, this type of perpetrator can be anyone with access to relevant data, including third parties. In some cases, the insider might be encouraged by (external) parties. The act is usually a one-time event during affiliation with the organization. 3.9 Politically Motivated Sabotage As with the whistleblower type, politically motivated saboteurs think that the victimized organization deserves the harm and that they act for the greater good (neutralization techniques "denial of victim" and "appeal to higher loyalties" (Sykes & Matza, 1957). But these individuals seek to actively harm the organization by (physical or digital) sabotage or causing reputational damage (e.g., by publishing compromising information) because they reject the organization. This primarily applies to public sector organizations or companies where the morality of the objectives or methods is in question. Perpetrators of this type can have any position in the organization, while the attack vector they choose depends on their position and the corresponding skil s and opportunities. M. Hofmeier, I. Haunschild, U. Lechner: Malicious Insider Threat Types – An Empirical Analysis 133 3.10 Extortion In the extortion type, insiders blackmail the organization with the help of their physical or digital access privileges. The target can be data (e.g., using ransomware) or physical goods (e.g., poisoning food in the supply chain). These perpetrators are typically financially motivated, which can reside in financial problems. They also might have a criminal background or psychological issues. In general, they have low job commitment and might misjudge the consequences of their actions. They can have any position in the organization, while the attack vector they choose depends on their position and the corresponding skil s and opportunities. 3.11 Il egal Use of IT Infrastructure In this type, individuals use the organization's IT infrastructure for their own illegal purposes for a long time, such as using storage devices for storing and providing illegal material or using infrastructure for illegal transactions (e.g., money laundering). Natural y, this type of perpetrator requires information technology skil s and access to IT infrastructure. IT staff such as system administrators are typical for this insider threat actor. 4 Conclusion We developed an analysis scheme from existing taxonomies and typologies for extracting malicious insider threat archetypes. We identified eleven types of malicious insider threats using this scheme in an empirical analysis of real-world and fictional but realistic (and plausible) adversarial insider roles and attack scenarios. These types consider multiple facets and thus exceed existing typologies. However, there are some limitations. The types are based on the analyzed material and the subjectivity of an analyzing researcher. Therefore, the typology may not be complete, and new types may be identified by analyzing additional material using the developed analysis scheme. 134 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY The identified types are important for risk analysis in practice (e.g., scenario-based risk assessment) and a basis for further research. In addition, we found that a creative approach (here in the form of a serious game) could extend the results, particularly regarding intentions and motivations. In comparison with existing taxonomies and typologies, the typology presented in this paper features more dimensions (especial y taxonomies are typical y one-dimensional) and more types (e.g., politically motivated insiders are missing in other typologies). This way, our work enables a better understanding of malicious insider threats, a better understanding of the variety of malicious insider threats, more awareness of insider threats, and a better fit of security measures to the insider threat variety. Acknowledgements This work originates from the LIONS research project. LIONS is funded by dtec.bw – Digitalization and Technology Research Center of the Bundeswehr, which we grateful y acknowledge. dtec.bw is funded by the European Union – NextGenerationEU. We would like to thank all game participants, all interview partners, and the LIONS project partners for fruitful discussions. References Bundesamt für Sicherheit in der Informationstechnik (2018). Industrial Control System Security – Innentäter. https://www.bsi.bund.de/SharedDocs/Downloads/Webs/ACS/DE/BSI- CS/BSI- CS_061.html. Bundeskriminalamt (2020). Monitoringbericht Innentäter in Unternehmen 2: Aktuelle inländische Forschungsbeiträge, wesentliche Ergebnisse und Handlungsempfehlungen. Technical Report. Cappel i, D., Moore, A., Trzeciak, R. (2012). The CERT Guide to Insider Threats: How to Prevent, Detect, and Respond to Information Technology Crimes (Theft, Sabotage, Fraud). Addison Weseley. Cole, E., Ring, S. (2005). Insider Threat: Protecting the Enterprise From Sabotage, Spying, and Theft. Syngress. ENISA (2020). ENISA Threat Landscape 2020 - Insider Threat. Technical Report. ENISA. Greitzer, F. L., Kangas, L. J., Noonan, C. F., Brown, C. R., Ferryman, T. (2013). Psychosocial Modeling of Insider Threat Risk Based on Behavioral and Word Use Analysis. e-Service Journal 9, 1. Hevner, A. R., March, S. T., Park, J., Ram, S. (2004). Design Science in Information Systems Research. Design Science in IS Research MIS Quarterly, 28(1), 75–105. Hofmeier, M. (2020). Operation Digital Ant. https://github.com/NutriSafe-DLT/operation-digital-ant. Hofmeier, M. (2021). Operation Digital Butterfly. https://github.com/LIONS-DLT/operation-digital-butterfly. Hofmeier, M., Lechner, U. (2021). Operation Digital Ant - A Serious Game Approach to Col ect Insider Threat Scenarios and Raise Awareness. European Interdisciplinary Cybersecurity Conference (EICC). ACM, New York. Homoliak, I., Toffalini, F., Guarnizo, J., Elovici, Y., Ochoa, M. (2019). Insight Into Insiders and IT: A Survey of Insider Threat Taxonomies, Analysis, Modeling, and Countermeasures. Comput. Surveys 52, 2. M. Hofmeier, I. Haunschild, U. Lechner: Malicious Insider Threat Types – An Empirical Analysis 135 Keeney, M., Kowalski, E., Cappeli, D., Moore, A., Shimeal, T., Rogers, S. (2005). Insider Threat Study: Computer System Sabotage in Critical Infrastructure Sectors. Technical Report. Carnegie Mellon University. Magklaras, G. B., Furnell, S. M. (2001). Insider Threat Prediction Tool: Evaluating the Probability of IT Misuse. Computers and Security 21, 1, 62–73. Mayring, P. (2008). Qualitative Inhaltsanalyse: Grundlagen und Techniken. 10th Ed., Weinheim/Basel. MITRE (2022). MITRE’s Human-Focused Insider Threat Types. https: //insiderthreat.mitre.org/insider- types. Moody, G. D., Siponen, M., Pahnila, S. (2018). Toward a Unified Model of Information Security Policy Compliance, MIS Quarterly, 42(1), pp. 285-311. Phyo, A. H., Furnell, S. (2004). A detection-oriented classification of insider IT misuse. Proceedings of the 3rd Security Conference. Shaw, E. D., Ruby, K., Post, J. (1998). The Insider Threat to Information Systems: The Psychology of the Dangerous Insider. Security Awareness Bul etin 2, 2–98. Sykes, G. M., Matza, D. (1957). Techniques of Neutralization: A Theory of Delinquency. American Sociological Review 22, 6, 664–670. 136 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY QUALITY DIMENSIONS FOR DIGITAL TWIN MATURITY IN THE CONTEXT OF DUTCH PUBLIC SPATIAL PLANNING KOEN SMIT,1 ROB PETERS,2 CHAIM DE GELDER,1 JOHAN VERSENDAAL1 1 HU University of Applied Sciences Utrecht, Digital Ethics, Ultrecht, Netherland koen.smit@hu.nl, chaim.degelder@hu.nl, johan.versendaal@hu.nl 2 Province of Utrecht, Ultrecht, Netherland rob.peters@provincie-utrecht.nl Digital Twin (DT) technology is being increasingly adopted by local and regional governmental organizations in the Netherlands to support spatial planning decision-making, balancing contradicting policy ambitions. These DT solutions are becoming more complex as more perspectives, based on sophisticated calculations, are added to one integrated view of the problem space. To be able to validate the quality of these DT solutions, quality dimensions are to be established. This study Keywords: digital focuses on identifying quality dimensions for DT solutions, twin, including legal and policy perspectives, and safeguarding public maturity value; thereby transcending the technical focus that most model, quality contributions on quality dimensions for DTs have in the current dimensions, body of knowledge. Based on empirical data collection and dutch analysis, 15 quality dimensions were identified. Future research spatial planning, should focus on further operationalizing these dimensions, public allowing for measuring DT solution quality on a maturity scale. values DOI https://doi.org/10.18690/um.fov.4.2023.9 ISBN 978-961-286-751-5 138 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction DT technology is gaining attention in academia and industry, particularly in the manufacturing and asset management domains (Uhlenkamp et al., 2022), as well as in healthcare (Fuller et al., 2020) and government (Peters et al., 2022). While there is no consensus on a standardized definition of a DT, most research views it as a cyber-physical system that shares the concepts of a physical entity, a virtual model, and connections between both (Liu et al., 2022). This study adheres to the definition of a DT as "a set of linked operation data artefacts and (simulation) models, which are of suitable granularity for their intended purpose and stakeholders and evolve throughout the product life-cycle" (Boschert & Rosen, 2016). However, this definition does not ful y comply with a true DT that should support bi-directional communication (Fuller et al., 2020; Liu et al., 2022). The government sector is using DT solutions more frequently to facilitate stakeholder discussions on operational asset management and permit processing for citizens during the spatial planning process (Marcucci et al., 2020; Peters et al., 2022; Wan et al., 2019). However, the growing use of DT solutions raises concerns about their design, development, implementation, and impact on society, as wel as their role in the democratic-legislative arena (Peters et al., 2022). Therefore, it is important to consider public values, particularly in the context of the ongoing climate discussion. An increasing number of municipalities, provinces, and national governmental agencies in the Netherlands are experimenting with DT solutions to support spatial planning. However, these experiments are mostly isolated, and governmental organizations aim to collaborate to increase the effectiveness of their DT solutions by exchanging relevant DT models, such as heat stress, flood risk, nitrogen emissions, biodiversity, or land/zoning usage (CROW, 2022). This col aboration reduces redundant creation of DT models by scarce Subject Matter Experts (SMEs) who are proficient in both spatial planning and data management and science. Uniformity is also required for larger-scale benchmarking using monitors such as emission, air quality, or water quality monitors to measure policy effects. K. Smit, R. Peters, C. de Gelder, J. Versendaal: Quality Dimensions for Digital Twin Maturity in the Context of Dutch Public Spatial Planning 139 Governmental organizations aim to increase the effectiveness of DT solutions by combining different DT models for variant analysis, simulation, and effect monitoring. Each DT model represents one or more measurable indicators related to specific policy ambitions, which may conflict with each other, such as noise level DT models relating to both mobility and quality of life (Deng et al., 2021; Ketzler et al., 2020; Schrotter & Hürzeler, 2020). Balancing these indicators in 3D spatial planning can better support scenario building and decision-making compared to printed 2D maps. The quality of DT models that curate policy ambitions into validated visualizations is the focus of this paper. To ensure the integrity of the digitalization process of policy ambitions, the Province of Utrecht is looking for ways to establish the quality of DT models made by colleagues from other Dutch governmental agencies and third parties involved in building real-estate projects. To the knowledge of the authors no work on quality dimensions for DT models from an integrated perspective, i.e., integrating technical, legal, and public values, exists in the current body of scientific and practical knowledge. Also, it is important to include the cultural and contextual aspects in constructing quality dimensions for DT models, that, to the knowledge of the authors, has not been done before. Therefore, provided a Dutch context, our research question in this paper is as follows: ‘ What quality dimensions for establishing DT maturity level for spatial planning are deemed important in the context of the Dutch governmental domain and the preservation of public values? ’ Because there is not much research on this topic yet, we intend to col ect data for answering from observations of professionals in practice, operationalizing an empirical point-of-view. The remainder of this paper is structured as follows. In section two, the background and related work are presented. This is followed by our research method in section three. Then, the results of our analysis are presented in section four. Next, section five discusses the limitations of this study with future research directions, which is followed by the conclusions in section six. 2 Background and Related Work Dutch governmental agencies responsible for spatial planning, such as provinces and municipalities, use steering mechanisms to execute law and regulations for implementations and policies for localization in specific areas of planning. The 140 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY related cyclic process is often referred to as the policy lifecycle (CROW, 2022) and comprises four phases, being: 1) Policy development, 2) Policy execution, 3) Policy enforcement, and 4) Policy monitoring. The policy development phase comprises the creation of spatial planning ambitions as wel as more detailed spatial plans based on measurable parameters and calculation of effects. It is important to note that DT solutions are increasingly being used in this phase to assist in determining a proper trade-off between policy ambitions represented by indicators to create spatial plans for a given area in a Dutch city or region (Future Insight, 2023; Gemeente Rotterdam, 2023; Vereniging van Nederlandse Gemeenten, 2022; Witsenburg, 2020). The second phase focuses on the implementation of the policies, e.g., so that concrete ambitions and goals are applied to planning areas in city programmes. During the third phase, governmental services are provided such as online permit checkers for civilians and organisations that need to comply with these ambitions and specific goals. This third phase also entails the enforcement of policies concerning the actual physical situation where the Dutch government has the responsibility to ensure that individuals or organizations act according to the implemented policies, e.g., do not cut down protected tree species or extend buildings without official permits. Lastly, the fourth phase focuses on monitoring the extent to which policy goals and ambitions are achieved, which is an important fundament for the continuous improvement of the policies implemented. For this study, we are looking for relevant quality dimensions that assist Dutch governmental organizations in determining the quality of their DT models throughout the whole policy lifecycle. Yet, DT models are, at this moment in time, generally used in the first phase; Policy development, by civil decision-makers from Dutch provinces and municipalities such as Amsterdam, Rotterdam, Utrecht & Amersfoort. A recent study that was performed on a large number of contributions with a broad focus on available DT Maturity Models (DTMMs) (77) revealed 31 dimensions across seven categories, albeit with a focus on production and logistics, see figure 6 in the work of (Uhlenkamp et al., 2022). We look at DTMMs as they are a good starting point and might contain relevant quality dimensions for DT models. In the context of this study, we refer to digital twin maturity levels as a hierarchical framework that describes the degree of advancement and integration of a digital twin into a system or process (Madni et al., 2019; Shahzad et al., 2022; Uhlenkamp et al., 2022). The analysis presented by Uhlenkamp et al., (2022) is recent, thorough and is entirely covering the technical perspective of a DTMM, e.g., levels of cognition, K. Smit, R. Peters, C. de Gelder, J. Versendaal: Quality Dimensions for Digital Twin Maturity in the Context of Dutch Public Spatial Planning 141 model maintenance, and computing capabilities, while largely lacking the consideration of explicit legal, policy, and preserving public values and the FAIR translation of indicators into the DTs. Results from our own narrative style literature review (Paré et al., 2015) reveal a similar direction, where contributions on DTMMs seem to predominantly focus on the technical capabilities of a DT and its underlying DT models, e.g., in the work of (Enders & Hoßbach, 2019; Fuller et al., 2020; Lim et al., 2020; Rasheed et al., 2020). Also, one should take into account that many DTMMs available in the current body of knowledge are designed to be either specifically tailored for a domain, e.g., manufacturing, healthcare or aerospace or designed to be universal y applicable (also referred to as cross-domain), see table 3 in the work of (Uhlenkamp et al., 2022). From a legal and policy point of view, it is important that the DT model is validated and holds when a decision made by spatial planners using a DT solution is disputed by individuals or organizations in court or permit processes. Chal enges concerning this decision-making process are revealed to be numerous (Peters et al., 2021, 2022) and given the importance of legal y-grounded technology (Peters et al., 2022), which represents important public values such as transparency, safety or fairness, the legal aspect should be taken into consideration when determining the value of DT model integration. For the successful implementation of DT solutions, acceptance by stakeholders in the context the DT is applied is deemed ''extremely critical'' (Rasheed et al., 2020). This involves trust from the spatial planners, decision-makers, legal advisors as well as the individuals impacted by the decisions made based on the DT models in the DT solution. Trust in the DT solution is the umbrel a term here, but it comprises a combination of many public values (Voas et al., 2021) represented by the DT solution, which also differs per stakeholder group. Because of this pluriformity, it is not sensible to explore al of these integral y in this study, at least not at this stage. As the maturity of the research domain is low, this stage should focus on identifying constructs and their relations (Edmondson & McManus, 2007). Therefore, we focus on exploring the concept of public value for policymakers and decision-makers in a legal and policy arena where disputes about climate goals, housing and SDGs are becoming increasingly severe and digital means are becoming part of the political arena. Public values are those providing normative consensus about 1) the rights, benefits, and prerogatives to which citizens should (and should not) be entitled, 2) 142 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY the obligations of citizens to society, the state and one another; and 3) the principles on which governments and policies should be based (Bozeman, 2007). As can be derived from the above definition, the normative consensus on the mentioned aspects in itself represent values relevant for individuals, e.g., privacy, equality, democracy, integrity, and honesty. A set of validated principles commonly used in the combination of ICT, the governmental domain and the use of DT solutions is FAIR. The FAIR framework is an abbreviation for 1) Findability, 2) Accessibility, 3) Interoperability and 4) Reusability (GoFAIR.org, 2016). It is used as a framework by different Dutch governmental agencies guiding the use of technology, thus also covering the use of DT solutions. Because of this we utilize the FAIR framework as a central framework to identify relevant public values to determine DT model quality, and with this its maturity. 3 Research Method The main aim of this study is to identify quality dimensions for determining DT maturity in the Dutch governmental domain. The maturity of research in this area is considered low, particularly for DT maturity research focusing on non-technical capabilities. Therefore, an exploratory and inductive approach is chosen (Edmondson & McManus, 2007) to extract relevant quality dimensions from practice, as an increasing number of Dutch provinces and municipalities are using DT models in the policy development phase. 3.1 Data Collection We focused on exploring the opinions of SMEs concerning two aspects: 1) which quality dimensions are relevant in establishing DT maturity? And 2) what criteria are relevant in the context of the quality dimensions identified? For this study, secondary data was used, originating from the end of 2020 until the end of 2022. In total, thirteen SME sessions with an average duration of 1,5 hours were included in the dataset, which were al in Dutch. In these digital sessions, Dutch SMEs discussed DT cases that were developed and applied in the Netherlands, predominantly in the context of the Dutch government. The SMEs were mostly employed at Dutch governmental agencies, though the expert sessions also hosted several other interested participants such as researchers, semi-governmental experts and commercial DT model suppliers. In total, 15 hours of video footage was col ected. K. Smit, R. Peters, C. de Gelder, J. Versendaal: Quality Dimensions for Digital Twin Maturity in the Context of Dutch Public Spatial Planning 143 The sessions were organized in a presentation style setting in which two to four presenters digital y presented to a crowd after which Q&A was organized and discussion about the DT model ensued. In total, an average of 45 participants were present during each of the sessions. Not al participants could interact with the speakers given the large number of participants and limited time for each expert session, however. In addition to the expert sessions, two documents were coded. These official documents were produced with and by the provincial council of Utrecht together with other SMEs focused on how to further develop their DT models to support spatial planning. 3.2 Data Analysis The data was analyzed using a thematic coding approach. Three coding cycles similar to Strauss and Corbin’s (1990) process of 1) open coding, 2) axial coding, and 3) selective coding was used. The first round, open coding, involved analyzing significant participant statements. In this process, we tried to identify what Boyatzis (1998) refers to as “codable observations”. Thus, we coded the data by identifying statements (both written and in audio) that discussed quality dimensions for DT models. For example, codable observations were: “Most suppliers of Digital Twin platforms and models are built and of ered in a closed sourcing agreement. ”, and “Depending on the policy phase the model supports, you need static or dynamic data. ” Synonyms as wel as redundant codes were al registered to ensure completeness. This process was conducted in several steps of which the first step was to ensure high intercoder reliability among the three research team members. Three researchers coded the first expert session separately from each other during the first step. This resulted in a similarity of 0.759 using the Krippendorfs Alpha coefficient, which is acceptable given the fact that the study is explorative (Lombard et al., 2004; Yeaton & Wortman, 1993). Because of this, two coders proceeded with distributing the data and coding them separately from each other. Additional y, the two coders compared one more random expert session from the data together to compare the results to see whether coder reliability was maintained, which was the case. Subsequently, we conducted axial coding. In this round, we categorized the open coding results so that quality dimensions can be identified. For example, we coded the following five codes under the quality dimension Visualisation capabilities: ' Visualisation of real-time data is bet er when done in symbols because textual information is 144 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY rather slow.' , 'Visualisation of fictive set ings must be possible next to realistic visualisation because of the closeness for the af ected user.' , ' Visualisation in a virtual reality' , ' One must deliberately choose which data to visualise and which data to omit depending on each stakeholder involved' , and ' Changing between 2-D and 3-D map layers' . This process was conducted by two researchers and resulted in 20 Digital Twin Quality Dimensions (DTQDs). Lastly, we conducted selective coding to normalize the identified categories as wel as to re-assess codes that could not be attributed to any of the existing DTMMs, in which approximately 50 codes could be assigned given the discussion involving the context in which a statement has been made. The normalization was discussed between the two researchers involved in the coding during rounds one and two. This process resulted in the recategorization of five DTQDs into existing ones. For example, the quality dimension 'Level of detail capabilities' has been merged with the quality dimension 'Visualisation capabilities' because, when analysing the context of the statements and semantics of the codes we learned that it concerns a visualisation criterion and does not warrant a separate quality dimension. In total, 103 codes were deemed not suitable enough to be included as supporting existing or new DTQDs after three rounds of thematic coding. The results of this stepwise process are presented in the next section. 4 Results Based on our analysis we identified 15 DT quality dimensions, which are further detailed in this section. The dimensions are sorted by the number of codes identified. Furthermore, we clustered the dimensions using 1) Technical, 2) Governance, or 3) legal labels. Figure 1 provides a visual summary of the identified DTQDs. Figure 1: Identified DTQDs K. Smit, R. Peters, C. de Gelder, J. Versendaal: Quality Dimensions for Digital Twin Maturity in the Context of Dutch Public Spatial Planning 145 Governance - DT model application area(s) and goal(s) (220 codes) The model in a DT for a certain area in a city or region consists of policy themes. The DT model itself serves the goal of balancing the policy themes and ambitions and covers a perspective, based on one or multiple data sources, that constitute indicators and their mutual interaction in business rules and the foundation for these rules, which must be made explicit. Examples of thematic indicators are air quality, road infrastructure, soil structure, elevation level, water storage, biodiversity, light emission, energy usage, social participation, public transport mobility, economic building function, etc. Technical - Visualisation capabilities (149 codes) Because DT models enable the visualisation of multiple integrated perspectives, knowledge elicitation and communication become more effective and efficient. SMEs state that, although not always required given the type of data, 3-D visualisation helps manage complexity for all stakeholders. Part of that complexity is the inherent capability of adjacency, where sound affects health, given a specific distance between source and housing quarters, for example. Many calculations among different indicators are set as regulations and permit requirements. Also, being able to work with different levels of detail is deemed important. Effective visualisation when integrating data in one view includes appropriate colouring, layers, dashboards and graphs. Also, augmented and virtual reality technology are mentioned as contributors towards the quality of visualisation by SMEs. An example of this is where 3-D above and below-ground data is integrated into one view using different colour markings for types of infrastructure. Using this visualisation has lowered the error margin for building accidents by 3-5%. Technical – DT model, data, algorithm openness capabilities (141 codes) Often mentioned by SMEs is that DT models and data should be openly maintained and shared. This increases transparency towards and auditability by stakeholders. Openness is a multi-faceted concept that could relate to the data being publicly available, availability of the calculation algorithm, open sensors to facilitate citizen science, re-use of data and DT models, etc. Other important criteria mentioned were, in a technical sense, the quality of the data and how data could be safely shared. Also, SMEs want to be able to modify and further develop third-party DT models 146 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY themselves after implementing them. For example, SMEs stated that most DT models currently being released are closed source, not enabling proper col aboration, which is deemed inefficient especial y because Dutch governmental organizations should collaborate to achieve targets and policy ambitions at higher administration levels (i.e., ministries) as well as to meet with public expectations. Technical - Simulation and prediction capabilities (120 codes) One of the core additions to the spatial planning policy lifecycle using DTs is its ability to simulate scenarios and predict outcomes. SMEs state that civil servants rely on simulating effects, probability calculations, stress testing, impact-assessment, etc. It is also important for them to be able to work with KPIs and thresholds so that the effectiveness of the simulation of the scenario can be established. Underlying the simulation and prediction is the promise of increasingly more effective and efficient utilization of artificial intel igence. An example mentioned is the ability to predict water damage in a city when riverbanks are flooded during storms, where mobility, demographics and weather data are included in the predictions. Technical - Infrastructure capabilities (104 codes) Using (a combination of) DT model(s) with large datasets in a 3-D environment requires proper infrastructure. SMEs urge that the DT platform and its infrastructure should support large sensor networks and APIs to interface web-based services and datasets used in the DT models. The amount of processing power of a DT model is also mentioned as an important criterion by SMEs because current DT models do not seem to scale appropriately. There is often a trade-off between the level of detail and the size of the area covered in the DT. The larger the size, the harder it becomes to maintain the same level of detail for that area. Also, depending on which policy lifecycle phase must be supported by the DT model, either static or dynamic data must be used. The update frequency of data is high when policies are monitored using automated sensor networks creating a near real-time datastream, requiring different resources compared to one-time-only imports of historical data. An example of an infrastructure choice is about how to provide the DT model to stakeholders, either web-based or in a client-server (local) configuration. This choice impacts the required infrastructure capabilities for governmental bodies. K. Smit, R. Peters, C. de Gelder, J. Versendaal: Quality Dimensions for Digital Twin Maturity in the Context of Dutch Public Spatial Planning 147 Technical - Perspective integration capabilities (100 codes) A (3D) DT platform offers the possibility to combine multiple perspectives in one overview. SMEs state that spatial planning complexity becomes more transparent once multiple layers of data are combined, on a technical level. This in turn improves their ability to consider different indicators as wel as to decrease error margins created by misalignment between planning and execution. For example, by combining noise, green and mobility data, one can use an assessment framework to balance a decrease in quality of life for occupants living near busy roads, keeping city segments accessible for public transport, maintaining biodiversity, and reducing heat-stress in the city, amongst other policy ambitions. SMEs stress that combining physical and social data, amongst others, is very important for efficient and effective spatial planning policy design. Legal - DT model/data legal integrity (93 codes) Using DT technology to support policy lifecycle management in spatial planning requires the DT model and data to be reliable and validated. SMEs stress that the DT model and its underlying data must have the legal power to hold up in court, or else it cannot be used effectively in spatial planning. Legal foundation is also multifaceted because one could refer to the sensors used, the dataset itself or the algorithm used to represent an indicator, e.g., nitrogen pollution. For example, some SMEs stated that DT models published by scientific national knowledge institutes are preferred over, e.g., commercial third-party organizations or from citizen science, while others pose that such data is just as sound. Legal – public value by explainability (65 codes) Each DT model inherently represents a combination of ambitions with indicators. SMEs stated that these indicators should be made explicit so that the users of the DT model can use the DT model in a way that these indicators are al considered in an assessment framework that is used to ground spatial planning decisions. For example, SMEs find that the DT models and underlying ambitions and indicators considered should be explicitly communicated and made explainable when presenting spatial planning policy decisions. Such an approach also forces third- 148 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY party/commercial DT model developers to be transparent about indicators served by their DT models as well as to take public values into account by design. Governance – Governance and col aboration (63 codes) Governance proved to be multi-faceted, however, is deemed an important cornerstone for SMEs. It is important that, when collaborating with other organizations, governmental bodies explicitly manage ownership of DT platforms, DT models and data. Also, because of the cal for standardization, the governance around the standardised components must be clear. Furthermore, governmental bodies aim to collaborate in a quadruple helix setting, which adds to the complexity of governance around DT models. For example, when a standardised federative national DT platform would enforce a standardised approach, it must be clear what local autonomy the governmental bodies have in building their own DT models. Technical - Interaction capabilities (59 codes) In addition to the visualisation capabilities of a DT model, SMEs also stressed the importance of interaction capabilities. Because multiple perspectives are combined, SMES find it useful to interact with the data so that layers of data can be enabled or disabled. Changing levels of detail by zooming or selecting specific areas on maps are examples of how interaction helps stakeholders utilize the data to support scenario building and simulations. SMEs also mentioned augmented and virtual reality, e.g., with a hololens, as ways to more realistically interact with the DT model. Other examples of using these technologies are walking, driving or flying through the DT. Legal - Stakeholder participation and expectation management (56 codes) Different stakeholders are involved during the policy lifecycle. SMEs state the importance of explicitly taking into account when each type of stakeholder is involved and to what extent. For example, involving citizens during the policy design phase provides them with the opportunity to influence the design of their neighbourhood, while it could also be the case that citizens are merely informed about the plans using the DT at a stage in which they could not influence the design K. Smit, R. Peters, C. de Gelder, J. Versendaal: Quality Dimensions for Digital Twin Maturity in the Context of Dutch Public Spatial Planning 149 at al. The way participation is perceived is also based on a few other criteria such as the level of detail, the fictiveness of the visualisation, and what data is presented. Technical - Standardization capabilities (34 codes) One dimension that is similar yet distinct from the openness of a DT model is that of standardisation. SMEs stated that for them to be able to effectively collaborate, DT models and data should be interchangeable as wel as their platforms be interoperable as much as possible. SMEs expressed their needs toward a national standard or principles, e.g., using an Object Type Library, reference architecture, and definition libraries. Using a more standardized approach enables importing DT models, standard building blocks, and more plug-and-play working using DT platforms. For example, because of the current lack of standardisation, DT models between most Dutch cities are not interchangeable, hampering collaboration between them. Governance - Knowledge and competencies (16 codes) When DT models are introduced, their integration within the DT platform and interpretation of the data must be secured. SMEs declared that to be able to do so, people must have the right competencies. Also, knowledge about the DT platform and models must be retained by the governmental bodies using the DT for spatial planning, as employees leaving organizations causes a knowledge drain. To guarantee proper competency development and knowledge management, SMEs think it is important to collaborate with universities, domain standardisation agencies, and other partners. Additionally, decision-makers must be trained in using DT technology for spatial planning purposes in a value-sensitive way. Legal – Temporal capabilities (13 codes) Again, using a DT model for spatial planning decision-making requires it to be recognized as a legal foundation. However, decision-making processes commonly take a long time and data changes over time as wel , which requires the DT platform and model to support proper archival capabilities. The moment a decision is made by stakeholders, its context, data, and algorithm used should be saved so that it can be reproduced when governmental bodies are requested to explain the decision in a 150 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY later stage of the policy lifecycle. For example, SMEs want to be able to explain to citizens what circumstances (and data) were relevant at the time when the decision was made. This is especial y relevant given the increasing level of monitoring of policy effects as part of the legal obligations of governments. There is a need for semantic consistency between policy ambitions and policy effects. Legal - Privacy sensitivity (9 codes) A combination of different datasets could lead to undesirable traceability towards individuals. SMEs stated that this could harm the privacy of individuals, especially when the data is accessible to the public. This risk is higher when data from the social domain is included in the model. An example would be where demographic health data is used to signify differences between streets, which could also be used to trace personal information towards a specific household or individual. 5 Discussion and future research The findings in this study have to be seen in light of some limitations. The first limitation concerns the depth of our empirical exploration of quality dimensions of DT models intending to construct a DTMM. While we established a list of dimensions with possible criteria for each dimension, based on the current data we cannot quantify a level of maturity on these dimensions, yet. Future research should further operationalize our findings and derive possible dependencies between them. The same holds for dependencies across the quality dimensions, which is an important aspect when constructing MMs (Mettler et al., 2010). The second limitation concerns the data that is used to arrive at the DTQDs presented in this paper. We used secondary data as our primary data source. This presents a disadvantage over data primarily collected for this study, because of its potential lack of focus during the SME sessions as wel as that coverage of al relevant DTQDs cannot be guaranteed. While theoretically, this is correct, our data analysis revealed that saturation was reached rather quickly when coding towards DTQDs in round two, meaning that most of the independent discussion amongst SMEs limited itself towards the concepts discussed in this paper. Another advantage of this approach is that it prevents some bias types. Future research should therefore include a larger scale direct approach in which SMEs are questioned about the DTQDs and their importance to determine the weight of the DTQDs at a later stage of development. K. Smit, R. Peters, C. de Gelder, J. Versendaal: Quality Dimensions for Digital Twin Maturity in the Context of Dutch Public Spatial Planning 151 Concerning our coding approach, some limitations could affect the validity of the DTQDs presented in this paper. In total three research team members engaged in coding, of which two primarily coded the data. Future research should focus on increasing the number of independent coders. On multiple occasions, we identified a large consensus amongst the audience on certain statements, which we could not explicitly code due to the online setting of the meetings. Therefore we argue that our codes could have more empirical value than the absolute number of counts in our dataset implies. Whether this is the case should be further explored in future research. 6 Conclusion To conclude this paper we revisit our research question: ‘What quality dimensions for establishing DT maturity level for spatial planning are deemed important in the context of the Dutch governmental domain and the preservation of public values?’ Based on a thorough analysis of secondary data featuring rich debate amongst SMEs in the context of the Dutch government we were able to derive 15 DTQDs along with possible criteria that constitute maturity. There are similarities between our findings and the results presented in the study presented in (Uhlenkamp et al., 2022). These similarities are predominantly focused on the technical dimensions of maturity, while the dimension of human-machine interaction is also present, though described at a high level of abstraction. Interestingly, the legal aspect as discussed in our data is not or barely mentioned in other studies related to DTQDs and should be further explored and operationalized in future research. From a theoretical perspective, this study presents new knowledge about the empirical point-of-view that SMEs have in the Dutch context. It also adds to the knowledge about how the maturity of DT models can be measured as, to the knowledge of the authors, most contributions ful y or predominantly focus on technical aspects of maturity. This study identified explicit empiric evidence concerning, e.g., the legal integrity of DT models, participation capabilities, and explicitation of policy ambitions. The legal integrity dimension points to two areas for further research. On the one hand, it is observed that effect monitoring gradual y becomes a new standard in legislation, which in turn generates a push for consistency between policy design and policy output in the democratic arena. Citizens demand that politicians achieve what they have promised, and DT solutions seem to provide 152 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY the means. On the other hand, we observe a merger between article or text-based legislation regarding those policy goals and the object-based area of application (effect). Further substantiation could support the theory that DT technologies enable such a merger between unstructured text-based administration and more object-oriented spatial data infrastructures as we already have observed with Inspire and Natura 2000 (Peters et al., 2009; Peters, 2016). This Merger implies a different approach towards legal processes from document versioning towards agile software development, This, in turn, requires a completely different approach or culture towards legislation. From a practical perspective, this study contributes towards Dutch governmental bodies about where to measure DT model quality on. Revealing the spectrum of DT model quality enables government agencies as wel as SMEs responsible for integrating DT models in DT platforms to start and guide the discussion with regards to their approach of collaborating with others on integrating increasingly larger (combinations of) DT models underlying the policy lifecycle and spatial planning decision-making process. Acknowledgements We would like to thank Luc de Horde, Jan van Lopik, Maarten van Helden, Michel Grothe, Wouter Heijnen, and all participants for supporting this research. References Boschert, S., & Rosen, R. (2016). Digital Twin—The Simulation Aspect. In Mechatronic Futures (pp. 59–74). Springer International Publishing. https://doi.org/10.1007/978-3-319-32156-1_5 Boyatzis, R. E. (1998). Transforming qualitative information: Thematic analysis and code development. SAGE Publication. Bozeman, B. (2007). Public Values and Public Interest: Counterbalancing Economic Individualism. Georgetown University Press. CROW. (2022). Routekaart Provincies en Datagedreven Assetmanagement - Interprovinciale Digitale Agenda. https://www.ipo.nl/media/533fo4yt/routekaart-en-datagedreven-assetmanagement.pdf Deng, T., Zhang, K., & Shen, Z.-J. (Max). (2021). A systematic review of a digital twin city: A new pattern of urban governance toward smart cities. Journal of Management Science and Engineering, 6(2), 125–134. https://doi.org/10.1016/j.jmse.2021.03.003 Edmondson, A. C., & McManus, S. E. (2007). Methodological Fit in Management Field Research. Academy of Management Review, 32(4), 1246–1264. Enders, M. R., & Hoßbach, N. (2019). Dimensions of digital twin applications-a literature review. AMCIS 2019 Proceedings, 20. Fuller, A., Fan, Z., Day, C., & Barlowc C. (2020). Digital Twin: Enabling Technologies, Challenges and Open Research. IEEE Access, 8, 952–971. K. Smit, R. Peters, C. de Gelder, J. Versendaal: Quality Dimensions for Digital Twin Maturity in the Context of Dutch Public Spatial Planning 153 Future Insight. (2023). Groningen: 3D Digital Twin belangrijk instrument om grote ruimtelijke vraagstukken mee te behandelen. https://futureinsight.nl/nieuws/groningen-3d-digital-twin-belangrijk-instrument-om-grote-ruimtelijke-vraagstukken-mee-te-behandelen/ Gemeente Rotterdam. (2023). Samenwerken aan de 3D Digital Twin. https://www.rotterdam.nl/digitale-stad GoFAIR.org. (2016). FAIR principles. https://www.go-fair.org/fair-principles/ Ketzler, B., Naserentin, V., Latino, F., Zangelidis, C., Thuvander, L., & Logg, A. (2020). Digital Twins for Cities: A State of the Art Review. Built Environment, 46(4), 547–573. https://doi.org/10.2148/benv.46.4.547 Lim, H. K. Y., Zheng, P., & Chen, C. H. (2020). A state-of-the-art survey of digital twin: Techniques, engineering product lifecycle management and business innovation perspectives. Intelligent Manufacturing, 31, 1313–1337. Liu, Y. K., Ong, S. K., & Nee, A. Y. C. (2022). State-of-the-art survey on digital twin implementations. Advances in Manufacturing, 10(1), 1–23. Lombard, M., Snyder-Duch, J., & Bracken, C. (2004). Intercoder Reliability in Content Analysis. Retrieved April, 2002, 1–18. Madni, A. M., Madni, C. C., & Lucero, S. D. (2019). Leveraging digital twin technology in model-based systems engineering. Systems, 7(1). Marcucci, E., Gatta, V., Le Pira, M., Hansson, L., & Bråthen, S. (2020). Digital Twins: A Critical Discussion on Their Potential for Supporting Policy-Making and Planning in Urban Logistics. Sustainability, 12(24), 10623. https://doi.org/10.3390/su122410623 Mettler, T., Rohner, P., & Winter, R. (2010). Towards a Classification of Maturity Models in Information Systems. In Management of the Interconnected World (pp. 333–340). Physica-Verlag HD. https://doi.org/10.1007/978-3-7908-2404-9_39 Paré, G., Trudel, M.-C., Jaana, M., & Kitsiou, S. (2015). Synthesizing information systems knowledge: A typology of literature reviews. Information & Management, 52(2), 183–199. https://doi.org/10.1016/j.im.2014.08.008 Peters, R., Hoekstra, R., van Engers, T., & Hupkes, E. (2009). Legal Simcity; Legislative Maps and Semantic Web Supporting Conflict Resolution. SDI Convergence, 63. Peters, R. M. (2016). The Law, the Map and the Citizen: Designing a legal service infrastructure where rules make sense again. Universiteit van Amsterdam. Peters, R., Smit, K., & Versendaal, J. (2021). Responsible AI and Power: Investigating the System Level Bureaucrat in the Legal Planning Process. 34th Bled EConference. Peters, R., Smit, K., & Versendaal, J. (2022). Validation Challenges for Legal Digital Twins in Dutch Climate Governance. EGOV-CeDEM-EPart, 22. Rasheed, A., San, O., & Kvamsdal, T. (2020). Digital Twin: Values, Chal enges and Enablers From a Modeling Perspective. IEEE Access, 8, 21980–22012. https://doi.org/10.1109/ACCESS.2020.2970143 Schrotter, G., & Hürzeler, C. (2020). The digital twin of the City of Zurich for urban planning. Journal of Photogrammetry, Remote Sensing and Geoinformation Sci-Ence, 88(1), 99–112. Shahzad, M., Shafiq, M. T., Douglas, D., & Kassem, M. (2022). Digital Twins in Built Environments: An Investigation of the Characteristics, Applications, and Challenges. Buildings, 12(2), 120. https://doi.org/10.3390/buildings12020120 Strauss, A., & Corbin, J. M. (1990). Basics of qualitative research: Grounded theory procedures and techniques. Sage Publications, Inc. Uhlenkamp, J. F., Hauge, J. B., Broda, E., Lütjen, M., Freitag, M., & Thoben, K. D. (2022). Digital twins: A maturity model for their classification and evaluation. IEEE Access, 10, 69605–69635. Vereniging van Nederlandse Gemeenten. (2022). Digital twin voor alle gemeenten. https://vng.nl/praktijkvoorbeelden/digital-twin-voor-alle-gemeenten Voas, J., Mell, P., & Piroumian, V. (2021). Considerations for digital twin technology and emerging standards. https://nvlpubs.nist.gov/nistpubs/ir/2021/NIST.IR.8356-draft.pdf Wan, L., Nochta, T., & Schooling, J. M. (2019). Developing a city-level digital twin–propositions and a case study. International Conference on Smart Infrastructure and Construction, 187–194. 154 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Witsenburg, F. (2020). Digital twin helpt bij ontwerp klimaatadaptief stationsgebied Amersfoort. https://www.gebiedsontwikkeling.nu/artikelen/digital-twin-helpt-bij-ontwerp- klimaatadaptief-stationsgebied-amersfoort/ Yeaton, W. H., & Wortman, P. M. (1993). On the Reliability of Meta-Analytic Reviews. Evaluation Review, 17(3), 292–309. https://doi.org/10.1177/0193841X9301700303 TOOLS AND TECHNOLOGIES UTILIZED IN DATA-RELATED POSITIONS: AN EMPIRICAL STUDY OF JOB ADVERTISEMENTS NURLAN MUSAZADE Åbo Akademi University, Faculty of Social Sciences, Business and Economics, and Law (FSEJ), Turku, Finland nurlan.musazade@abo.fi Role, value and amount of data and related tools, technologies, companies and professions in society is rising. Since the required skills for data-related professions are predicted to experience changes, and labor market mismatches create chal enges for stakeholders, this research focuses on changes in the required tools and technologies for data-related positions. The presented research defines trends and changes in frequencies of the tools utilized in the data-related professions by applying quantitative Keywords: content analysis on collected data from job advertisements of data analytics, Finland, Denmark and Poland. The research findings show that data tools used in data-related professions experience significant science, changes over time. For example, AI and cloud computing-related data professions, skills, and SQL started to be required more, whereas Excel, SPSS skill and similar tools are less expected from the candidates. requirements, Furthermore, while R programming language utilization rises in labor market, analytics related positions, Python is more common in positions job related to data science. advertisements DOI https://doi.org/10.18690/um.fov.4.2023.10 ISBN 978-961-286-751-5 156 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Data volumes have an increasing trend influenced by automation, technological shifts and pandemic (Sheng et al., 2020), which makes it crucial to study big data, related tools and professions. Furthermore, organizations have started to utilize big data, related technologies and Artificial Intel igence (AI) more extensively compel ed by pandemic regulations (McKinsey, 2021). As a result, data context and labor market have started to experience significant shifts (McKinsey, 2021). Field experts and related studies indicate that technologies utilized in the workplace for accomplishing tasks, and corresponding required skills, are experiencing changes, and predicted to change in the future on a wider scale (Musazade, 2022, World Economic Forum, 2020). In other words, digitalization drives uncertainty regarding the tools and skills required for professions in the future (Koh, 2020). For example, increasing data volume and unstructured data requires working on new technologies, such as cloud computing (Marr, 2019), and corresponding skills therefore. Some professions, e.g. data entrance clerks, are expected to be completely replaced by emerging technologies (World Economic Forum, 2020). As a result of distribution of some work to machines, the emergence of different professions (e.g. AI specialist) specialized in the new tools and possessing corresponding skil s can be observed (World Economic Forum, 2020). Human resource management in organizations experiences chal enges in recruiting to vacant positions because of the existence of skill gaps (Koh, 2020). Furthermore, data-related professions have common definition deficiency, as they are typical y specified via personal perspectives of different specialists, which increases the knowledge gap (Miller, 2014, Granville, 2014). This is also a problem sporadically addressed in the literature (Granville, 2014). In other words, there is evident need for the study of data-related professions, their role in organizations, as well as related tools that are utilized by them. Another factor that increases the relevance of this study is the rising trend of data companies (Cattaneo et al., 2020). Furthermore, data volume has a continuous rising trend, which generates the need for usage of advanced tools that are capable of processing big data (Power & Heavin, 2017, Sheng et al., 2020, Kenett & Redman, 2019). While there are some existing studies on this topic, they are focused on specific markets with limited data analysis techniques. N. Musazade: Tools and Technologies Utilized in Data-Related Positions: An Empirical Study of Job Advertisements 157 This study concentrates on the Danish, Polish and Finnish labor markets, and aims to answer fol owing research questions: RQ1: What are the most important skil s organizations look for when hiring data professionals? RQ2: How have the skil s required by companies for data-related positions changed in the recent years? The rest of the paper is structured as fol ows. Section 2 presents the literature review on the previous similar research in the domain. In Section 3, the research methodology, data and data analysis are described. Section 4 discusses trends and the findings of the study, and Section 5 summarizes the study, and presents contributions and limitations. 2 Literature Review Data-related positions have been previously analyzed by researchers as listed in Appendix A, typical y by means of empirical data col ected from job advertisements by web scraping and analyzes have been performed with different techniques of text analytics. For example, the study of De Mauro et al. (2018) shows that big data developer (BDD) and engineer (BDE), business analyst (BA) and data scientist (DS) are related professions in the big data domain. Furthermore, research done by Verma et al. (2019) compares data analyst (DA) and DS, and BA and business intel igence (BI) positions, and concludes that a DS profession requires more technical skil s. Yet, in the findings of these two articles there is a mismatch (e.g. in coding skill requirements). Organization, Excel, SQL, Reporting are the skills that have been found required for the DA profession in multiple studies (Verma et al., 2019, Jiang & Chen, 2021). At the same time, Python and R programming languages, Teamwork and Statistics are the common required skills for DS positions (Verma et al., 2019, Jiang & Chen, 2021). For the BA profession, teamwork, management and project management, testing and analytics are the keywords that studies have identified in job advertisements (De Mauro et al., 2018, Verma et al., 2019, Jiang & Chen, 2021). 158 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Persaud (2020) has analyzed big data analytics-related professions with the aim of defining main in demand competencies and provision of students with them. The study has utilized both quantitative and qualitative research methodology, including text mining of job advertisements and education institutions’ courses, and interviewing senior executives. The framework presented consists of cognitive, functional, meta-competence and social competence extensions. Python, SQL, Spark, Hadoop are some of the tools, teamwork, customer relationships, business domain knowledge, communication, presentation skil s the research found to be required from candidates. Moreover, for entry-level vacant positions, candidates are required to have a degree from quantitative field, as wel as experience in relevant positions. Yet, the research presents general findings without detailed description of the frequencies of the skills. In a similar research, Smaldone et al. (2022) analyzed job advertisements of the US market. After data preprocessing and text analysis, skil s have been categorized in thematic groups. Similar to the study of the Persaud (2020), skills have been classified into different groups, whereas in the research of Persaud (2020) the themes have been compared with the particular theoretical framework. Moreover, Smaldone et al. (2022) queried only the presence of “data science” during data collection. Big data, data processing and AI are the top three frequent terms. Topic modeling has resulted in five topics, including generic skills, hard skills, big data and AI, analytical and team-related skills. Python, R, data analysis, deep learning, communication, teamwork, problem-solving, data management and data mining are some of the skills that research has found existing in job advertisements. Fang and Zhou (2021) have analyzed DA positions in China’s market via data col ection from recruitment websites and applying text mining. Data have been collected during 2017. Among the programming languages, Python, R and Java have been found as most utilized in DA positions, whereas authors underline R as the most demanded language. Furthermore, Hadoop, Spark, SQL, Hive, Oracle, SPSS, MATLAB, TABLEAU and SAS are other tools that have been found having high frequency in the job postings, whereas the highest is Excel among the tools that serve for “visualization of statistical analysis” or “application” (i.e. latter four tools) as defined by authors. Yet, soft skills and other domain-related terms are completely missing in the study. N. Musazade: Tools and Technologies Utilized in Data-Related Positions: An Empirical Study of Job Advertisements 159 Another research, in which data have been colected from Chinese websites, is conducted by Cao and Zhang (2021). The search query has been limited to the “data mining” and “data science” keywords. In the research named entity recognition, as wel as Bert-BILSTM-CRF model has been built and trained for extraction of information or named entities. Based on the findings, Computer Science, Statistics and Mathematics are the most required and co-occurred majors by recruiters, based on the data collected in 2020. Yet, majors such as Economic, Finance, Management, Electronics, Information Management, Physics also exist in the requirement list of some advertisements, which mostly co-occur with the mentioned top three majors. Findings indicate that communication, teamwork, logical thinking and responsibility, as well as SQL, Python, R, SPSS, Excel tools are having highest frequencies in the job postings. Furthermore, authors have categorized positions into three groups, which are big data mining, data analysis and data administration, data based on the findings in advertisements, that differ based on the programming languages and softwares (e.g., for data processing and management). The studies present considerably differing findings. To compare findings, SPSS has only around 4% and SAS 11% frequency in the study of Verma et al. (2019) for the DA profession, whereas they are in the most required skil list in the study of Fang and Zhou (2021). Moreover, different from other studies, Excel keyword lacks in presented findings of Smaldone et al. (2022) and Fang and Zhou (2020). Although two studies, in which data have been col ected from China’s job advertisements, have analyzed similar positions, the search query and the data col ection years have been different. Yet, all the skills that are described in the study of Fang and Zhou (2021) are mentioned in the study of Cao and Zhang (2021), except Matlab and Tableau. Non-existence of precise frequencies in the studies unable to make accurate observations on the change of or variations in skil requirements. As a summary, while previous studies have concentrated on current labor market requirements, this study expands literature by focusing on additional different perspectives. Among others, lack of accurate information on frequencies of skil s, focus mostly on US and China’s markets, existence of mismatches among the previous studies’ findings have been the rationale of this study. 160 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3 Methodology 3.1 Content analysis and data col ection In total, eight different data-related professions have been studied in the research, as shown in Figure 1. Figure 1: Studied positions Query and data have been limited to the existence of the term in the title of vacancy advertisements. The titles have been selected by referring to the previous studies, Future of Jobs Report (World Economic Forum, 2020) and LinkedIn’s (2020) career explorer portal. Derived from the domain problems that are listed above, including lack of common definition, personal interpretations in the literature and similarity of the professions, frequency technique of quantitative content analysis has been utilized for conducting the research. Quantitative content analysis has enabled generalization of the findings and elimination of possible biases, caused by subjective interpretations of the positions by field experts, of qualitative analysis (White & Marsh, 2006, Macnamara, 2018). Moreover, the research has been designed to find changes in required skil s and utilized tools by differentiating study results with the previous research, and authors N. Musazade: Tools and Technologies Utilized in Data-Related Positions: An Empirical Study of Job Advertisements 161 of previous research utilized content analysis and job advertisements as a methodology and research data, correspondingly. Factors, such as permission requirement for web scraping, usage commonness or number of advertisements, technical feasibility or architecture simplicity of the website, have been decisive for choosing Indeed.com job portal as a data source. Data have been col ected between November 2021 and January 2022 by utilizing Python programming language and BeautifulSoup library (i.e. web scraping). Common libraries, such as “requests”, “numpy” and “pandas”, have been deployed during data collection, preparation and processing. In total, 2658 advertisements of Denmark, Finland and Poland have been extracted. 3.2 Data Analysis As shown in Figure 2, data analysis has been conducted in three interrelated stages based on the research questions and objective. Skil s and tools have been col ected from previous studies. Among eight studied positions, six of the titles have matched with the titles from the previous research. Subsequently, Spacy library (i.e. Matcher class) has been applied to the col ected data for retrieving frequencies of the tools and skills in the job advertisements. Figure 2: Data Analysis Process 162 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 4 Discussion and Result The study focuses on understanding the required skills for data-related professions, changes in required skills, as well as actual skill requirements in the market. The results show that utilized tools have changed considerably. For instance, AI, cloud computing, Hadoop and SQL skills requirements started to rise, whereas Excel, SPSS, SAS and similar tools less expected from the candidates. Moreover, Python is more common in data science related professions, while R programming language usage increases in analytics related positions. 4.1 Fal ing trends An analysis of job advertisements shows that the demand for some skil s has decreased. For instance, Microsoft Excel has become less required for data professionals, particularly for DA and BI occupations. In contrast, most of the previous studies indicate Excel as one of the most required tools (e.g., Jiang & Chen, 2021; Verma et al., 2019; Cao & Zhang, 2021; Fang & Zhou, 2021). Based on research, some tools and programs, including Microsoft Word or Office, SPSS, and SAS, are being utilized less frequently in data-related jobs, as evidenced by a decrease in their requirement frequencies in job vacancies. For instance, in the study by Verma et al. (2019), SPSS has a frequency of around 4% for DA positions, while in this study it has a frequency of around 1%. Similarly, SAS has a frequency of 11% in Verma et al.'s study, and around 5% in this study. However, Fang and Zhou (2021), who studied China's market, found these tools to be among the most required for DA positions. A similar trend is applicable for Microsoft Office and PPT. 4.2 Rising trends In contrast, the requirement for some skil s and tools has shown an increasing trend. The matching trend to the utilization of cloud computing by data professionals is the rise of usage of keyword “cloud” in job advertisements, in particular for data engineer (DE) and DAR professions. Not only the “cloud” keyword, but also specific cloud-based database management systems, e.g., Azure and AWS, have increased their presence in the job advertisements, in particular for the DAR and DE positions. N. Musazade: Tools and Technologies Utilized in Data-Related Positions: An Empirical Study of Job Advertisements 163 The findings show that AI, cloud, deep learning, Mathematica, R, and Python have started to appear in one or more data-related vacancies, which is different from al of the previous studies except for the most recent US market study by Smaldone et al. (2022). In the study by Smaldone et al. (2022), "artificial intelligence" is the third most commonly appearing keyword after "big data" and "data processing," whereas it appears in only around one-third of advertisements in this study. While "deep learning" in Smaldone et al.'s study has a higher frequency than Python, in this study, around 20% of advertisements include "deep learning" compared to 87% for Python. The same trend applies to the term "machine learning." A possible interpretation of this differentiation is that tasks directly related to big data and AI have become the main duties and requirements for data scientists in the US market in recent years. In addition, Power BI and Tableau have become required and utilized more frequently in DA, BA and BI positions. Fog computing’s components and big data tools Spark and Hadoop have started to be utilized more as well, especially in DS, DE and DAR positions. Persaud (2020) lists Spark and Hadoop as one of the most required tools in the data related professions. Increase in the utilization of Hadoop may also be a sign of more extensive application of predictive analytics by organizations, since it is one of the enabling tools of predictive analytics (Pathak et al., 2018). The latter hypothesis is compatible with Power and Heavins’ (2017) findings, in which they identify more incorporation of predictive analytics into organizational operations and decision makings. 4.3 Differing trends of programming languages (R, Python, SQL) There are tools, in particular programming languages, that have become more required for some professions, while less demanded in other data-related positions. To elaborate, Python programming language has become more required in DS and DE positions; R programming language has become more in demand in BI and DA positions; Structured Query Language (SQL) requirement has a rising trend in DS, DE and DA positions. Jiang and Chen (2021) have found R in one in two job postings, whereas Python in three of four job postings titled as a “data scientist”. In China’s job advertisements, for DA positions R is more required than Python (Fang and Zhou, 2021), whereas 164 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY job postings with “data science” or “data mining” keywords have higher frequency for Python than R (Cao & Zhang, 2021). In other words, inferences match with the findings of Musazade (2022). Findings of Smaldone et al. (2022) and Verma et al. (2019) shows that R programming language appears more in advertisements than Python. For instance, in the study of Verma et al. (2019) R is higher in the Statistics skill category, whereas Python is higher in the Programming skill category of data scientist positions. Topic model ing of Smaldone et al. (2022) shows that R more frequently appears with the analytics, data management and business intel igence keywords, whereas Python more appears with the keywords such as data mining and processing, or artificial intelligence. 4.4 Background requirement/specialization In addition to the particular tools, the study shows that “technical”, “engineering” and “computer science” terms have started to appear less in vacancies of data related positions compared to previous studies (e.g., Verma et al., 2019, Jiang & Chen, 2021), which may imply a diminishing role of technical background in data professionals. For instance, for DE positions “computer science” term has 22% frequency, whereas in the study of Jiang and Chen (2021) the term appears in half of the advertisements. The study of Cao and Zhang (2021) has defined “computer science” and other related terms (e.g., mathematics, statistics) as one of the required background majors from candidates. Keyword “statistics” for the DS profession has started to appear 10-15% less: it has around 60% frequency in study of Verma et al. (2019) and Cao and Zhang (2021). Moreover, the study suggests that data professionals have started to be required to possess a narrower competence and specialization in particular common, unique and extensive tools and technologies. For instance, different from findings of previous research, Python and Python libraries (e.g. Pytorch) have started to occur more frequently in advertisements. Yet, for a business intel igence role, for instance, the same judgment cannot be applied, in which tools utilized are distributed more broadly and shifts are more thorough. N. Musazade: Tools and Technologies Utilized in Data-Related Positions: An Empirical Study of Job Advertisements 165 4.5 Grouping of professions Based on the findings on required tools some professions can be categorized. Firstly, DAR, consultant and partly DE positions involve extensive labor on databases and different cloud platforms. Jiang and Chen’s (2021) findings are similar with the findings of this study, except instead of a data consultant position, data scientist profession is a part of the group. Secondly, professions such as DA and BI frequently use BI, data visualization and analytics products, such as Tableau or Power BI. Thirdly, SQL and Python programming languages are frequently used in DE, DS and DA positions. Findings indicate that BI and DA professions are similar to each other in terms of skil s they require. Verma et al. (2019) analyze the existence of similarity between DS and DA positions, and between BI and BA professions, whereas the study lacks comparison of positions al together. Final y, both findings of this research and Jiang and Chen (2021) indicate the distinctiveness of the BA profession, which may not be ful y matched to a particular category. 5 Conclusions The research findings show that the skills required and tools utilized in data-related professions have experienced substantial changes in the last two-four years. As the presented results address RQ1, while some tools have common trends in most or al of the data-related professions, differences in frequencies of skills and tools, and of particular tools’ trends may enable differentiation between similar data-related professions. Furthermore, regarding RQ2, this study shows that skill requirements for data related positions have changed in the last two-four years, which may imply the change of tools and technologies used in data-related positions, as well as usage objectives of big data by organizations. For example, since Python and machine learning are highly correlated in job ads (Smaldone., 2022), the rising requirement for Python may suggest the rising usage of big data for machine learning. Fal ing trend of most of the tools belonging to the statistics category defined by Verma et al. (2019), such as Excel, SAS, SPSS may imply less usage of big data for statistical analysis only. More specifically, cloud computing tools, AI and deep learning skills and knowledge have started to be required more from candidates. SQL, Power BI, R and Tableau have become more common in the job postings for analytics related professions. 166 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Moreover, SQL, Spark, Hadoop and Python have become more in demand for data scientist and engineer professions. However, Microsoft Word or Office, SPSS, SAS, Excel have started to be utilized less in data related positions. Although comparison of finding with the previous studies’ results and of results of previous studies outline existence of differences in required tools (e.g. Excel), some common observations can be derived. For example, SQL both in this study and in previous studies is the most required database management tool. While Python is mostly used in data scientist and related positions, data analyst and similar professions utilize R more than Python. Firstly, the research contributes to the literature by expanding study of data-related professions to the new geographical area and studying the labor market of three countries: Denmark, Finland and Poland. Secondly, the study not only defines current market requirements, but also studies and defines an approach for exploring changes in the market. Thirdly, findings of the research may assist authors in accurate definition of data-related professions based on the skill required for them. Finally, our results can also help in recognizing the importance of the existence of human resources and skills, in addition to the cost, scalability and functionality, security and availability, as the component of multi-criteria decision-making task of tool and system selection (Kachaoui & Belangour, 2019, Grandhi & Wibowo, 2018). Findings of the research may contribute to the closing skil gap, re-skilling, correctly defining data-related positions by the job advertisers, specialization for a particular data-related profession, search of relevant vacancies based on the possessed skil s by candidates, as well as up-skilling for the professions in the field. Furthermore, education institutions may utilize findings in defining curriculum for the studied specializations based on the market requirements and trends. Shifts in the skill requirements can be influenced by country-level variations, considering previous studies have been conducted in other markets. Frequency of the words may be influenced by the terms in the language of the advertisements. Moreover, real expectations of recruiters may be represented incompletely or inaccurate in the job postings. Number of collected data for some professions may be considered insufficient. Finally, utilization only quantitative analysis may inhibit possible interpretation of terms, in particular soft skil s, and conjunctions among terms have been dismissed. N. Musazade: Tools and Technologies Utilized in Data-Related Positions: An Empirical Study of Job Advertisements 167 Future research may study possessed (i.e. supplied) skills for refining the existing skill gap. Secondly, studying each high in demand tool and relationships between them, may present insights on usage objectives of the tools and possible need for a shift between them. Thirdly, categorization of job advertisements based on the position level, company size and different industries may depict a clear picture of the current required skills for each specialization, as well as possible differences depending on listed or other variables. This paper focused on a limited number of professions, whereas the number studied positions can be extended to other professions. References Cao, L., & Zhang, J. (2021, April 1). Skill Requirements Analysis for Data Analysts Based on Named Entities Recognition. IEEE Xplore. https://doi.org/10.1109/ICBDIE52740.2021.00023 Cattaneo, G., Micheletti, G., Glennon, M., La Croce, C., & Mitta, C. (2020). The European data market monitoring tool : key facts & figures, first policy conclusions, data landscape and quantified stories : d2.9 final study report. Publications Office of the European Union. De Mauro, A., Greco, M., Grimaldi, M., & Ritala, P. (2018). Human resources for Big Data professions: A systematic classification of job roles and required skill sets. Information Processing & Management, 54(5), 807–817. https://doi.org/10.1016/j.ipm.2017.05.004 Fang, F., & Zhou, Y. (2021). A Study on Recruitment of Data Analyst Based on Text Mining and Visualization Technology. Journal of Physics: Conference Series, 1952(4), 042017. https://doi.org/10.1088/1742-6596/1952/4/042017 Gardiner, A., Aasheim, C., Rutner, P., & Wil iams, S. (2018). Skil Requirements in Big Data: A Content Analysis of Job Advertisements. Journal of Computer Information Systems, 58(4), 374–384. https://doi.org/10.1080/08874417.2017.1289354 Grandhi, S., & Wibowo, S. (2018). A Multi-criteria Group Decision Making Method for Selecting Big Data Visualization Tools. Journal of Telecommunication, Electronic and Computer Engineering (JTEC), 10(1-8), 67–72. Retrieved from https://jtec.utem.edu.my/jtec/article/view/3737 Granville, V. (2014). Developing analytic talent: Becoming a data scientist. John Wiley & Sons, Inc. Jiang, H., & Chen, C. (2021). Data Science Skills and Graduate Certificates: A Quantitative Text Analysis. Journal of Computer Information Systems, 1–17. https://doi.org/10.1080/08874417.2020.1852628 Kachaoui, J., & Belangour, A. (2019). A Multi-criteria Group Decision Making Method for Big Data Storage Selection. Networked Systems, 381–386. https://doi.org/10.1007/978-3-030-31277-0_25 Kenett, R., & Redman, T. C. (2019). The real work of data science : turning data into information, better decisions, and stronger organizations. Wiley. Koh, A. (2020). Anticipating and preparing for the future—one example from higher education: The Singapore Management University (SMU) experience. In B. Panth & R. Maclean (Eds.), Anticipating and Preparing for Emerging Skil s and Jobs: Key Issues, Concerns, and Prospects (pp. 41–46). Springer. https://doi.org/10.1007/978-981-15- 7018-6_6 LinkedIn. (2020). https://linkedin.github.io/career-explorer/. Linkedin.github.io. https://linkedin.github.io/career-explorer/ 88 168 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Macnamara, J. (2018). 11. Content Analysis. Mediated Communication, 191–212. https://doi.org/10.1515/9783110481129-012 Retrieved from: https://www.researchgate.net/publication/327910121_Content_Analysis Marr, B. (2019). Artificial intelligence in practice: How 50 successful companies used artificial intelligence to solve problems (First Edition). Wiley. McKinsey. (2021). The future of work after COVID-19. Retrieved October 23, 2021, from https://www.mckinsey.com/featured-insights/future-of-work/the-future-of-workafter-covid-19 Miller, S. (2014). Collaborative approaches needed to close the big data skills gap. Journal of Organization Design, 3(1), 26–30. https://doi.org/10.7146/jod.9823 89 Musazade, N. (2022). Understanding the relevant skills for data analytics-related positions: An empirical study of job advertisements. www.doria.fi. https://urn.fi/URN:NBN:fi-fe2022031323297 Pathak, A. R., Pandey, M., & Rautaray, S. (2018). Construing the big data based on taxonomy, analytics and approaches. Iran Journal of Computer Science, 1(4), 237–259. https://doi.org/10.1007/s42044-018-0024-3 Persaud, A. (2020). Key competencies for big data analytics professions: a multimethod study. Information Technology & People, ahead-of-print(ahead-of-print). https://doi.org/10.1108/itp-06-2019-0290 Power, D. J., & Heavin, C. (2017). Decision support, analytics, and business intel igence. Business Expert Press. Sheng, J., Amankwah-Amoah, J., Khan, Z., & Wang, X. (2020). Covid-19 pandemic in the new era of big data analytics: Methodological innovations and future research directions. British Journal of Management, 32(4), 1164–1183. https://doi.org/10.1111/1467-8551.12441 Smaldone, F., Ippolito, A., Lagger, J., & Pellicano, M. (2022). Employability skills: Profiling data scientists in the digital labour market. European Management Journal, 40(5). https://doi.org/10.1016/j.emj.2022.05.005 Verma, A., Yurov, K. M., Lane, P. L., & Yurova, Y. V. (2019). An investigation of skil requirements for business and data analytics positions: A content analysis of job advertisements. Journal of Education for Business, 94(4), 243–250. https://doi.org/10.1080/08832323.2018.1520685 White, M. D., & Marsh, E. E. (2006). Content Analysis: A Flexible Methodology. Library Trends, 55(1), 22–45. https://doi.org/10.1353/lib.2006.0053 World Economic Forum. (2020). The Future of Jobs Report 2020. https://www3.weforum.org/docs/WEF_Future_of_Jobs_2020.pdf N. Musazade: Tools and Technologies Utilized in Data-Related Positions: An Empirical Study of Job Advertisements 169 Appendix A: Literature Review of Data-related Positions Study - Scope or “Title” Methodology (tools) Study focus or findings Frequent tools Web Scraping (Octoparse), Text Thematic areas for Smaldone et al. (2022) Mining (R), LDA, skills, correlated skills, Python, R - DS Total frequency (TF), top skills Network analysis Skills and their Experts for skil s’ categories’ Verma et al. (2019) - categories, Content comparative SAS, R, Python, SQL, DS, DA, BI, BA analysis, Web scraping importance, Excel (R) comparison of professions Web crawling, text Domain knowledge, analysis (job postings, Persaud (2020) -Big kinds of skills are university programs), Hadoop, sql, python, DA-related expected, e.g., social, Interviews, semantic spark professions functional, cognitive, relationships technical. (Leximancer) Overlapping skills and grouping of Text analysis with professions, match Jiang & Chen the created list of of skills of graduate (2021) - DA, BA, SQL, Excel, R, dictionary, total certificated with the DAR, DE, DS, DS Python frequency (Python labor market Researcher SpaCy) requirements, comparison of top skil s of professions 170 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY R, Python, Java, Crawling MySQL, Hive, Fang & Zhou (Bazhuayu), Text Spark, Oracle, Top skills (2021) - DA mining (dictionary) MATLAB, (Python -Jieba) Hadoop, SPSS, Tableau, SAS Required majors and their cooccurance, Excel, SQL, SPSS, Cao & Zhang (2021) - Web crawler, BERT - soft and hard skil s, Python, R, hive, PPT, “data science”, “data BiLSTM -CRF for kinds and SPARK, Haddop, mining” NER (Python) comparisons of Java, SAS; positions. Identifying and Web Scraping Hadoop, SQL, comparing job De Mauro et al., (Portia), Text mining Azure, AWS, Python, families, skill sets and (2018) - “Big Data” and expert judgment, SQL, Java, Ruby, R, their match with each LDA (R) SAS, MATLAB job family Defining frequent Text analysis (R), terms and studying Gardiner et al. (2018) Consensus pile-sort term or skil s Hadoop, Java; - “big data” protocol with domain categorization per experts concept, presenting developed framework CUSTOMERS’ QR CODE USAGE BARRIERS IN A BRICK-AND-MORTAR STORE: A QUALITATIVE STUDY TIINA PAANANEN,1 MATILDA HOLKKOLA,1 MARKUS MAKKONEN,1 LAURI FRANK,1 TIINA KEMPPAINEN2 1 University of Jyväskylä, Faculty of Information Technology, Jyväskylä, Finland ti na.e.paananen@jyu.fi, matilda.i.holkkola@jyu.fi, markus.v.makkonen@jyu.fi, lauri.frank@jyu.fi 2 University of Jyväskylä, Jyväskylä University School of Business and Economics, Jyväskylä, Finland ti na.j.kemppainen@jyu.fi The purpose of this study is to explore the factors that hinder customers from utilizing QR codes when they visit brick-and- Keywords: mortar (B&M) stores. The research was conducted at a Finnish QR codes, fashion retailer, where three types of QR codes were displayed omnichannel, channel, for customers to use. In order to identify the barriers that retail, customers face, two different sets of data were col ected and customer perspective, analyzed: open-ended survey data (n = 101) and interview data commerce, (n = 16). The findings showed two main categories of barriers, QR code customer and company related, with seven sub-barriers. The barriers, usage customer-related barriers included lack of interest in QR codes, barrier, user-related usage issues, desire for a device-free B&M store, and brick-and-dislike toward QR codes. The company-related barriers included mortar store, unnoticeable QR codes in the B&M store, service personnel in B&M, the B&M store, and QR code-related technical problems. Finland DOI https://doi.org/10.18690/um.fov.4.2023.11 ISBN 978-961-286-751-5 172 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The customer journey is going through a digital transformation (Zimmermann, 2016), and due to the change from a multichannel to omnichannel environment, the digital and physical worlds are evolving into an single entity instead of being separate channels. Physical channels, such as brick-and-mortar (B&M) stores, include digital elements, such as electronic displays (Roggeveen et al., 2016) and quick response (QR) codes (Tiwari, 2016) to name a few. QR codes are becoming more and more common, as they al ow for easy and quick access to value via a smartphone. Thus, QR codes are a way to guide customers to engage with the digital world as part of their shopping experience in a B&M store, but customers are not utilizing this opportunity (Lau et al., 2022). Previous research on customers’ QR code usage has focused on QR codes’ visual appearance (Gao et al., 2015; Tsai & Peng, 2023); acceptance (Kim & Woo, 2016); use alongside with mobile trust and technology acceptance in omnichannel (Lawry & Choi, 2013); perceptions of codes on milk packages (Lau et al., 2022); impact on customer satisfaction and purchase intention (Hossain et al., 2018); customer experience (Shin et al., 2012); and print advertisements with codes (Trivedi et al., 2019). Even though QR codes have been investigated from multiple perspectives, there is a gap in our understanding of why customers do not use QR codes in B&M stores. This knowledge may help companies to understand the customers’ perspective and make strategic decisions when implementing QR codes as a source of value in B&M stores to enable omnichannel shopping. Hence, this study focuses on identifying the barriers that prevent customers from using QR codes while shopping in B&M stores. The research was conducted at a Finnish fashion retailer, where three types of QR codes were displayed for customers to use. To gather information about the chal enges that customers face when using QR codes, we col ected and analyzed two sets of data: open-ended survey data (n = 101) and interview data (n = 16). T. Paananen, M. Holkkola, M. Makkonen, L. Frank, T. Kemppainen: Customers’ QR Code Usage Barriers in a Brick-and-Mortar Store: A Qualitative Study 173 2 Literature Review A QR code is a two-dimensional matrix code which enables high-speed decoding via a smartphone’s built-in camera. It was developed in 1994 by Denso Wave (a Toyota subsidiary) to track vehicle parts, and it has been implemented in marketing campaigns since 2011 (Tiwari, 2016). Ever since, QR codes have been a way for customers to navigate to the desired digital end point. The visual appeal of QR codes has been noted as a usage barrier, as Gao et al. (2015) criticized that the tedious black-and-white code distracts from the quality of the advertisement. Thus, beautified (Tsai & Peng, 2023), invisible (Gao et al., 2015), stylized aesthetic (Xu et al., 2019), and colorful and decorated QR codes have been invented in an attempt to make the codes more attractive (Lin et al., 2013). In the luxury context, the visual appeal of QR codes was linked to customers’ QR code scanning intentions (Lawry & Choi, 2013). Additionally, Singhal and Pavithr (2015) explained that the optimal size of a QR code for scanning should be calculated. Sometimes, QR codes are not accepted by customers. Shin et al. (2012) studied QR code acceptance and found that while users might have a positive perception toward QR code features, this does not lead to QR code usage because confirmation of what wil happen upon scanning the QR code is missing. The major technology acceptance drivers are perceived usefulness, enjoyment, and usability (Pucihar, 2020). Scanning intentions can prevent acceptance of QR codes and lead to their rejection. Lawry and Choi (2013) found that mobile trust was not related to QR code scanning intentions in the luxury retail context and called for experimental studies in real-life retail settings. Oppositely, Liébana-Cabanillas et al. (2015) found that attitude, innovation, and subjective norms were linked with intentions to use QR codes in the future in the context of mobile payment acceptance. Similarly, Kim and Woo (2016) found that customers have a positive intention toward using QR codes as an information tool. Additional y, QR codes have been criticized as unclear and cumbersome from customers’ perspectives. Lau et al. (2022) found that in low-cost items (e.g., milk), the ‘hassle cost’ of scanning many QR codes to get a discount was too much effort. The study found that the perceived value of or trust toward QR codes did not lead to scanning and that the novelty of QR codes resulted in a lack of interest toward them. 174 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Some studies have confirmed QR codes’ feasibility. QR code promotions with product information in magazines were found to be motivating for customers (Ertekin & Pelton, 2014). Also, QR codes can be a bridge to move traditional push-based marketing (e.g., print) to a digitalized environment (Trivedi et al., 2019). This idea for using QR codes follows the omnichannel logic by Brynjolfsson et al. (2013), in which digital and physical spaces become one entity instead of channel clutter. Internal and external touchpoints in the customer journey can support each other (Mali et al., 2022; Paananen et al., 2022a; Holkkola, 2022a) with the aim of creating a seamless customer experience (Lemon & Verhoef, 2016). Digital transformation might increase the familiarity of QR codes. A study by Zhong and Moon (2022) found that the COVID-19 pandemic increased the use of contactless services, leading to familiarity with contactless technologies, such as QR codes, which are most likely to continue being used after the pandemic. Additionally, Rotsios et al. (2022) found that QR codes’ contents can be entertaining and enriching for customers. The problem lies in QR code scanning barriers, which previous research has failed to address. As there are justified reasons for using QR codes, such as omnichannel operations, it is crucial to understand what prevents the use of QR codes in B&M stores. 3 Methodology To map customers’ real-life experiences of QR code usage barriers during a B&M store visit, we put QR codes on display at one of the case company’s B&M stores and asked about customers’ thoughts on them. The case company was a Finnish fashion retail company and clothing brand. Three different types of QR codes (see Appendix A) were on display from May 2022 to September 2022: 1. The first type of QR code was placed next to a clothing rack that contained the brand ambassador’s (long term influencer) favorite items. The QR code led to the ambassador’s Instagram post about the company’s products. 2. The second type of QR code was placed next to jackets and trousers. The QR code led to the product page of the company’s online store. 3. The third type of QR code was placed close to a sale outlet clothing rack, inside the fitting booth, at the checkout, and at the front door. In addition, the salesperson placed a promotional leaflet with the QR code on it in the shopping bag with the purchased items. The QR code led to the survey. T. Paananen, M. Holkkola, M. Makkonen, L. Frank, T. Kemppainen: Customers’ QR Code Usage Barriers in a Brick-and-Mortar Store: A Qualitative Study 175 The survey data colection was carried out using both a pen-and-paper survey and an online survey that was implemented with the LimeSurvey service and could be accessed by either entering a web address or scanning a QR code. The survey questionnaires of both the surveys were identical and inquired the customers about their background information, B&M store visit frequency, use of the brand’s other channels, customer experience, previous QR code use experience, and QR code usage during the B&M store visit. Those who answered that they had not used a QR code during the B&M store visit were asked why via an open question. After filling out the survey, customers could take part in prize drawing of a €50 gift card and, also, volunteer for an interview to receive an additional reward (a bag valued at €40) and tel their experience in more depth. The interview data col ection was carried out using semi-structured interviews with 16 selected volunteers (see Appendix B): three males and 13 females (referred to as Participants 1–16: P1–P16). The selection criteria were persons who had visited the B&M store between May 2022 and September 2022 while the QR codes were on display (see Appendix C). Data saturation was reached after 16 interviews, when new information was no longer obtained. One test interview was conducted. The interviews were conducted via Microsoft Teams in September 2022 with an average duration of 59 minutes. The participants were asked to describe their shopping experiences when visiting the B&M store, their IT skills, their QR code usage before and during the visit, and their general brand relationship with the company. The order of the interview questions was not strictly defined in advance, which enabled additional questions, following Myers and Newman’s (2007) advice for semi-structured interviews. All the interviews were recorded and transcribed. The quotes were translated from Finnish to English. Inductive coding was selected to process the data (Thomas, 2003). The survey data showed the themes, and the interviews explained them. First, the survey data’s open answers to why customers did not use QR codes in the B&M store (i.e., usage barriers) were imported into Microsoft Excel starting with col ecting the QR code usage related barriers. Second, after multiple rounds, two main categories were identified: company-related and customer-related QR code usage barriers. Third, after several rounds of examination, seven sub-categories emerged under these two main categories. After careful familiarization with the interview data, we took the seven identified sub-categories from the survey data’s open answers and coded the 176 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY interview data with these barriers using the ATLAS.ti qualitative analysis software. Finally, seven-sub barriers with the deeper explanations were identified from the interview data. 4 Findings In total, 101 valid survey responses were obtained. The demographics (gender, age, and socioeconomical status) of respondents (referred to as R1–R101) are reported in Table 1. Most of them were women (74.3%). Many were 30–39 years old (28.7%). The average age was 39.9 years and the standard deviation 14.6 years. The respondents presented varying socioeconomic statuses, but the most of them were employed (72.3%). The respondents were able to pick multiple socioeconomic status choices. All the respondents were from Finland. Almost al answered yes to having previous experience with QR code usage (93.1%). Table 1: The survey respondents’ demographics (n = 101) n % n % Gender Socioeconomic status Male 26 25.7 Student 23 22.8 Female 75 74.3 Employed 73 72.3 Self-employed 5 5.0 Age Unemployed or unable to work 3 3.0 15–29 years 26 25.7 Stay-at-home parent 1 1.0 30–39 years 29 28.7 Pensioner 7 6.9 40–49 years 20 19.8 Has used QR codes before 50–59 years 11 10.9 Yes 94 93.1 60+ years 15 14.9 No 6 5.9 No response 1 1.0 Despite their previous experiences with QR codes, only 11 (10.9%) of the respondents reported using QR codes during their B&M store visit. In contrast, 83 of the respondents (82.2%) did not use the QR codes during the store visit, and seven (6.9%) did not respond to the question. For all respondents who answered that they did not use QR codes while visiting the B&M store, we asked the open question: “Why did you not use QR codes when visiting the B&M store?” Of the 83 non-users, 78 (93.98%) answered this question. The non-users’ open answers were T. Paananen, M. Holkkola, M. Makkonen, L. Frank, T. Kemppainen: Customers’ QR Code Usage Barriers in a Brick-and-Mortar Store: A Qualitative Study 177 analyzed and divided into two main categories of usage barriers (company and customer related) with seven sub-barriers (see Table 2). These are presented in more detail in Table 2 and expanded upon using insights from the interviews. Table 2: Themes and open answer counts from the survey Theme Count Customer-related QR code usage barriers 54 Lack of interest in QR codes 32 User-related usage issues 12 Desire for a device-free B&M store 6 Dislike toward QR codes 4 Company-related QR code usage barriers 36 Unnoticeable QR codes in the B&M store 26 Service personnel in the B&M store 8 QR code-related technical problems 2 4.1 Customer-Related QR Code Usage Barriers Lack of interest in QR codes. The lack of interest in QR codes could be seen through the open answer data, appearing as a passive resistance to QR codes. This passive resistance emerged in the form of ignoring QR codes but not directly resenting them. This topic was explained by the survey respondents as, for instance, too tired to use them (R12, R63), no need (R28, R55, R58, R72, R87, R91, R98, R99), or I wasn’t buying products right now (R7, R17). The interview data revealed that QR codes were uninteresting for some interviewees because the codes did not seem rewarding or clear enough to be scanned. For example, some interviewees were too confused to get excited about scanning the QR code. “Wel , maybe. I don’t know what [value] these QR codes bring for me or why should I read them.” (P3) “I don’t know. Maybe those [QR codes] were nothing worth mentioning, so I just saw that there is one and maybe wondered what that was but haven’t started to investigate or look into it any further.” (P5) 178 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Some interviewees found the design of the QR codes and tags unappealing, describing them as “just black-and-white” (P9) or “just some paper with that QR code” (P6). Also, the QR codes were associated with products that were irrelevant to the interviewees. “Maybe I wasn’t interested enough because they were not included with products I would buy anyways. So, then, I didn’t see it as relevant.” (P2) User-related usage issues. The data revealed that customers entering the B&M store had varying technical skills and devices, leading to some being unable to utilize the QR codes even if they wanted to, as indicated by the responses “I did not have the application to scan the QR codes” (R43), “lack of Internet” (R25), “lack of skill” (R73), “low battery” (R67), “I did not have a mobile phone with me” (R75), and “unfamiliar” (R92). Problems with the user’s skills or their device’s functionality created barriers to QR code use. In the interviews, however, many interviewees described QR code usage as an “easy thing to do” (P2) and an “everyday thing” (P13) in a technical sense. Most of the interviewees described their general IT skil s as advanced. Despite this, some interviewees faced issues when figuring out how to use the QR code scanner. “I remember that it was like ‘al fingers and thumbs’ moment when you don’t use them every day, like, how does this work.” (P7) The interviews also revealed that dysfunctional user devices created a barrier to QR code usage. For example, one interviewee’s phone did not have a QR scanner directly in the camera application. Additional y, the QR code scanner application’s flood of ads was distracting to another interviewee. “These QR code scanners that I have, those have so many ads in them, that there is such a great danger of hitting an ad and then going to the wrong page.” (P9) Desire for a device-free B&M store. The reluctance to adopt QR codes was further justified by the preference for a B&M store to be free of digital devices. One of the survey respondents indicated that “in a B&M store, I want to concentrate on looking and trying out physical products. In a B&M store, I don’t surf on the Internet” (R47). Another survey respondent reported that they just “prefer trying clothes on me instead of looking for inspiration from the Internet” (R70). This did not encourage the use of QR codes since phones were viewed as a distraction and T. Paananen, M. Holkkola, M. Makkonen, L. Frank, T. Kemppainen: Customers’ QR Code Usage Barriers in a Brick-and-Mortar Store: A Qualitative Study 179 not part of a B&M shopping moment. These respondents wanted to experience, feel, and touch the products: “products in the B&M store were interesting and nicely on display” (R53). Thus, a barrier was created because some of the respondents did not want to experience products through the phone in the B&M store. This phenomenon of desire for a device-free B&M store was also reported in the interviews because “these products can be enjoyed without any digital aspects” (P12). Some interviewees even left their phone in the car to ensure their device-free shopping moment in the B&M store. They indicated that they would rather experience the store without digital encounters: “When I saw those [QR codes], I thought, “Is this necessary, when I come here to the B&M store?”…. I don’t want to dig out my phone. Often my phone might be left in the car, and it isn’t always accompanying me. Then you don’t assume that you need it there, and then you hope that you can reach the information in the store without using digital devices. There is a webstore etc. when you can use technology.” (P12) Dislike toward QR codes. Disliking QR codes emerging as active resistance to QR codes, which means direct resentment towards QR codes in particular. In the open answers from the survey, certain respondents reported dislike and strong negative emotions rather than just a lack of interest, as exemplified by the responses “QR codes feel a little bit complicated and useless in general” (R14), “no additional value” (R32), and irrelevant for them right now (R44, R72). Some of the respondents indicated that they “haven’t learned how to use them” (R57) or “not everyone is a “ ‘QR-code person’ ” (R77). This resistance toward QR codes created a barrier to their use and sometimes even to learning how to use them. Additional y, some interviewees felt that QR codes were unnatural because B&M store visits “do not require any phone” (P13) or they were not their “cup of tea” (P13) and, therefore, created a usage barrier. QR codes seemed to even represent an unwanted change or trend phenomenon. “They represent the new era, but I don’t care for them or need them” (P13) “They are useless, or they are put up without purpose. They surely are useful for someone, but I feel that they seem out of place and bit “glued on.” It’s like they are put up because its trendy or new even though obviously they are no longer a new thing. Regardless, I think they are not as needed there [in the B&M Store] as in other places.” (P4) Some interviewees just wanted to “shop without scanning them” (P12) because the QR codes did not feel necessary or provide any additional value for them. 180 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 4.2 Company-Related QR Code Usage Barriers Unnoticeable QR codes in the B&M store. Many survey respondents reported not noticing the QR codes during their B&M store visit (R8, R16, R18, R20, R27, R33, R34, R36, R39, R43, R48, R50, R51, R52, R55, R61, R64, R69, R70, R83, R84, R101). Almost all of the survey respondents had used QR codes before (see Table 1), and therefore, they understood what a QR code is and what it looks like, although some respondents reported noticing only the survey QR code (R16, R29, R31): “I thought al codes were the same; separating different codes could have helped” (R31). The interview data exposed issues with the QR codes, such as that they were too small (P2, P3, P4, P6, P9, P10, P16) or lacking color (P9) to be noticeable. Sometimes the QR codes were too wel -camouflaged in the store. Additionally, the interviewees questioned themselves, wondering if perhaps they saw the QR codes, but their minds did not even register the codes, thus barring usage. “Either those were so wel -integrated into the product description, or the eye is already so used to seeing QR codes around that they are not noted in that sense” (P11) “They blend in really well with the environment, maybe a little too well” (P6) Service personnel in the B&M store. Some of the survey respondents described usually asking for customer service help when they needed information, such as, “I would rather ask a customer servicer” (R49) or that customer service personnel were “easy to approach” (R93). Customer service even outplayed the QR codes as a source of information: “I didn’t need it. Great service and salespersons who know their business!” (R81). Thus, the customer service seemed to be of too high of a level to encourage survey respondents to use the QR codes. One interviewee clarified the matter: “one can ask a salesperson, at least in a B&M store” (P1). This created a QR code usage barrier because customer service as an information channel came to their minds before using QR codes. Some interviewees also trusted the information and answers from the store’s salespeople more than trying to find information in other ways. This created a barrier to using the QR codes because customer service was the first choice to find answers: “If I wanted more information about those products, I’d rather ask the salesperson when I’m in the store. And I trust that the salesperson can best tell me about it and answer questions.” (P12) T. Paananen, M. Holkkola, M. Makkonen, L. Frank, T. Kemppainen: Customers’ QR Code Usage Barriers in a Brick-and-Mortar Store: A Qualitative Study 181 QR code-related technical problems emerged when positioning and visual design prevented their usage. It was noted in the survey, for example, that the QR codes did not work (R36, R65) when some of the survey respondents tried to scan the codes. One of the interviewees shed light on this technical problem by stating that the QR codes were placed too far away to be scanned with some devices, therefore creating a usage barrier: “I noticed the QR code related to this study, which was honestly a bit far behind the counter, so I had to type in the visible URL that was below it to get there.” (P6) Also, some participants experienced QR code dysfunction due to their smal size (P2, P3, P4, P6, P9, P10, P16), which prevented them from being scanned. 5 Discussion This study investigated customers’ barriers to QR code usage in a B&M store in a real-life setting. The identified barriers are summarized in Figure 1. Figure 1: Summary of the findings This study deepens the previous knowledge by identifying and describing QR code usage barriers relating to both customers and companies in the B&M store setting. The study responds to Lawry and Choi’s (2013) call for experimental studies on real brands and QR codes. As in the study by Zhong and Moon (2022), participants in this study were familiar with QR codes but chose not to use them despite their previous experience. This study suggests that QR code usage barriers in the B&M store setting can be customer related. Firstly, the finding of a lack of interest in QR codes is in line with Lau et al.’s (2022) finding of a ‘hassle cost’ when scanning QR codes and with Gao et al.’s (2015) criticism toward QR codes’ appearance, which are perceived as 182 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY unappealing and requiring too much effort to use. Also, as in the study by Shin et al. (2012), a lack of confirmation regarding what would happen upon scanning the QR code was also shown in this study; the QR codes were not clear enough despite information in the QR code tags. Because they did not signal to the customer what would happen after a successful scan, customers were not motivated to take action and scan the tags. Perhaps a promise of more exciting content from a customer’s point of view (e.g., sustainability information, product manufacturing information, or product color options) could make the QR codes more attractive and thus worth of the ‘hassle cost’. Secondly, sometimes the customers did not expect to encounter QR codes in a B&M store. User-related usage issues revealed customers who were unprepared to scan the QR codes because they lacked the required usage skills, suitable devices, or Internet connection. These are novel findings because QR code usage barriers in terms of devices and skil s have been scarcely reported in prior research. QR codes should not complicate the buying of products in the B&M setting but rather add extra confirmation to the purchase intent. Thirdly, as a novel finding, this study found that customers’ desire for a device-free B&M store visit, without the need to use their digital devices, is a notable usage barrier. Some customers had even left their devices in their cars and did not expect digital encounters before stepping into the B&M store. This finding shows aspects of technostress (e.g., Salo et al., 2019; Pirkkalainen, et al., 2019) because annoyance toward using a phone in the B&M store was noticed in the study. The stressor could be, for example, a shattered expectation of only a physical encounter during the B&M store visit. This is an interesting finding because the study by Holkkola et al. (2022b) found that certain customers tend to showroom, meaning they actively search for more information about products and better deals via their phone while physical y experiencing products in a B&M store. Lastly, the findings of customer-related QR code usage problems suggest active resistance. This dislike toward QR codes could be due to a failure to meet the technology acceptance drivers, such as providing enjoyment, usability, or usefulness (Pucihar, 2020) or perceived value or trust (Lau et al., 2022). The QR codes were not accepted; these actively resisting QR code dislikers did not even have an intention of using QR codes and showed resistance toward the technology. T. Paananen, M. Holkkola, M. Makkonen, L. Frank, T. Kemppainen: Customers’ QR Code Usage Barriers in a Brick-and-Mortar Store: A Qualitative Study 183 In turn, this study suggests that a company can involuntarily cause QR code usage barriers in the B&M setting. First, the findings of unnoticeable QR codes in the B&M store showed novel findings concerning QR codes may be too small, blend into the background, or be difficult to tel apart from each other. In line with Lau et al.’s (2022) finding’s inattention, some customers did not register or notice the QR codes. It is possible that they did not expect QR codes in the B&M store. The visibility of QR codes and their tags should be improved to make them more noticeable. Visibility could be improved with conspicuous design choices to make a pop-out effect and not to camouflage them. Also, improving QR codes’ visual appearance (Gao et al., 2015; Lin et al., 2013; Tsai and Peng, 2023; Xu et al., 2018) should be considered. In cases where multiple different QR codes are used, they should be clearly separated from each other. Second, this study showed that the service personnel in the B&M store outplayed the QR codes because customer service was available and was too good. Thus, QR codes may be useful if customer service is not available or if additional content (e.g., videos) is provided via the QR codes. Lastly, non-working QR codes were reported. Singhal and Pavithr (2015) previously noted that QR codes should be of an optimal size and scanning distance. Thus, a usage barrier was created by the QR codes’ small size and big distance. We suggest planning where QR codes will be located and their size. 6 Limitations and Future Research This study considered the QR code usage barriers for customers in a real-life setting of a fashion retail store in Finland. This study provides an authentic in-depth understanding based on individual participants’ experiences and thus is not to meant to be generalized. Notably, of participants, Finnish female B&M visitors were mostly represented in the findings. Future studies could research supporting factors of customers’ QR code usage in a B&M store in different contexts. Retail markets should examine ways to ensure sustainable consumption following the barriers noted by Kemppainen et al. (2021). Last, QR code as an intermediary to a brand relationship could be investigated as stressed by Paananen et al. (2022b). 184 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY References Brynjolfsson, E., Hu, Y. J., & Rahman, M. S. (2013). Competing in the age of omni-channel retailing. MIT Sloan Management Review, 54(4), 23–29. Ertekin, S., & Pelton, L. E. (2014). An empirical study of consumer motivations to use QR codes on magazine ads. American International Journal of Contemporary Research, 4(5), 47-55. Gao, Z., Zhai, G., & Hu, C. (2015). The invisible QR code. In Proceedings of the 23rd ACM International Conference on Multimedia (pp. 1047–1050). ACM. Holkkola, M., Frank, L., Kemppainen, T., Paananen, T., & Luhtanen, V. (2022a). The role of sustainability in online customer experiences: a qualitative study on female fashion shoppers. In Proceedings of the 14th Mediterranean Conference on Information Systems (MCIS). Holkkola, M., Nyrhinen, J., Makkonen, M., Frank, L., Karjaluoto, H., & Wilska, T.-A. (2022b). Who are the showroomers? Socio-demographic factors behind the showrooming be-havior on mobile devices. In A. Pucihar, M. Kljajić Borštnar, R. Bons, A. Sheombar, G. Ongena, & D. Vidmar (Eds.), Proceedings of the 35th Bled eConference: Digital restructur-ing and human (re)action (pp. 113–128). University of Maribor Press. Hossain, M. S., Zhou, X., & Rahman, M. F. (2018). Examining the impact of QR codes on purchase intention and customer satisfaction on the basis of perceived flow. Interna-tional Journal of Engineering Business Management, 10, 1–10. Kemppainen, T., Frank, L., Makkonen, M., & Hyvönen O.-I. (2021). Barriers to responsible consumption in e-commerce: Evidence from fashion shoppers. In A. Pucihar, M. Klja-jic Borstnar, R. Bons, H. Cripps, A. Sheombar, & D. Vidmar (Eds.), Proceedings of the 34th Bled eConference: Digital support from crisis to progressive change (pp. 327–340). Univer-sity of Maribor Press. Kim, Y. G., & Woo, E. (2016). Consumer acceptance of a quick response (QR) code for the food traceability system: Application of an extended technology acceptance model (TAM). Food Research International, 85, 266–272. Lau, S., Wiedmann, M., & Adalja, A. (2022). Consumer perceptions of QR code technology for enhanced fluid milk shelf-life information provision in a retail setting. JDS Communica-tions, 3(6), 393–397. Lawry, C. A., & Choi, L. (2013). The omnichannel luxury retail experience: Building mobile trust and technology acceptance of quick response (QR) codes. Marketing ZFP, 35(2), 144–154. Lemon, K., & Verhoef, P. (2016). Understanding customer experience throughout the customer journey. Journal of Marketing, 80, 69–96. Liébana-Cabanil as, F., Ramos de Luna, I., & Montoro-Ríos, F. J. (2015). User behaviour in QR mobile payment system: The QR payment acceptance model. Technology Analysis & Strategic Management, 27(9), 1031–1049. Lin, Y.-S., Luo, S.-J., & Chen, B.-Y. (2013). Artistic QR code embel ishment. Computer Graphics Forum, 32(7), 137–146. Mali, E., Paananen, T., Frank, L., & Makkonen, M. (2022). A Customer Perspective on Omni-channel Customer Journey and Channel Usage: A Qualitative Study. In P. Bednar, A. S. Islind, H. Val o-Hult, A. Nolte, M. Rajanen, F. Zaghloul, A. Ravarini, & A. M. Braccini (Eds.), Proceedings of the 8th International Workshop on Socio-Technical Perspective in Information Systems Development (STPIS 2022) (pp. 299-310). RWTH Aachen. CEUR Workshop Proceedings, 3239. Myers, M. D., & Newman, M. (2007). The qualitative interview in IS research: Examining the craft. Information and Organization, 17(1), 2–26. Paananen, T., Kemppainen, T., Frank, L., Holkkola, M., & Mali, E. (2022a). Reinforcement of brand relationships in an omnichannel environment: A qualitative study on clothing shopping. In Proceedings of the 14th Mediterranean Conference on Information Systems (MCIS). Paananen, T., Frank, L., & Kemppainen, T. (2022b). Customer-Brand Relationships in the Context of Digital Brands. In A. Pucihar, M. Kljajić Borštnar, R. Bons, A. Sheombar, G. Ongena, & D. T. Paananen, M. Holkkola, M. Makkonen, L. Frank, T. Kemppainen: Customers’ QR Code Usage Barriers in a Brick-and-Mortar Store: A Qualitative Study 185 Vidmar (Eds.). Proceedings of the 35th Bled eConference: Digital Restructur-ing and Human (Re)action (pp. 495-510). University of Maribor. Pirkkalainen, H., Salo, M., Tarafdar, M., & Makkonen, M. (2019). Deliberate or instinctive? Proactive and reactive coping for technostress. Journal of Management Information Sys-tems, 36(4), 1179–1212. Pucihar, A. (2020). User acceptance of electronic commerce: Contributions from the Bled eConference. Electronic Markets, 30(1), 29–37. Roggeveen, A. L., Nordfält, J., & Grewal, D. (2016). Do digital displays enhance sales? Role of retail format and message content. Journal of Retailing, 92(1), 122–131. Rotsios, K., Konstantoglou, A., Folinas, D., Fotiadis, T., Hatzithomas, L., & Boutsouki, C. (2022). Evaluating the use of QR codes on food products. Sustainability, 14(8). Salo, M., Pirkkalainen, H., and Koskelainen, T. (2019). Technostress and social networking services: Explaining users’ concentration, sleep, identity, and social relation problems. Information Systems Journal, 29(2), 408–435. Shin, D.-H., Jung, J., & Chang, B.-H. (2012). The psychology behind QR codes: User experience perspective. Computers in Human Behavior, 28(4), 1417–1426. Singhal, A., & Pavithr, R. (2015). Degree certificate authentication using QR code and smartphone. International Journal of Computer Applications, 120(16), 38–43. Thomas, D. R. (2003). A general inductive approach for qualitative data analysis. School of Population Health, University of Auckland. Tiwari, S. (2016). An introduction to QR code technology. In Proceedings of the 2016 International Conference on Information Technology (ICIT) (pp. 39–44). IEEE. Trivedi, R., Teichert, T., & Hardeck, D. (2019). Effectiveness of pul -based print advertising with QR codes: Role of consumer involvement and advertisement appeal. European Journal of Marketing, 54(1), 145–167. Tsai, M.-J., & Peng, S.-L. (2023). QR code beautification by instance segmentation (IS-QR). Digital Signal Processing, 133. Xu, M., Hao, S., Yafei, L., Xi, L., Jing, L., Jianwei, N., Pei, L., & Bing, Z. (2019). Stylized aes-thetic QR code. IEEE Transactions on Multimedia, 21(8), 1960–1970. Zhong, Y., & Moon, H.-C. (2022). Investigating customer behavior of using contactless pay-ment in China: A comparative study of facial recognition payment and mobile QR-code payment. Sustainability, 14(12). Zimmermann, H. D. (2016). Digital transformation—The emerging digital economy. In Proceed-ings of International Conference: Liberec Informatics Forum: ICT in the Role of Services: State of the art and perspectives (pp. 138–146). Tech University Liberec. 186 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Appendix A: Examples of QR code tags Figure 2: Promo material of R-Col ection brand (https://www.r-collection.fi/) T. Paananen, M. Holkkola, M. Makkonen, L. Frank, T. Kemppainen: Customers’ QR Code Usage Barriers in a Brick-and-Mortar Store: A Qualitative Study 187 Appendix B: Details about the study’s interview participants Interview Duration Gender Age Employment Status (Min) P1 Female 37 Employed 43 P2 Female 30 Student, employed 53 P3 Female 23 Student 64 P4 Female 40 Employed 84 P5 Female 29 Employed 47 P6 Male 29 Student, employed 65 P7 Male 24 Student, employed 52 P8 Female 33 Employed 58 P9 Female 42 Employed 49 P10 Female 25 Student 40 P11 Male 38 Employed 69 P12 Female 33 Employed 70 P13 Female 59 Employed 69 P14 Female 24 Student 51 P15 Female 32 Employed 65 P16 Female 29 Employed 64 188 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Appendix C: Setting of the experiment in the brick-and-mortar store USING THE APPLE WATCH TO TEACH AND LEARN ABOUT HEART RATE VARIABILITY WHILE VACATIONING ONDREJ MITAS, MARENNA VAN REIJSEN, NADIA CARREIRA OLIVEIRA Breda University of Applied Sciences, Academy for Tourism, Breda, Netherland mitas.o@buas.nl, reijsen.m@buas.nl, oliveira.n@buas.nl Vacations are known to contribute to wel -being. Wearable sensing technologies associated with 'the quantified self' hold promise for designing experiences such as vacations based on evidence to optimize their wel -being outcomes. Tourism organizations may also ask to collect data from customers to help with experience management. The present project shows the potential of heart rate variability data, measured using an Apple Watch, during vacation to produce such insights, and to do so in cooperation with students for the purpose of inspiring them to state-of-the-art research ideas in the context of tourism. Findings show that heart rate variability during vacation appears to follow Keywords: a conversation-of-resources pattern, whereby days with low life tourism, emotions, satisfaction on vacation feature significant increases in heart rate experience, variability. While students gained insights from providing these quantified data and becoming familiar with the Apple Watch, future self, heart iterations of this project may feature students learning to work rate with their data themselves. variability DOI https://doi.org/10.18690/um.fov.4.2023.12 ISBN 978-961-286-751-5 190 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The term the quantified self has been used to refer to the use of personal technologies to digitize and analyze the workings of one’s own body and mind (Lupton, 2016; Swan, 2013). The original impetus for exploring the quantified self was development of biosensing technologies such as smartwatches with photoplethysmogram sensors to measure heart rate and sensors which recorded temperature, location, and movement. There was hope that such data would positively inform individuals wishing to improve their own health and wel -being. A part of life important to wel - being is vacationing (Mitas & Kroesen, 2019). The present contribution applies this logic the use of the Apple Watch to measure six students’ autumn vacation experiences, with two aims: 1) to assess the role of heart rate variability in vacation wel -being outcomes, and 2) to reflect on this project as a way of teaching students about the quantified self. Vacations, defined here as travel away from one’s usual place of residence for leisure purposes of between one night and one year in duration, have demonstrable positive effects for wel -being. We use a subjective wel -being conceptualization in this paper, which means that we define wel -being as a judgment an individual makes about their own life, and that this judgment comprises an affective dimension, operationalized as positive and negative emotions and moods, as wel as a cognitive dimension, operationalized as life satisfaction (Diener & Oishi, 2005). In the short term, vacations produce boosts in positive emotions (de Bloom et al., 2010; Mitas et al., 2012); over a period of several years, vacation frequency is associated with higher life satisfaction as well (Mitas & Kroesen, 2019). The theoretical foundation we use to examine this association comprises the broaden-and-build theory, which holds that brief experiences of positive emotion (which often occur on vacation) accumulate to personal resources, eventually contributing to wel -being as a whole (Fredrickson, 2001); and conservation of resources, which alleges that workaday life drains individuals’ personal resources in a way that only a respite from work––often a vacation––can allow to naturally regenerate (Hobfoll, 2011). Both theories would suggest that vacations contribute to wel -being via increased positive emotions and/or decreased negative emotions. However, while broaden-and-build theory suggests this process would be running across al emotional experiences, conservation of resources theory suggests that O. Mitas, M. van Reijsen, N. Carreira Oliveira: Using the Apple Watch to Teach and Learn About Heart Rate Variability While Vacationing 191 vacation experiences would be at their most powerful when personal resources had been diminished. Vacation experiences have occasional y been studied using the kind of mobile sensing technology that brings to mind the quantified self. For example, wristbands which measure skin conductance have been used to measure peaks of emotional arousal during city walks (Kim & Fesenmaier, 2015; Shoval et al., 2018) and theme park visits (Strijbosch et al., 2021). These measurements can demonstrate, for example, which emotional content of a tour improves an experience, and which degrades it (Mitas, Mitasova, et al., 2020). Skin conductance is an indicator of emotional arousal independent of valence, however, so it cannot indicate whether an experienced emotion peak is positive or negative (Braithwaite et al., 2015). For this purpose, heart rate variability is a more promising metric, though it is not wel understood (Levenson, 2014; Ragot et al., 2017). Unlike skin conductance, which indicates sympathetic arousal in response to emotional stimuli, heart rate variability is associated with the parasympathetic down-regulation of this response. In other words, heart rate variability increases when the mind judges an emotional situation to be non-threatening and begins to return physiological responses to emotion back to baseline levels. Therefore, heart rate variability is associated with well-being over a course of days to weeks (Kok & Fredrickson, 2010). Very few examples of heart rate variability measurement in the context of vacationing exist. Heart rate variability, alongside heart rate and skin conductance, was assessed in a sample of museum visitors in the Netherlands (Mitas, Cuenen, et al., 2020) as well as in a Spanish city walk (i Agustí et al., 2019). The assumption of these studies is that heart rate variability is associated minute-to-minute with emotional valence or, at least, negatively associated with unpleasant emotional stress. While that assumption remains untested, the validity of heart rate variability as a metric of favorable wel -being on a scale of days or weeks is somewhat more widely accepted. It is unknown whether it is associated with the subjective wel -being changes that occur during vacationing. Therefore, in the present study, we used daily measurements of heart rate variability and self-reported subjective wel -being before, during, and after six students’ vacations during the autumn of 2022 to determine if 1) subjective well-being self- 192 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY reports changed during vacation; 2) heart rate variability changed during vacation, and 3) these changes interacted. The choice to work with students as participants was made not only based on convenience, but also served an important educational goal. After providing their data, students were debriefed on study variables and hypotheses, and chal enged to generate research proposals based on wearable sensing technologies such as the Apple Watch as tools answer applied research questions in tourism experience management. Thus, by reflecting on being research participants, students were triggered to design new project ideas with practical implications based on technology that had just been used to measure their experiences. 2 Methods In the present study, we invited six students to borrow an Apple Watch paired with an iPhone using the Apple Health application to monitor heart rate variability at three to five random, recurring 60 second daily intervals. This approach is native within the Apple Health software environment. We asked participants to use the device over a period of three weeks. The second of these weeks was the annual ‘October break’ during which students are free from classes and exams, and many take vacations, including camping or visiting family. Besides recording their heart rate variability, we asked the student-participants to respond to daily questionnaires measuring their positive and negative emotions, life satisfaction, and whether or not they were on vacation. To record emotions using the SPANE instrument (Diener et al., 2010), we asked participants to indicate how strongly they felt each of 8 different emotions ( happy, joyful, content, positive, positively surprised, negative, sad, angry, and afraid) on a 5-point scale from Not at all to Extremely. This emotion list represents a slight adjustment from the original SPANE, in that general and redundant items are omitted, and positive surprise, which is important to tourism experiences, was added. Positive emotion items were averaged together into a positive emotion index, while negative emotion items were averaged together into a negative emotion index. We measured life satisfaction using the almost universal Satisfaction With Life Scale (SWLS; Diener et al., 1985), which comprises five statements such as I am most satisfied with my life and a 7-point Likert-type scale from Strongly disagree to Strongly agree. Instead of using the raw heart rate variability data, we asked the students to report O. Mitas, M. van Reijsen, N. Carreira Oliveira: Using the Apple Watch to Teach and Learn About Heart Rate Variability While Vacationing 193 the daily average generated by Apple Health, in order to keep the data structured at a daily level. In line with our three research questions, we conducted three analyses in a within-persons random intercept framework. It was necessary to use random intercept models as much variation in longitudinal well-being data can be explained by within-participant autocorrelation. Thus, each data point had to be analyzed with respect to the baseline of the participant it came from. With this in mind, we first created a boxcar regressor which was coded 1 for days on vacation and otherwise 0. This was used to predict subjective wel -being and heart rate variability measures. Then, we entered subjective wel -being variables as predictors to the model of heart rate variability as a function of vacation. Finally, we allowed the boxcar regressor for vacationing to interact with subjective well-being variables. 3 Findings Participants reported being slightly happy on average during the data collection period. Means of positive emotions (3.24) and life satisfaction (4.69) were just over the scale midpoints of each respective variable, while negative emotions were just below (2.26). Average daily heart rate was 71.12 beats per minute, while inter-beat intervals varied by a daily average of 59.53 miliseconds, which is how the Apple Watch measures heart rate variability (Table 1). Table 1: Descriptive statistics Response scale Mean Proportion Standard deviation Positive 1 - 5 3.24 0.71 emotions Negative 1 - 5 2.26 0.58 emotions Life 1 - 7 4.69 1.13 satisfaction Heart rate 71.12 35.41 Heart rate 59.53 36.48 variability Days on 0 – 100% 20% vacation 194 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY As expected, being on vacation improved participants’ positive emotions (difference = 0.32 (0.19); p = 0.099) at a marginally significant level, while life satisfaction remained unchanged (difference = 0.04 (0.15); p = 0.781). Their negative emotions were also marginally higher (difference = 0.28 (0.16); p = 0.0743). Heart rate variability was higher on vacation, but the difference was not significant (difference = 6.025 (5.026); p = 0.2344). Day-to-day variation in heart rate variability was also unrelated to all subjective well-being variables (all p’s > 0.3). Table 2: Random intercept model coefficients Outcome variable Predictor(s) Coefficient (Standard error) Model AIC; BIC Positive Being on vacation emotions 3.19 (0.19) * 170.5; 180.1 Negative Being on vacation emotions 0.28 (0.16) * 138.9; 148.4 Life Being on vacation 0.042 (0.153) 139.1; 148.5 satisfaction Heart rate Being on vacation 6.025 (5.026) 707.3; 716.8 variability Heart rate Positive emotions -2.687 (2.939) 707.8; 717.4 variability Heart rate Negative emotions 2.638 (3.595) 708.1; 717.7 variability Heart rate Life satisfaction -3.375 (3.933) 692.8; 702.2 variability Heart rate Positive emotions -2.587 (3.031) 709.2; 723.5 variability Being on vacation 38.806 (35.176) Positive emotions X -9.180 (10.031) Being on vacation Heart rate Negative emotions 3.378 (3.696) 709.0; 723.3 variability Being on vacation 44.243 (28.029) Negative emotions X -13.634 (9.723) Being on vacation Heart rate Life satisfaction -2.299 (3.842) 691.4; 705.5 variability Being on vacation 51.506 (22.686)** Life satisfaction X Being on vacation -8.915 (4.335)** Note: * = p < 0.1; ** p < 0.05 O. Mitas, M. van Reijsen, N. Carreira Oliveira: Using the Apple Watch to Teach and Learn About Heart Rate Variability While Vacationing 195 Heart rate variability was also unrelated to positive and negative emotions when these were al owed to interact with the boxcar regressor representing vacation days (all p’s > 0.1). However, the interaction between being on vacation and life satisfaction significantly predicted heart rate variability (coefficient = -8.915 (4.335); p = 0.0433), over and above a significant positive effect of vacation (difference = 51.505 (22.686); p = 0.0262). In other words, heart rate variability was significantly higher on vacation on days with low life satisfaction. As life satisfaction increased, the positive effect of vacation on heart rate variability decreased, and above a life satisfaction of 5.78––approximately a point higher than the mean in our sample–– the sign of the vacation effect on heart rate variability is modeled to become negative (Figure 1). Figure 1: Simple slopes plot of relationship of vacationing (1 = vacation; 0 = at home), life satisfaction, and heart rate variability 4 Discussion We used Apple Watches to measure six students' heart rate variability before, during, and after the autumn holiday week. We aimed to examine the effects of vacationing on heart rate variability and subjective well-being. Within-individual, between-day analyses showed that vacationing during the holiday week featured slightly improved emotions and, for days with low life satisfaction, improved heart rate variability as wel . These findings have theoretical implications for understanding the effects of 196 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY vacationing on wel-being, as wel as for teaching about the quantified self in the context of tourism. 4.1 Theoretical implications It has been posited that vacations contribute to wel -being by creating positive emotions (Mitas et al., 2012), thus building up life satisfaction over time by 'broadening and building' resources (Fredrickson, 2001; Mitas & Kroesen, 2019). The present data do not represent a time span which can demonstrate effects of positive emotion on life satisfaction, as that normal y takes a number of years, but did show improvement in positive emotions. Thus, our findings are not confirmatory but consistent with this account. Other scholars have demonstrated that vacations improve occupational wel -being outcomes by restoring or conserving attentional and physical resources (Fritz & Sonnentag, 2006). Our finding that life satisfaction negatively moderates the effect of vacationing on heart rate variability are accordingly consistent with conservation of resources theory (Hobfoll, 2011). On days with high life satisfaction, participants did not experience elevated heart rate variability if they were on vacation. Vacationing did raise their heart rate variability on days with low life satisfaction, however. In sum, it is possible that days with low life satisfaction were marked by physical, attentional, or emotional exhaustion. To be on vacation those days, without the demands of school, facilitated a restoration of those resources reflected by elevated heart rate variability. 4.2 Reflections on teaching Students collected data in six groups of four students. We instructed them to consider various ways to record experiences using quantified self approaches which tourism companies can, in turn, ask their customers to share as a source of feedback and quality monitoring. Thus, within each group, only one student had the Apple Watch and reported their heart rate variability. Others recorded their location or considered their photos or social media posts as a source of potential data. As a result of working in groups, with a breadth of data sources in each group, students had a more diverse view on possibilities than if they merely heard about these research methods secondhand. While they were nominal y studying tourism, wherein they considered possible implications for »marketing and advertising efforts,« »offsites or workations for employees,« or »determine how travel packages perform,« O. Mitas, M. van Reijsen, N. Carreira Oliveira: Using the Apple Watch to Teach and Learn About Heart Rate Variability While Vacationing 197 students were also able to consider applications to universities to »give better insights of work/study lifestyle« or »insights in person's life satisfaction…for people that have a lot of pressure at work.« Nevertheless, from a teaching point of view, the present methods could be improved, namely in three ways. First, the sample size could be larger if more devices were available. This would obviously improve the sensitivity of the analyses, but would also give the students the choice to use multiple data streams, for example combining their physiological data together with their photos. Second, some of the students found the process rather unappealing, having to keep devices charged, and wearing during potential y dangerous or appearance-sensitive activities. For example, they asked if we had any other watch bands available so the watches would look better. Also, students emphasized that security of the data, and decoupling location tracking from the physioloigcal and self-response data collection, would be important for the confidentiality of the data. Third, rather than proposing research based on the data they had col ected, students could be taught to inspect and visualize their own data, which would empower them to not only propose but also conduct research based on such devices once working in the travel industry. It is widely appreciated that excel ent vacation experiences contribute to wel -being (Mitas et al., 2017) and that wearable sensing technologies hold promise for designing such experiences based on evidence (Bastiaansen et al., 2019). We acknowledge that the sample size of six students x 21 days is too smal for a robust analysis of the examined phenomena of vacation experiences and wel -being. This study should be seen as a test of data collection and modeling methodology on the way to a larger and more diverse sample which would address these questions more substantially. Despite the smal sample size, the present project shows the potential of heart rate variability measurement during vacation to produce such evidence, and to do so in cooperation with students for the purpose of inspiring them to state-of-the-art research ideas in the context of tourism. References Bastiaansen, M., Lub, X., Mitas, O., Jung, T. H., Passos Acenção, M., Han, D., Moilanen, T., Smit, B., & Strijbosch, W. (2019). Emotions as core building blocks of an experience. International Journal of Contemporary Hospitality Management, 31. Braithwaite, J. J., Watson, D. G., Jones, R., & Rowe, M. (2015). A Guide for Analysing Electrodermal Activity (EDA) & Skin Conductance Responses (SCRs) for Psychological Experiments. 198 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY de Bloom, J., Geurts, S. A., Taris, T. W., Sonnentag, S., de Weerth, C., & Kompier, M. A. (2010). Effects of vacation from work on health and wel -being: Lots of fun, quickly gone. Work & Stress, 24(2), 196-216. Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The satisfaction with life scale. Journal of personality assessment, 49(1), 71-75. Diener, E., & Oishi, S. (2005). The nonobvious social psychology of happiness. Psychological Inquiry, 16(4), 162-167. Diener, E., Wirtz, D., Tov, W., Kim-Prieto, C., Choi, D.-w., Oishi, S., & Biswas-Diener, R. (2010). New wel -being measures: Short scales to assess flourishing and positive and negative feelings. Social Indicators Research, 97(2), 143-156. Fredrickson, B. L. (2001). The role of positive emotions in positive psychology: The broaden-and-build theory of positive emotions. American psychologist, 56(3), 218. Fritz, C., & Sonnentag, S. (2006). Recovery, wel -being, and performance-related outcomes: The role of workload and vacation experiences. Journal of Applied Psychology, 91(4), 936. Hobfol , S. E. (2011). Conservation of resources theory: Its implication for stress, health, and resilience. i Agustí, D. P., Rutl ant, J., & Fortea, J. L. (2019). Differences in the perception of urban space via mental maps and Heart Rate Variation (HRV). Applied Geography, 112, 102084. Kim, J., & Fesenmaier, D. R. (2015). Measuring emotions in real time: Implications for tourism experience design. Journal of Travel Research, 54(4), 419-429. Kok, B. E., & Fredrickson, B. L. (2010). Upward spirals of the heart: Autonomic flexibility, as indexed by vagal tone, reciprocal y and prospectively predicts positive emotions and social connectedness. Biological psychology, 85(3), 432-436. Levenson, R. W. (2014). The autonomic nervous system and emotion. Emotion review, 6(2), 100-112. Lupton, D. (2016). The quantified self. John Wiley & Sons. Mitas, O., Cuenen, R., Bastiaansen, M., Chick, G., & van den Dungen, E. (2020). The War from both Sides: how Dutch and German Visitors Experience an Exhibit of Second World War Stories. International Journal of the Sociology of Leisure, 3(3), 277-303. Mitas, O., & Kroesen, M. (2019). Vacations Over the Years: A Cross-Lagged Panel Analysis of Tourism Experiences and Subjective Wel -Being in the Netherlands. Journal of Happiness Studies, 1-20. Mitas, O., Mitasova, H., Mil ar, G., Boode, W., Neveu, V., Hover, M., van den Eijnden, F., & Bastiaansen, M. (2020). More is not better: The emotional dynamics of an excellent experience. Journal of hospitality & tourism research, 1096348020957075. Mitas, O., Nawijn, J., & Jongsma, B. (2017). Between Tourists: Tourism and Happiness. In M. K. Smith & L. Puczko (Eds.), The Routledge Handbook of Health Tourism (pp. 47-64). Routledge. Mitas, O., Yarnal, C., Adams, R., & Ram, N. (2012). Taking a “peak” at leisure travelers’ positive emotions. Leisure Sciences, 34(2), 115-135. Ragot, M., Martin, N., Em, S., Pal amin, N., & Diverrez, J.-M. (2017). Emotion recognition using physiological signals: laboratory vs. wearable sensors. International Conference on Applied Human Factors and Ergonomics, Shoval, N., Schvimer, Y., & Tamir, M. (2018). Real-time measurement of tourists’ objective and subjective emotions in time and space. Journal of Travel Research, 57(1), 3-16. Strijbosch, W., Mitas, O., van Blaricum, T., Vugts, O., Govers, C., Hover, M., Gelissen, J., & Bastiaansen, M. (2021). Evaluating the Temporal Dynamics of a Structured Experience: Real-Time Skin Conductance and Experience Reconstruction Measures. Leisure Sciences, 1-25. Swan, M. (2013). The quantified self: Fundamental disruption in big data science and biological discovery. Big data, 1(2), 85-99. A VIRTUAL COMPANION FOR LIFELONG LEARNING – DESIGN PRINCIPLES FOR MOTIVATION, SOCIAL LEARNING, AND EXAM PREPARATION PIA GEBBING,1 BIJAN KHOSRAWI-RAD,2 TIMO STROHMANN,2 CHARLOTTE WINDOLF,2 CHRISTOPH LATTEMANN1 1 Constructor University gGmbH, Bremen, Germany pgebbing@constructor.university, clattemann@constructor.university 2 Technische Universität Braunschweig, Germany b.khosrawi-rad@tu-braunschweig.de, t.strohmann@tu-braunschweig.de, c.windolf@tu- braunschweig.de Lifelong vocational learning in a digital context frequently falters due to a lack of motivation, structure, time management, and attention to adult students’ work-life balance. In remote settings, students have further little contact with peers and feel disconnected. This paper answers how a Virtual Learning Companion (VLC) can be designed to address these chal enges and fulfill the specific needs of vocational students. Following a Design Science Research (DSR) approach, a meta-requirement mapping process combines insights from a literature review and ten semi-structured interviews with vocational students. A focus Keywords: group with experts from the field of vocational training, online virtual learning, chatbot design, and DSR evaluated the results. As a companionship, education, result, five Design Principles are presented: (1) Motivational design science Goal Setting, (2) Context and Learner Adaptation, (3) Focus and research, Control, (4) Promoting Resilience, and (5) Enabling Social virtual collaboration, Interactions and Feedback. Exemplary Design Features further Bled il ustrate the VLC development. eConference DOI https://doi.org/10.18690/um.fov.4.2023.13 ISBN 978-961-286-751-5 200 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Lifelong learning is the pursuit of continuous learning and skil development beyond formal education (Cropley & Knapper, 1983). Advanced technologies in distance and hybrid learning environments have opened up new opportunities for lifelong learning. Pedagogical Conversational Agents (PCA) present such a novel technology (Khosrawi-Rad et al., 2022; Wollny et al., 2021) and refer to intelligent dialog systems, e.g., chatbots, which interact with learners using natural language (Gnewuch et al., 2017; Hobert & von Wolff, 2019). A specific form of PCA are Virtual Learning Companions (VLC), designed to facilitate learning and establish a close bond with their users (Greenwald et al., 2016; Strohmann et al., 2022). They are available regardless of time or location and can motivate learners, assist in time management, and foster self-reflection (Gubareva and Lopes, 2020; Khosrawi-Rad, et al., 2022; Wollny et al., 2021). They differ from PCAs by acting relationship-oriented and being designed for long-term use, often inspired by artificial intelligence techniques (Khosrawi-Rad, et al., 2022b). Vocational training in lifelong learning differs from formal education as it involves adults with diverse backgrounds studying alongside their work commitments (Cropley & Knapper, 1983), covering a wider age range, having limited contact with fel ow students, and requiring effective time management between work, family, and studies (Rinn et al., 2022). In vocational training, students must develop self-regulation and acquire structures, skills, habits, knowledge, and attitudes (Cropley & Knapper, 1983; Rinn et al., 2022). To design and develop VLCs for this specific learning context, it is necessary to understand and prioritize the needs and values of learners in vocational training (Rinn et al., 2022). In recent years, research on PCAs has increased (Khosrawi-Rad et al., 2022) and authors proposed design knowledge for various specific application contexts such as argumentative writing support (Wambsganss et al., 2020), avoiding procrastination (Rodriguez et al., 2019), or designing a PCA as a tutor (Winkler & Roos, 2019). However, existing design principles are inadequate for the novel VLC approach and fail to cater to learners in vocational training, highlighting a research gap in VLC design knowledge and facilitating the practical implementation of long-term learning support (Khosrawi-Rad et al., 2022). This study refers to this research gap by answering the research question of how to design VLCs, which address the needs of students in lifelong learning in terms of (1) learning behavior, exam preparation, and time management, P. Gebbing, B. Khosrawi-Rad, T. Strohmann, C. Windolf, C. Lattemann: A Virtual Companion for Lifelong Learning – Design Principles for Motivation, Social Learning, and Exam Preparation 201 (2) motivation, and (3) social learning. To gather design knowledge, we utilize Design Science Research (DSR) as a practical and scientifical y rigorous approach (Hevner et al., 2004). Design principles are derived by integrating kernel theories, VLC design features, and user requirements from ten interviews. 2 Theoretical Grounding This paper aims to highlight and address three specific chal enges related to selfregulated learning in vocational education, which encompasses cognitive, metacognitive, behavioral, motivational, and emotional aspects of learning (Panadero, 2017; Rinn et al., 2022): (1) learning strategy, including exam preparation and time management, (2) motivation and goal setting, and (3) social learning. Table 1 represents a theoretical framework derived from a literature review of kernel theories from psychology which form the theoretical basis of the DP formulation (Kuechler and Vaishnavi, 2008; Möller et al., 2022). Design Features (DF) highlight applicable learning techniques and approaches for implementing and applying kernel theories in a VLC (Möller et al., 2020). Table 1: Conceptualization and References to Theoretical Foundations Learning Strategy, Exam Preparation & Time Management Kernel Cognitivism: Long-, Short-term, and Working Memory (Cowan, 2008); Cognitive Bottleneck Theory Theory (Saxe et al., 2019); Multitasking (Rosen, 2008); Shallow & Deep Knowledge (Bennet & Bennet, 2008; Jackson & Graesser, 2007); Chunking (Miller, 1956; Rosenbloom et al., 1989); Bloom’s Taxonomy (Bloom, 1956); Selective Attention (Johnston & Dark, 1986); Expertise (Gobet, 2019); Transfer and Situated Learning (Greeno et al., 1993); Deliberate Practice (Anders Ericsson, 2008); DF Flipped Classroom (Bergmann & Sams, 2014); Pomodoro Technique (Cirillo, 2018); Microlearning (Hug, 2006; Javorcik & Polasek, 2019); Motivation Kernel Intrinsic and Extrinsic Motivation (Reiss, 2012; Serin, 2017) Flow (Nakamura & Theory Csikszentmihalyi, 2014); Self-Determination (Deci & Ryan, 1980); The Model of Goal Directed Behaviour (Ajzen & Madden, 1986); Goal-setting (Locke & Latham, 2012); Exam-Anxiety (Stöber, 2004); Attribution (Kelley, 1967); Self-directed learning (Merriam & Baumgartner, 2020; Tekkol & Demirel, 2018); Social Identity perspective on motivation (Ellemers et al., 2002; Mirbabaie et al., 2021) DF SMART Goal Setting (Doran, 1981); Rubicon Model of Action Phases (Achtziger & Gollwitzer, 2009); Social Learning KernelT Vygotsky’s Sociocultural Learning Theory (Hall, 2007; Vygotsky, 1978); Social Facilitation heory (Bond & Titus, 1983; Zajonc, 1965); Social Learning (Bandura & Walters, 1977); Theory of Social Comparison (Festinger, 1954; Michinov, 2012); Media Synchronicity Theory (Dennis & Valacich, 1999); DF Three Types of Feedback (Jackson & Graesser, 2007) 202 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 2.1 Learning Strategy, Exam Preparation, and Time Management Cognitivism theories inform the selection of efficient learning strategies and effective learning materials by revealing how humans process information. Effective learning strategies align workloads with an individual's cognitive abilities, facilitating exam preparation and time management. The working memory's limited capacity can suffer from cognitive overload by processing excessive information and reduced accuracy and increased errors, known as "choking" (Saxe et al., 2019; Slonim, 2002; Welford, 1952), as the brain rapidly switches between tasks (Rosen, 2008). Similar performance losses occur in multitasking (Welford, 1952). Breaking down content into small, focused learning units (Rosenbloom et al., 1989) as applied in microlearning (Yin et al., 2021) and learning in short, sequenced intervals, as applied in the pomodoro technique (Cirillo, 2018), foster more sustainable learning outcomes. The depth of processing affects information retention (Graesser et al., 2017). Learning by heart is easier but less sustainable than profound understanding (Bloom, 1956). Content that has been studied briefly and superficial y is stored in short-term memory and quickly forgotten (Cowan, 2008). Deep learning includes mechanisms like repetition and deliberate practice (Anders Ericsson, 2008; Cowan, 2008). Transferring existing knowledge to new situations through associations and integration expands knowledge networks (Greeno et al., 1993). Interactive teaching concepts, like the flipped classroom, promote deep learning as, students act as teachers (Bergmann & Sams, 2014; Huang et al., 2019). Vocational training benefits from connecting new information to existing knowledge, expertise, and experience (Greeno et al., 1993). In summary, a VLC should assist learners in optimizing efficiency, deep learning, and exam performance by aiding in workload planning, organization, and prioritization based on individual capacities and timelines. Incorporating existing learning strategies like micro-learning, the pomodoro technique, or flipped classroom as design features in a VLC can provide advantages such as moderation, feedback, personalization, and active learning. 2.2 Motivation Motivation is gradual and can involve varying levels of extrinsic or intrinsic motivation, although the distinction between the two polarities is fluid and there is a continuum between intrinsic and extrinsic motivation (Reiss, 2012). Extrinsic motivation is determined by external factors such as rewards or punishments, while P. Gebbing, B. Khosrawi-Rad, T. Strohmann, C. Windolf, C. Lattemann: A Virtual Companion for Lifelong Learning – Design Principles for Motivation, Social Learning, and Exam Preparation 203 intrinsic motivation is driven by personal interest and passion for an activity (Serin, 2017). Working on an intrinsically motivating task is the prerequisite for the experience of flow and the complete immersion and absorption in an activity (Nakamura & Csikszentmihalyi, 2014), often seen as the ideal state of learning. Further, a task or activity that elicits a flow experience must be slightly demanding and extend existing capacities (Nakamura and Csikszentmihalyi, 2014). Learners need clear proximal stretching goals and immediate feedback about the progress that is being made (Nakamura and Csikszentmihalyi, 2014). The goal-setting theory confirms that high and specific goals can motivate increased performance (Locke & Latham, 2012). In this context, the SMART Method (Doran, 1981) helps formulate specific, measurable, attainable, realistic, and time-related goals. However, setting adequate goals might not be enough for long-term motivation (Locke & Latham, 2012). The theories of planned behavior (Ajzen, 2011) and goal-directed behavior (Aarts & Elliot, 2011; Ajzen & Madden, 1986) state how beliefs shape intentions and behavior. Doubts, fears, and lack of control are the biggest obstacles to goal achievement (Ajzen & Madden, 1986). Exam anxiety is caused by self-doubts or negative self-efficacy beliefs and increases with pressure, excitement, or stress before the exam (Morris & Liebert, 1970; Stöber, 2004). To counteract negative beliefs, the Self-Determination Theory (Deci & Ryan, 1980) suggests that effective learning fulfills three fundamental psychological needs: competence, autonomy, and relatedness. The feeling of competence can be enhanced by adopting a favorable attribution style that enhances a positive self-image (Kelley, 1967): Attributing positive outcomes and achievements to internal capacities, efforts, and progress, while considering negative results as situational and changeable (e.g. caused by lack of time), can enhance self-confidence. Autonomy can be increased by setting motivational goals that are in line with one’s capacities. Learning should further be self-directed (Tekkol & Demirel, 2018), and learners should be in control of planning, continuing, and evaluating their learning experiences (Merriam & Baumgartner, 2020). Didactic models such as the Rubicon model of action phases (Achtziger & Gollwitzer, 2009), guide the learner through different phases of task accomplishment, i.e. goal setting, planning, realization, and evaluation of goal achievement. To conclude, a VLC should aid learners in enhancing their motivation through the utilization of strategies that promote goal-setting (Locke & Latham, 2012), autonomy, confidence (Deci & Ryan, 1980), and self-efficacy beliefs (Tärning & Silvervarg, 2019). Incorporating established didactical methods such as the SMART-Method for goal setting (Doran, 1981) and the Rubicon model of action 204 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY phases (Achtziger & Golwitzer, 2009) can be beneficial features to foster motivation and flow. The use of a VLC al ows for more meaningful and self-directed learning, enabling learners to apply these strategies independently. 2.3 Social Learning As seen before, learning is highly context-dependent and influenced by the social environment, culture, and task (Hall, 2007; Lave & Wenger, 1991; Vygotsky, 1978). Socio-constructivism suggests learning is best as a shared social rather than an individual experience (Allport, 1920; Rosenberg, 2009). The groups, peers, or social cycles people belong to strongly define their social identity and how they see themselves (Ellemers et al., 2002; Mirbabaie et al., 2021). People learn through observing, modeling, and imitating the behaviors, attitudes, and emotional reactions of others (Bandura & Walters, 1977). Learners form an image of their self and competencies through social comparison with relevant others, for example, fellow students, teachers, or role models (Festinger, 1954). Yet, in remote learning, receiving adequate feedback is often an unmet need (Rinn et al., 2022). It is difficult to foster natural social exchange (Kock, 2005). Even rich media cannot ensure optimal collaboration, as the communication channel must be adapted to the task (Dennis et al., 2008; Dennis & Valacich, 1999). Thereby three types of feedback can be distinguished (Jackson & Graesser, 2007): First, task-based feedback gives practical and domain-specific advice on how to complete an activity and avoid mistakes; second, progress-based feedback evaluates the overall learning progress; and third empathic feedback offers emotional support. In digital settings, task-based feedback can be provided in a fast, synchronous way, helping learners to identify errors and proceed (Jung et al., 2010). Students further benefit from recoding common questions of other students (Dennis et al., 2008). Concerning emotional or procedural feedback, findings suggest a positive effect of exchanges on social media in knowledge sharing and building trust among peers (Cao et al., 2012). Yet not all learners are comfortable sharing their emotions and data online (Atske, 2021). To conclude, a VLC should provide feedback to support a positive self-image and competence. While automated scoring and task-based feedback are common practices, the provision of emotional support is stil in its nascent stages. P. Gebbing, B. Khosrawi-Rad, T. Strohmann, C. Windolf, C. Lattemann: A Virtual Companion for Lifelong Learning – Design Principles for Motivation, Social Learning, and Exam Preparation 205 3 Method Design Principles (DPs) contribute to design knowledge and guide effective and innovative conceptual design at a meta-level, (Gregor et al., 2020; Möller et al., 2020). Following the DSR approach, the research plan follows six steps for formulating DPs (Möller et al., 2020): (1) Definition of a practical y relevant problem, formulated as a research question (see introduction section). (2) Identification of kernel theories as a justificatory knowledge base (see theory section). (3) Development of User Stories (US) from a thematic analysis of semi-structured interviews with students from three vocational learning institutions in Germany, promoting a user-centered approach to artifact development (Abras et al., 2004). (4) Meta-requirement mapping, clustering similar US into distinct meta-requirements. (5) Formulation of preliminary DPs, based on the meta-requirements and kernel theories, following a predefined framework from Gregor et al. (2020). Suggestions for DP implementation through exemplary design features (DFs) are derived from both the interviewees' suggestions and the literature. (6) Evaluation of the DPs in an expert focus group. Ten semi-structured interviews were conducted with vocational education students from three German institutes offering hybrid or online courses, selected through a systematic stratified sampling approach to representing various demographic backgrounds (Gläser & Laudel, 2010). Six men and four women aged 22 to 50 were interviewed. The structured interview guideline provided deeper insights into seven themes identified in a prior quantitative survey (Rinn et al., 2022): (1) respondent's life and (2) learning situation, (3) methods and techniques used, (4) learner's motivation, (5) social learning, (6) time management, and (7) learning success and learning challenges. Open-ended questions were used to generate creative, original, and atypical answers. Final y, the interviewees were asked to draw conclusions based on their answers about the design of a VLC. The interviews were conducted by two independent researchers (Dec 2021-Mar 2022) and lasted from 33 to 84 minutes. Interviews were voice recorded and transcribed using the software Happy Scribe. Personal data was anonymized. A qualitative thematic content analysis (Mayring, 2015) of the ten transcribed interviews was carried out with the software MAXQDA 2020. The coding scheme was deductively generated and refined based on the interview guideline and participants' statements. The analysis yielded 844 codes, highlighting the top three codes for each theme. The findings were synthesized in 206 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 36 user stories (US), illustrating a specific user's needs and preliminary requirements for VLC-supported learning. US were formulated as follows: "As , I want , so ." Throughout a meta-requirement mapping process, four researchers used an adapted 1-2-4-al method (Lipmanowicz & McCandless, 2014) to link user stories and kernel theories and create a consolidated set of meta-requirements. Researchers independently clustered a subset of meta-requirements, which were then peer-reviewed and finally discussed and consolidated by all researchers. Based on these meta-requirements and prescriptive knowledge derived from the literature, a set of preliminary DPs was formulated following a predefined framework from Gregor et al. (2020), specifying the aim, implementor and user, context, underlying mechanisms and rationale for each DP (see the digital appendix: https://bit.ly/eBled23). The DPs were evaluated by a focus group of five experts from different domains (online- and remote teaching and learning, chatbot design, DSR). The DPs and prototypical examples of DF were presented to the experts using the online collaboration platform Miro. Following the presentation of each DP, participants individual y completed an online survey based on the light reusability framework (Iivari et al., 2018), assessing five factors: accessibility, importance, novelty and insightfulness, actability and guidance, and effectiveness. The effectiveness of each DP in addressing the targeted construct was measured using a reflective self-assessment. (e.g., “Compared to my current situation, I believe that a VLC that incorporates this DP would improve motivation.”). After the survey completion, each DP was openly discussed in the focus group and qualitative feedback was recorded. 4 Artifact Description The meta-requirement mapping of the kernel theories and interview results were summarized in five DPs for VLCs in digital teaching, which are explained below. P. Gebbing, B. Khosrawi-Rad, T. Strohmann, C. Windolf, C. Lattemann: A Virtual Companion for Lifelong Learning – Design Principles for Motivation, Social Learning, and Exam Preparation 207 4.1 Motivational Goal Setting (DP1) The goal of DP1 is to design a VLC which promotes motivational goal setting for the learner in the planning stage of the learning process (Achtziger & Gol witzer, 2009) to build up competence awareness, autonomy (Deci & Ryan, 1980), self-efficacy (Ajzen & Madden, 1986), and motivation (Locke & Latham, 2012; Nakamura & Csikszentmihalyi, 2014). Boundary conditions are addressed in DP5, namely the need for adaptation to the learner's individual needs and a realistic self-assessment. To manage time and tasks appropriately constraints and workload need to be anticipated (see DP3). The following mechanisms ensure motivational goal setting: (1) Set long-term goals that inspire the learner and that reflect personal development, skill acquisition, and career opportunities (Hall, 2007; Vygotsky, 1978). (2) Breaking down tasks into short-term learning goals that are specific, challenging, and attainable (Doran, 1981; Locke & Latham, 2012). (3) Constantly evaluate the progress towards a learning goal to increase the learners’ autonomy (Achtziger & Gollwitzer, 2009; Deci & Ryan, 1980). (4) Increase the perception of competence, and self-efficacy by celebrating success, goal achievement, and chal enges (Deci & Ryan, 1980; Kim, 2001). The following DFs support these mechanisms: The SMART concept (Doran, 1981) helps to formulate motivational goals. The Rubicon model (Achtziger & Gollwitzer, 2009) is a framework for planning, tracking, and evaluating goal achievement and milestones. Goals and progress can be visibly displayed, e.g., in the form of a personal mission statement, progress bar, or success record. 4.2 Enabling Social Interactions and Feedback (DP2) The goal of DP2 is to provide valuable feedback and social support to the learner while performing a task, especially during the progress evaluation phase (Hall, 2007; Vygotsky, 1978). Feedback should be provided by peers and instructors on three levels to promote the perception of competence and social relatedness (Jackson & Graesser, 2007): (1) Task-related feedback, (2) empathic feedback for social connection, and (3) procedural feedback on long-term goals. Further, feedback- givers must be matched according to the needs and experience of the learner to foster a motivational, upward comparison (Festinger, 1954). Task-related feedback should be provided by knowledgeable colleagues or instructors through a rich, low-threshold, and highly synchronized communication channel (Dennis & Valacich, 208 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1999). Procedural feedback should be given by experts in the field who have more expertise and experience than the learner. Different channels should be used: Personal networks, such as Instagram, are better for reaching out to friends and family and getting empathetic feedback. Professional networks, such as LinkedIn, al ow networking with experts for procedural feedback. Task-based feedback and the ability to interact and discuss with instructors and peers should be available via a direct, built-in chat feature. 4.3 Focus and Control (DP3) DP3 aims to foster focus, concentration and perceived control while limiting disruption to al ow an experience of flow (Nakamura & Csikszentmihalyi, 2014), throughout different work phases (Achtziger & Gollwitzer, 2009). The following mechanisms are suggested help to avoid cognitive overload and enhance concentration (Saxe et al., 2019): The VLC should help divide content into smal er work packages aligned with the learning goal (Aarts & Elliot, 2011), and individual circumstances (Miller, 1956; Rosenbloom et al., 1989), e.g., through time-boxing or micro-learning (Hug, 2006). The VLC can further instruct learners on coping strategies for managing external factors, such as deadlines and unforeseen events, to improve their sense of control and competence. 4.4 Fostering Resilience (DP4) The goal of DP4 is to equip the learner with coping strategies for stress and exam anxiety (Stöber, 2004). The aim is to foster the learner's sense of competence, autonomy, and resilience (Deci & Ryan, 1980; Stöber, 2004). Since learning difficulties are a personal matter, it is necessary that the user accepts the VLC as a mentor and strives for long-term interaction. The following underlying mechanisms can help foster resilience: (1) Recording the learner's mental state and stress level through the collection of personal and behavioral data, (2) Providing an initial assessment and suggesting professional contacts if needed. A VLC cannot replace specialist intervention for severe psychological symptoms, (3) Suggestion of personalized interventions based on learner's perceived competence, agency, and self-efficacy beliefs. (4) The VLC supports planning by estimating the required time for preparation based on historical data. P. Gebbing, B. Khosrawi-Rad, T. Strohmann, C. Windolf, C. Lattemann: A Virtual Companion for Lifelong Learning – Design Principles for Motivation, Social Learning, and Exam Preparation 209 4.5 Context- and Learner Adaptation (DP5) DP5 aims to provide individualized support to the learner throughout the entire learning process. The VLC should collect information about individual needs, preferences, and circumstances. Further, the VLC should have an extensive knowledge base about the learning content to determine the right level for examination protocols and learning time (see also DP3). The VLC should provide a transparent communication of results and provision of feedback about learning behavior (see also DP4), and should suggest learning materials and methods based on learners’ preferences and learning progress. However, such an approach needs the user’s consent before implementing any adaptations. 5 Evaluation A focus group of five experts assessed the DPs. All DPs were deemed understandable and accessible, indicating no misinterpretation due to non- comprehension. The experts further differentiated between "must-have” and "nice-to-haves" DPs, which would be useful but less urgent. The focus group unanimously chose Motivational Goal Setting (DP1) as the most relevant DP. The experts found that the SMART-goal-setting method, including the visualization of goals, was an effective feature. However, limitations were noted in the actability and responsiveness of the VLC to the learner. DP2, “focus and flow”, was rated as the second priority. There were controversial opinions about the exemplary design feature, the pomodoro technique. A pedagogical expert noted that time restrictions may disrupt the flow. Developers suggested improving the user experience of VLC by pooling different methods to enhance flow and focus into a single, multi-device interface. Context- and learner adaptation (DP3) was rated as “relatively” important. One participant said that this DP is highly relevant because it best expresses what makes the VLC a true "companion". Yet, the complex, individualized design may be overwhelming for inexperienced learners, who may need time to learn how to use it effectively. Therefore, a "basic" and a "pro" version are proposed to provide suitable guidance and actability. Overall, the focus group viewed the ability to facilitate social interaction and feedback (DP4) as beneficial but not essential. The VLC should inquire if learners are wil ing and able to connect and provide support to their peers. Most experts valued feedback and comparisons to other students. Controversy surrounds the involvement of external feedback givers for emotional and procedural 210 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY feedback. The option to disable social media usage is crucial to address privacy concerns related to sharing information about learners' progress. The importance of stress reduction and the promotion of resilience and wel -being through a VLC (DP5) was discussed, with some controversy over the provision of in-depth psychological care. Experts agreed on the relevance of stress reduction, but some argued that in-depth psychological care might be beyond the competence and objectives of a VLC. The experts suggested that it may be more feasible for a VLC to focus on preventive measures, such as improving time management skil s and promoting breaks, physical activity, and social contact. The VLC could track and analyze behavioral data to assess workload and support healthy habits, as long as data protection is ensured. 6 Discussion In recent years, research on PCA has increased (Khosrawi-Rad et al., 2022) and researchers contributed design knowledge on PCAs in different application contexts (Rodriguez et al., 2019; Wambsganss et al., 2020; Winkler & Roos, 2019). We extend this knowledge to vocational training by user-centered DPs for VLCs. The DPs were derived from user interviews and expert evaluations, highlighting its relevance and validity to the target group. However, the missing instantiation is a limitation that future studies should further evaluate. Based on the evaluation of the DPs, a VLC should address the diverse needs of individual learners, such as their learning strategies, exam preparation, time management, motivation, and social learning. Each identified kernel theory (see Table 1) contributes to the understanding and description of selected learners’ requirements. However, our research indicates a need for more transdisciplinary research as developing complex information systems like VLCs incorporates insights from diverse fields like cognitive and social psychology and pedagogy. Further, we identified 23 scientific concepts (see Table 1), which help to extract concrete applications, such as methods or guidelines, and are identified as design features (DF). The results of the evaluation of the DPs show that some of the identified concepts are very helpful, such as the formulation and visualization of SMART goals (Doran, 1981) and the provision of feedback on three levels (Jackson & Graesser, 2007). However, some aspects are controversial y discussed, e.g., the pomodoro technique (Cirillo, 2018), and privacy issues. P. Gebbing, B. Khosrawi-Rad, T. Strohmann, C. Windolf, C. Lattemann: A Virtual Companion for Lifelong Learning – Design Principles for Motivation, Social Learning, and Exam Preparation 211 7 Conclusion The transformative power of technology in lifelong vocational education cannot be denied. As distance and hybrid learning become more prevalent, the need for personalized support increases. This paper presents a set of design principles for Virtual Learning Companions (VLCs) that can promote resilience, goal setting, content focus, and control over the learning process while providing tasked-based feedback to address the lack of structure, time management, and work-life balance support. Additionally, a VLC can address the issues of disconnection, lack of peer contact, and motivation commonly experienced by adult students, by facilitating social interaction and emotional feedback. The results contribute to the research stream on digital lifelong vocational education and offer practical insights for developers and providers of learning platforms, ultimately leading to better learning outcomes and success for adult learners. Acknowledgments: This research was funded by the German Federal Ministry of Education and Research (BMBF) under the project Study Buddy (Grant # 21INVI06). References Aarts, H., & El iot, A. (Eds.). (2011). Goal-Directed Behavior. Psychology Press. https://doi.org/10.4324/9780203869666 Abras, C., Maloney-Krichmar, D., & Preece, J. (2004). User-centered design. Bainbridge, W. Encyclopedia of Human-Computer Interaction. Thousand Oaks: Sage Publications, 37(4), 445–456. Achtziger, A., & Gol witzer, P. (2009). Rubikonmodell der Handlungsphasen. In V. Brandstätter (Ed.), Handbuch der Al gemeien Psychologie: Motivation und Emotion. https://kops.uni-konstanz.de/handle/123456789/17176 Ajzen, I. (2011). The theory of planned behaviour: Reactions and reflections. Psychology & Health, 26(9), 1113–1127. https://doi.org/10.1080/08870446.2011.613995 Ajzen, I., & Madden, T. J. (1986). Prediction of goal-directed behavior: Attitudes, intentions, and perceived behavioral control. Journal of Experimental Social Psychology, 22(5), 453–474. https://doi.org/10.1016/0022-1031(86)90045-4 Al port, F. H. (1920). The influence of the group upon association and thought. Journal of Experimental Psychology, 3, 159–182. https://doi.org/10.1037/h0067891 Anders Ericsson, K. (2008). Deliberate Practice and Acquisition of Expert Performance: A General Overview. Academic Emergency Medicine, 15(11), 988–994. https://doi.org/10.1111/j.1553-2712.2008.00227.x Atske, S. (2021, April 7). Social Media Use in 2021. Pew Research Center: Internet, Science & Tech. https://www.pewresearch.org/internet/2021/04/07/social-media-use-in-2021/ Bandura, A., & Walters, R. H. (1977). Social learning theory (Vol. 1). Englewood Cliffs Prentice Hal . Bennet, D., & Bennet, A. (2008). The depth of knowledge: Surface, shal ow or deep? VINE, 38(4), 405–420. https://doi.org/10.1108/03055720810917679 212 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Bergmann, J., & Sams, A. (2014). Flipped learning: Gateway to student engagement. International Society for Technology in Education. Bloom, B. S. (1956). Learning domains or Bloom’s taxonomy. In Handbook: The Cognitive Domain. David McKay. Bond, C. F., & Titus, L. J. (1983). Social facilitation: A meta-analysis of 241 studies. Psychological Bul etin, 94(2), 265–292. https://doi.org/10.1037/0033-2909.94.2.265 Cao, X., Vogel, D. R., Guo, X., Liu, H., & Gu, J. (2012). Understanding the Influence of Social Media in the Workplace: An Integration of Media Synchronicity and Social Capital Theories. 45th Hawai International Conference on System Sciences, 3938–3947. https://doi.org/10.1109/HICSS.2012.618 Cirillo, F. (2018). The Pomodoro Technique: The Life-Changing Time-Management System. Random House. Cowan, N. (2008). What are the differences between long-term, short-term, and working memory? Progress in Brain Research, 169, 323–338. https://doi.org/10.1016/S0079-6123(07)00020-9 Cropley, A. J., & Knapper, C. K. (1983). Higher education and the promotion of lifelong learning. Studies in Higher Education, 8(1), 15–21. https://doi.org/10.1080/03075078312331379081 Deci, E. L., & Ryan, R. M. (1980). Self-determination Theory: When Mind Mediates Behavior. The Journal of Mind and Behavior, 1(1), 33–43. Dennis, A. R., Ful er, R. M., & Valacich, J. S. (2008). Media, Tasks, and Communication Processes: A Theory of Media Synchronicity. MIS Quarterly, 32(3), 575–600. https://doi.org/10.2307/25148857 Dennis, A. R., & Valacich, J. S. (1999). Rethinking media richness: Towards a theory of media synchronicity. Proceedings of the 32nd Annual Hawai International Conference on Systems Sciences, 10-pp. Doran, G. T. (1981). There’s a SMART way to write management’s goals and objectives. Management Review, 35–36. El emers, N., Spears, R., & Doosje, B. (2002). Self and social identity. Annual Review of Psychology, 53(1), 161–186. Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7(2), 117–140. Gläser, J., & Laudel, G. (2010). Experterninterviews und qualitative Inhaltsanalyse als Instrumente rekonstruierender Untersuchungen (4. Auflage). VS Verlag für Sozialwissenschaften. Gnewuch, U., Morana, S., & Maedche, A. (2017). Towards Designing Cooperative and Social Conversational Agents for Customer Service. Proceedings of the 38th International Conference on Information Systems. Proceedings of the 8th International Conference on Information Systems (ICIS). Gobet, F. (2019). The Classic Expertise Approach and Its Evolution. In P. Ward, J. Maarten Schraagen, J. Gore, & E. M. Roth (Eds.), The Oxford Handbook of Expertise (pp. 35–55). Oxford University Press. https://doi.org/10.1093/oxfordhb/9780198795872.013.2 Graesser, A.C., Lippert, A.M., Hampton, A.J. (2017). Successes and Failures in Building Learning Environments to Promote Deep Learning: The Value of Conversational Agents. In: Buder, J., Hesse, F. (eds) Informational Environments. Springer, Cham. https://doi.org/10.1007/9783-319-64274-1_12 Greeno, J. G., Moore, J. L., & Smith, D. R. (1993). Transfer of situated learning. Greenwald, S. W., Funk, M., Loreti, L., Mayo, D., & Maes, P. (2016). EVA: Exploratory Learning with Virtual Companions Sharing Attention and Context. 2016 IEEE 16th International Conference on Advanced Learning Technologies (ICALT), 26–30. https://doi.org/10.1109/ICALT.2016.56 Gregor, S., Chandra Kruse, L., & Seidel, S. (2020). The Anatomy of a Design Principle. Journal of the Association for Information Systems, 21, 1622–1652. https://doi.org/10.17705/1jais.00649 Gubareva, R., & Lopes, R. P. (2020). Virtual assistants for learning: A systematic literature review. CSEDU 2020 - Proceedings of the 12th International Conference on Computer Supported Education, 1, 97– 103. Scopus. Hall, A. (2007). Vygotsky goes online: Learning design from a socio-cultural perspective. In Learning and socio-cultural Theory: Exploring Modern Vygotskian Perspectives international workshop 2007 (Vol. 1, No. 1, p. 6). P. Gebbing, B. Khosrawi-Rad, T. Strohmann, C. Windolf, C. Lattemann: A Virtual Companion for Lifelong Learning – Design Principles for Motivation, Social Learning, and Exam Preparation 213 Hevner, A. R. (2021). The Duality of Science: Knowledge in Information Systems Research. Journal of Information Technology, 36(1), 72–76. https://doi.org/10.1177/0268396220945714 Hevner, A. R., March, S. T., Park, J., & Ram, S. (2004). Design Science in Information Systems Research. MIS Quarterly, 28(1), 75–105. https://doi.org/10.2307/25148625 Hobert, S., & von Wolff, R. M. von. (2019). Say Hel o to Your New Automated Tutor – A Structured Literature Review on Pedagogical Conversational Agents. Wirtschaftsinformatik 2019 Proceedings. Huang, W., Hew, K. F., & Gonda, D. E. (2019). Designing and evaluating three chatbot-enhanced activities for a flipped graduate course. International Journal of Mechanical Engineering and Robotics Research, 8(5), 813–818. Scopus. https://doi.org/10.18178/ijmerr.8.5.813-818 Hug, T. (2006). Microlearning: A New Pedagogical Challenge (Introductory Note). In T. Hug, M. Lindner, & P. A. Bruck, (Eds.), Microlearning: Emerging Concepts, Practices, and Technologies After E-Learning: Proceedings of Microlearning Conference 2005: Learning & Working in New Media (pp. 8-11). Innsbruck, áustria: Innsbruck University Press. Iivari, J., Hansen, M. R. P., & Haj-Bolouri, A. (2018). A Framework for Light Reusability Evaluation of Design Principles in Design Science Research. Proceedings of the 13th International Conference on DESRIST. International Conference on Design Science Research and Information Systems and Technology. Jackson, G. T., & Graesser, A. C. (2007). Content matters: An investigation of feedback categories within an ITS. Frontiers in Artificial Intel igence and Applications, 158, 127. Javorcik, T., & Polasek, R. (2019). Transformation of e-learning into microlearning: New approach to course design. AIP Conference Proceedings, 2116(1), 060016. Johnston, W. A., & Dark, V. J. (1986). Selective Attention. Annual Review of Psychology, 37(1), 43–75. https://doi.org/10.1146/annurev.ps.37.020186.000355 Jung, J. H., Schneider, C., & Valacich, J. (2010). Enhancing the motivational affordance of information systems: The effects of real-time performance feedback and goal setting in group col aboration environments. Management Science, 56(4), 724–742. Kelley, H. H. (1967). Attribution theory in social psychology. Nebraska Symposium on Motivation, 15, 192– 238. Khosrawi-Rad, B., Rinn, H., Schlimbach, R., Gebbing, P., Yang, X., Lattemann, C., Markgraf, D., & Robra-Bissantz, S. (2022). Conversational Agents in Education – A Systematic Literature Review. ECIS 2022 Conference Proceedings. Khosrawi-Rad, B., Schlimbach, R., Strohmann, T., & Robra-Bissantz, S. (2022). Design Knowledge for Virtual Learning Companions. Proceedings on the International Conference on Information Systems Education and Research. Kim, B. (2001). Social constructivism. Emerging Perspectives on Learning, Teaching, and Technology, 1(1), 16. Kock, N. (2005). Media richness or media naturalness? The evolution of our biological communication apparatus and its influence on our behavior toward E-communication tools. IEEE Transactions on Professional Communication, 48(2), 117–130. https://doi.org/10.1109/TPC.2005.849649 Kuechler, B., & Vaishnavi, V. (2008). On theory development in design science research: Anatomy of a research project. European Journal of Information Systems, 17(5), 489–504. Lave, J., & Wenger, E. (1991). Situated Learning: Legitimate Peripheral Participation. Cambridge University Press. Lipmanowicz, H., & McCandless, K. (2014). The Surprising Power of Liberating Structures: Simple Rules to Unleash A Culture of Innovation (1st edition). Liberating Structures Press. Locke, E. A., & Latham, G. P. (2012). Goal Setting Theory: The Current State. In New Developments in Goal Setting and Task Performance. Routledge. Mayring, P. (2015). Qualitative Inhaltsanalyse: Grundlagen und Techniken (12., überarbeitete Auflage). Beltz Verlag. Merriam, S. B., & Baumgartner, L. M. (2020). Learning in Adulthood: A Comprehensive Guide. John Wiley & Sons. Michinov, N. (2012). Is electronic brainstorming or brainwriting the best way to improve creative performance in groups? An overlooked comparison of two idea-generation techniques. Journal of Applied Social Psychology, 42, E222–E243. 214 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Miler, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63(2), 81. Mirbabaie, M., Stieglitz, S., Brünker, F., Hofeditz, L., Ross, B., & Frick, N. (2021). Understanding Col aboration with Virtual Assistants – The Role of Social Identity and the Extended Self. Business & Information Systems Engineering, 63(1), 21–37. Möl er, F., Guggenberger, T. M., & Otto, B. (2020). Towards a Method for Design Principle Development in Information Systems. In S. Hofmann, O. Mül er, & M. Rossi (Eds.), Designing for Digital Transformation. Co-Creating Services with Citizens and Industry (pp. 208–220). Springer International Publishing. https://doi.org/10.1007/978-3-030-64823-7_20 Möller, F., Schoormann, T., Strobel, G., & Hansen, M. (2022, December 1). Unveiling the Cloak: Kernel Theory Use in Design Science Research. ICIS 2022 Proceedings. https://aisel.aisnet.org/icis2022/adv_methods/adv_methods/2 Morris, L. W., & Liebert, R. M. (1970). Relationship of cognitive and emotional components of test anxiety to physiological arousal and academic performance. Journal of Consulting and Clinical Psychology, 35, 332–337. https://doi.org/10.1037/h0030132 Nakamura, J., & Csikszentmihalyi, M. (2014). The Concept of Flow. In M. Csikszentmihalyi (Ed.), Flow and the Foundations of Positive Psychology: The Col ected Works of Mihaly Csikszentmihalyi (pp. 239– 263). Springer Netherlands. https://doi.org/10.1007/978-94-017-9088-8_16 Panadero, E. (2017). A Review of Self-regulated Learning: Six Models and Four Directions for Research. Frontiers in Psychology, 8:422. doi: 10.3389/fpsyg.2017.00422 Reiss, S. (2012). Intrinsic and Extrinsic Motivation. Teaching of Psychology, 39(2), 152–156. https://doi.org/10.1177/0098628312437704 Rinn, H., Khosrawi-Rad, B., Schlimbach, R., Masurek, M., Robra-Bissantz, S., & Markgraf, D. (2022). Needs of Students in Further Education -A Mixed Methods Study. Gemeinschaft Neuer Medien (GeNeMe) 2022. Rodriguez, J., Piccoli, G., & Bartosiak, M. (2019). Nudging the Classroom: Designing a Socio-Technical Artifact to Reduce Academic Procrastination. Hawai International Conference on System Sciences. https://doi.org/10.24251/HICSS.2019.533 Rosen, C. (2008). The myth of multitasking. The New Atlantis, 20, 105–110. Rosenberg, M. J. (2009). The Conditions and Consequences of Evaluation Apprehension. In R. Rosenthal, R. L. Rosnow, & Alan E. Kazdin (Eds.), Artifacts in Behavioral Research (pp. 211–263). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195385540.003.0007 Rosenbloom, P. S., Laird, J. E., & Newell, A. (1989). The chunking of skill and knowledge. Working Models of Human Perception, 391–410. Saxe, A. M., Bansal, Y., Dapel o, J., Advani, M., Kolchinsky, A., Tracey, B. D., & Cox, D. D. (2019). On the information bottleneck theory of deep learning. Journal of Statistical Mechanics: Theory and Experiment, 2019(12), 124020. Serin, H. (2017). The Role of Passion in Learning and Teaching. International Journal of Social Sciences & Educational Studies, 4(1), Article 1. Slonim, N. (2002). The information bottleneck: Theory and applications [Ph.D. Thesis]. Hebrew University of Jerusalem, Israel. Stöber, J. (2004). Dimensions of test anxiety: Relations to ways of coping with pre-exam anxiety and uncertainty. Anxiety, Stress & Coping, 17(3), 213–226. Strohmann, T., Siemon, D., Khosrawi-Rad, B., & Robra-Bissantz, S. (2022). Toward a design theory for virtual companionship. Human–Computer Interaction, 1–41. https://doi.org/10.1080/07370024.2022.2084620 Tärning, B., & Silvervarg, A. (2019). “I Didn’t Understand, I´m Really Not Very Smart”—How Design of a Digital Tutee’s Self-Efficacy Affects Conversation and Student Behavior in a Digital Math Game. Education Sciences, 9(3), Article 3. https://doi.org/10.3390/educsci9030197 Tekkol, İ. A., & Demirel, M. (2018). An Investigation of Self-Directed Learning Skills of Undergraduate Students. Frontiers in Psychology, 0. https://doi.org/10.3389/fpsyg.2018.02324 Vygotsky, L. S. (1978). Mind in Society: Development of Higher Psychological Processes. Harvard University Press. https://doi.org/10.2307/j.ctvjf9vz4 P. Gebbing, B. Khosrawi-Rad, T. Strohmann, C. Windolf, C. Lattemann: A Virtual Companion for Lifelong Learning – Design Principles for Motivation, Social Learning, and Exam Preparation 215 Wambsganss, T., Sölner, M., & Leimeister, J. M. (2020, December 13). Design and Evaluation of an Adaptive Dialog-Based Tutoring System for Argumentation Skil s. International Conference on Information Systems (ICIS). International Conference on Information Systems (ICIS), Hyderabad, India. https://www.alexandria.unisg.ch/261297/ Welford, A. T. (1952). The “psychological refractory period” and the timing of high-speed performance—A review and a theory. British Journal of Psychology, 43, 2–19. Winkler, R., & Roos, J. (2019). Bringing AI into the Classroom: Designing Smart Personal Assistants as Learning Tutors. ICIS 2019 Proceedings. https://aisel.aisnet.org/icis2019/learning_environ/learning_environ/10 Wol ny, S., Schneider, J., Di Mitri, D., Weidlich, J., Rittberger, M., & Drachsler, H. (2021). Are We There Yet? - A Systematic Literature Review on Chatbots in Education. Frontiers in Artificial Intel igence, 4, 654924. https://doi.org/10.3389/frai.2021.654924 Yin, J., Goh, T.-T., Yang, B., & Xiaobin, Y. (2021). Conversation Technology With Micro-Learning: The Impact of Chatbot-Based Learning on Students’ Learning Motivation and Performance. Journal of Educational Computing Research, 59(1), 154–177. https://doi.org/10.1177/0735633120952067 Zajonc, R. B. (1965). Social Facilitation: A solution is suggested for an old unresolved social psychological problem. Science, 149(3681), 269-274. 216 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY EMPIRICAL VERIFICATION OF DIFFERENT RANK DYNAMIC ROLES IN INFORMAL HIERARCHIES PETER VATTER,1 ANDREAS WEISBECK2 1 FOM University of Applied Sciences for Economy and Management, Institute for Work and Human Resources (iap), Nuremberg, Germany peter.vatter@fom.de 2 FOM University of Applied Sciences for Economy and Management, Nuremberg, Germany andreas.weisbeck@fom-net.de Organizations today face a more volatile, uncertain, complex and ambiguous market environment, referred to by the term »VUCA world«. To this, they commonly respond with management approaches that comprise flat hierarchies and a decentralization of decision-making structures. But while some companies are very successful with such approaches, others struggle significantly. One reason is that conflicts about ranks in the informal hierarchy arise as the formal hierarchy diminishes. Hence, understanding group dynamics and the different strategies of team members in the social competition becomes increasingly important. In previous works, a theoretical model Keywords: new work, containing five different roles was presented. These roles result working from specific psychological capabilities, values, and needs of culture, each individual and include specific contributions to the social group dynamics, functioning of the group. In this paper, participants of an online informal survey have been asked about their contribution to team success. hierarchy, A subsequent factor analysis of n = 421 datasets proofed the rank dynamics, existence of five different roles and, thus, the validity of the leadership, model. status DOI https://doi.org/10.18690/um.fov.4.2023.14 ISBN 978-961-286-751-5 218 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 The History of Leadership and Group Dynamics Research After World War II, research on sociology and psychology was highly motivated by recent experiences and focused on explaining the phenomena of the war. Against this background, Milgram published his works on obedience which were broadly discussed not only in the scientific community (1963). At the same time, Schindler worked on the »bifocal group therapy« for schizophrenic patients (Schindler 2016, Spal er 2018). An important component were his observations of the »rank order positions« within the group, which was how he found the negative influence of authority on the psychological wel -being of the patients. As a byproduct he developed the rank dynamic model, becoming the epitome of group dynamic research. Nevertheless, it took half a century until empirical verification took place (Bachmann 2022). In general, the idea that aggression and authority should be accepted as intrinsic parts of human nature has been a difficult strand in academic debates. It is to the credit of behavioral psychologists like Lorenz that the reason and the benefits of aggression did not vanish completely from scientific attention. In his studies of animal behavior, he considered aggression to be »an element within the concert of drives« (Lorenz 1963), meaning that it fulfills certain functions that a species cannot do without. However, the more social a species is, the more aggression is expressed through a complex set of signals. This allows conflicts to escalate in a controlled manner, and the individuals involved can surrender before suffering serious injury. In consequence, the knowledge about the opponents’ strengths leads to a rank order within a pack that additional y prevents perpetual conflicts. In later decades, evolutionary psychologists identified many similarities between animals and humans, and legitimately chal enged the view that humans could be outstandingly more rational and civilized than animals in social groups (Buss 2015). However, publications in this field have focused mainly on mating and only little on the importance of rank dynamics for human behavior in the context of work and leadership. In business practice, the attitudes were ambivalent. In 1960, McGregor defined Theory X and Theory Y (1960), suggesting that managers would just have to trust their employees to create a better and more productive work environment. In 1978, P. Vatter, A. Weisbeck: Empirical Verification of Dif erent Rank Dynamic Roles in Informal Hierarchies 219 Bergmann first wrote about New Work (2019). But for a long time, management thinking was stil dominated by a Tayloristic stance. Command-and-control was widely seen as the only way to coordinate large undertakings. It was in the nineties, when the New Economy and the IT skil shortage led to a revolution in project management and leadership (Duhigg 2016). Creativity, flexibility, and employer attractiveness gained higher importance. Liberal approaches, such as Design Thinking and agile project management, began to replace the overly excessive bureaucracy and rigidity of previous approaches, slowly shifting responsibility to the employee level. At the same time, formal hierarchy was being reduced, the informal hierarchy kicked in and became more important for successful project implementation (Diefenbach & Sillince 2011). This aspect has been widely neglected and might be one of the possible reasons why numerous companies have been struggling a lot with New Work formats (Busch & Link 2021). 2 Rank Dynamic Mechanisms and Strategies in Recent Works For a deeper understanding of the influence and the mechanisms of informal hierarchies, the characteristic behavior of individuals in gaining a best-possible rank must be known. Schindler’s model may be a clue but seems not yet specific enough. The applicability of Schindler’s role definitions, derived from psychiatric patients to business teams may suffer from certain weaknesses – especial y since the patients had no reason to compete for salary or promotion. It is therefore unclear whether the role definitions accurately reflect the roles that occur in the business context, or whether there are more roles that should be considered. In previous work, the author presented a model describing how ranks are negotiated within groups (Vatter & Kugler 2022). According to this, the root of rank dynamics is the competition for resources, like food, influence on group decisions, mating partners, and the psychological pleasantness of attention or admiration. As they are limited, group members compete for them against each other. Simultaneously, they must cooperate with each other to gain resources and ensure the continued existence of the group. This balancing between competition and cooperation is an ongoing decision each individual has to make. 220 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY An individualś decision is based on the factors (Fig. 1): 1. the individual’s psychological capabilities, like the ability to endure conflicts or act in a socially valued manner, 2. the individual’s values, expressed in the behavior regarded as acceptable or socially appropriate. 3. the individual’s psychological needs, like the need for admiration or being admitted to a group in a harmonious manner, and Depending on these factors, a group member forms a strategy on how to compete in the rank dynamic within a team. This strategy manifests itself in certain patterns of behavior that can be interpreted as signals about what kind of contributions others can expect. Individual Group Psychol. Psychol. Rank Capabilities Values Needs Acceptance Strategy Group Decision Denial Behavioral Signalized Request Patterns Contrib. / Claims Figure 1: Rank Dynamic Negotiation Model Source: According to Vatter & Kugler 2022 The group then decides on behalf of three questions if it trusts the signals given: 1. Is the individual able to make the promised contribution? 2. Is the individual willing to make the promised contribution? P. Vatter, A. Weisbeck: Empirical Verification of Dif erent Rank Dynamic Roles in Informal Hierarchies 221 3. Is the need for the promised contribution higher than the resources claimed (or are other team members providing this kind of contribution sufficiently)? The first two questions are highly based on uncertainty. Therefore, any signal associated with costs, such as the risk of punishment, is valued more than the plain verbal statement (McAndrew 2002, 2018). If all three questions are answered with yes, the individual is granted the desired rank in the informal hierarchy. If not, chal enged role owners receive the support of the group and the chal enger typical y must resign to a lower rank. Therefore, not all individuals thrive for the highest rank, as they may lack the needed abilities or rather avoid the associated costs and risks. Some individuals prefer a lower rank which stil provides enough resources to survive. On this behalf, Vatter & Kugler reviewed Schindler's role definitions and compared them to observations in business environment (2022). Hence, the roles have been revised as following: • The Alpha role was kept unchanged the leader of the group. Its contribution is to coordinate the activities of the group and al ocate resources. Moreover, it defines the rules within the group and takes uncomfortable but necessary decisions, e.g., the expulsion of a member from the group (see below). By assuming responsibility, it lowers the psychological stress of the other team members. • The Beta role was defined as Alpha’s assistant. 8 Therefore, it is assumed to have good social skills and the ability to be subordinate and loyal. In addition, it contributes by providing positive emotions and by carrying out Alpha’s orders. The contribution to the group is to provide order and maintain social cohesion. • A new Delta role forms the middle class of a group, preferring a secure and stable rank without prospect of a leadership position. 9 Its contribution lies in its labor, both manual and intellectual. It is characterized by a preference for structure and therefore advocates authority. 8 In contrast to Schindler's definition, Beta is not primarily providing workforce (see Delta). 9 Best comparable to Schindler’s Beta 222 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY • The Gamma role represents the opposition to Alpha and embodies the alternative perspective of various matters. 10 In its rebellious attitude it is convinced it was the better leader but lacks the followership of the other members. By providing an additional point of few, Gamma limits Alpha’s scope of action and prevents arbitrary decisions. Thus, Gammas most important contribution is to hold up a regulatory mechanism. • The Omega role is on the last rank of the informal hierarchy. The group does not recognize any useful contribution from its side. It is unable to defend against the group’s aggression and is therefore at risk of being excluded from the group. 3 Hypothesis and Objective of This Work Even though social behavior is complex and may vary significantly in different situations and over time, there are arguments that make it seem plausible that the number of roles is limited. This is, because a group’s social needs for psychological contributions may be manifold, but still finite. Additionally, roles should be easy to detect, so other group members can easily understand what kind of contribution an individual is promising. Moreover, pretending a certain behavior consistently is very difficult. So, coherent behavior makes the signaled proposal more reliable. For these reasons, roles should be Nash equilibria and corresponding strategies most successful if consistently adhered to (Holt & Roth 2004). If these assumptions are true, a specific set of roles should be detectable. Hence, the aim of this paper is to verify the existence of the mentioned roles in the informal hierarchy. 4 Data Collection and Analysis Based on the given role definitions, a self-assessment questionnaire has been designed. Each of the five scales was represented by six items (Alpha: A1 – A6, Beta: B1 – B6, …). To increase participant acceptance of the questionnaire and the 10 We agree with Bachmann that Schindler's Omega role should be divided into an »active Omega« and a »passive Omega«. The switch in names seems appropriate, as they are loacted on different rank levels and »Omega« should be used for the lowest rank in the group. P. Vatter, A. Weisbeck: Empirical Verification of Dif erent Rank Dynamic Roles in Informal Hierarchies 223 discrimination, a seven-point Likert scale was used for the response category, ranging from "totally agree" to "totally disagree". A seven-point Likert scale is easy to understand, reduces the "middle effect", and positively contributes to the discriminative power of the items. The items were constructed targeting an individual’s contribution to the group. This was considered meaningful as asking for the contribution is positively connotated and truthful answers can be expected. Additional y, the contributions seemed to be a good distinguishing feature of a role. When finding questions for contributions was found impossible, especial y for the Omega role, other significant characteristics of the role are determined. A pretest was carried out with five test persons which led to some adjustments in the wording of the final items (Tab. 1). The survey was implemented as online questionnaire and was distributed among students and on various social media platforms. The author's university is mainly aimed at working students. It can therefore be assumed that the participants have several years of professional experience. In total, 457 subjects, mainly German citizens, completed the test. The resulting datasets had been undergone several plausibility checks, such as the relative speed index RSI < 2. So, dubious entries were excluded beforehand the analysis and no further statistical outlier handling was pursued. This resulted in n = 421 usable datasets in total. 4.1 Descriptive Analysis and Assessment of Model Quality In the resulting data 53 % of the test persons were male and 47 % female. 54 % of al test persons were between 20 and 29 years old, 19 % between 30 and 39, 11 % between 40 and 49 and 14 % over 50 years old. 5 % had a secondary school diploma (ger.: Qualifizierter Hauptschulabschluss), 19 % medium maturity (ger.: Mittlere Reife), 28 % a high-school diploma (ger.: Abitur) and 48 % a university degree. PhD and no diploma both ranked below 1 %. 65 % of al participants reported to be working, while 34 % were studying. This may be inaccurate as most students at that university work regularly and study part-time. In the first step of the analysis, the model validity was determined. Cronbach’s Alpha showed values between 0.45 and 0.90, the discriminatory power was between 0.06 and 0.80. In particular, some items of the Delta and Gamma scales showed weak 224 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY values. To correct this, the analysis was continued with an exploratory factor analysis (EFA). 4.2 Exploratory Factor Analysis (EFA) The applicability of the EFA was tested using the Kaiser-Meyer-Olkin test, which showed values between 0.72 and 0.96 (Bühner 2021). Hence, the EFA could be applied. The EFA was performed in form of a paral el analysis with resampling and 2,000 repetitions, applying the Weighted Least Square method (WLS). The resulting eigenvalues showed a value within the confidence interval for a five-factor model (Fig. 2). So, the presence of five factors has been proven. Then, the EFA was proceeded to determine the factor loadings (Tab. 2). The Promax rotation and the WLS estimation have been used. The item complexity turned out to be 1.5. According to the results, three items (B5, D4 and D5) have been removed from the list of items. Two further items have been assigned to different scales as the cross-loadings were significantly higher than the loadings on the initially intended scale (B6 to Alpha, G1 to Omega). By this, the quality characteristics improved significantly. None of the discriminatory powers were below 0.20. By this, the Cronbach’s Alpha of the Delta scale increased to 0.51. P. Vatter, A. Weisbeck: Empirical Verification of Dif erent Rank Dynamic Roles in Informal Hierarchies 225 Table 1: Rank dynamic self-assessment questionnaire Items11 A1 In the team, I have the final say in decisions. A2 I am more responsible for the team's results than the other team members. A3 I set the goals in the team and take over the organization. A4 I take charge of the team in critical situations. A5 I act as a spokesperson between my team and external stakeholders. A6 I see my contribution to the team as taking on leadership responsibilities. B1 I mediate in interpersonal conflicts within the team. B2 I exert a positive influence on the team leadership. B3 I promote social cohesion in the team. B4 I maintain a good relationship with the team leader. B5 I contribute my professional expertise to the team. B6 I enforce the team's rules when someone steps out of line. D1 My actions in the team represent the team's labor power. D2 As a team member, I primarily increase the team's labor power. D3 I perform the intel ectual or manual work tasks in the team. D4 My contribution to the team is mainly the completion of work. D5 It’s ok if I do not have to participate in decisions and only do the tasks. D6 I prefer direct task accomplishment rather than discussion about it. G1 The team does not have clear goals and structures for me. G2 I criticize the coordinated approach in the team. G3 I exhibit unconventional working methods compared to colleagues. G4 In the team, I critical y question the current approach. G5 I constantly think about the improvement of the team. G6 I question the status quo of the team. O1 The team does not let me participate in decision-making processes. O2 My views are ignored by the team. O3 The team does not cooperate with me. O4 I am the scapegoat of the team. O5 I am not a ful team member. O6 I cannot speak freely and openly in the team. 11 For the items’ original German wording, see Weisbeck 2023 226 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 2: Factor loadings of rank dynamic self-assessment survey* Item Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 A1 0.80 A2 0.75 A3 0.81 A4 0.78 A5 0.59 A6 0.87 B1 0.40 0.56 B2 0.34 B3 0.80 B4 0.43 -0.36 B5 -0.31 B6 0.61 D1 0.42 D2 0.52 D3 0.43 D4 -0.32 0.36 D5 -0.36 0.31 D6 0.43 G1 0.43 G2 0.49 G3 0.46 G4 0.87 G5 0.42 G6 0.64 O1 0.40 O2 0.71 O3 0.67 O4 0.40 0.60 O5 0.51 O6 0.65 Alpha Beta Delta Gamma Omega * Factor loadings < 0.3 have been removed. P. Vatter, A. Weisbeck: Empirical Verification of Dif erent Rank Dynamic Roles in Informal Hierarchies 227 Parallel Analysis Scree Plots 8  FA Actual Data --- FA Resampled Data actors 6 al Fciprin 4 f Pe o Eigen values of sample data within alu 95 %-quantile for up to five factors en V 2 igE 0 0 5 10 15 20 25 30 Number of Factors Figure 2: Paral el Analysis Scree Plots 4.3 Calculation of Higher Order EFA After this, the correlations between the five factors were calculated and showed significant values between -0.43 and 0.62. Therefore, a general factor g was considered and a higher order EFA was calculated. The results showed the fol owing loadings onto the general factor: Alpha: 0.7, Beta: 0.7, Delta: 0.4, Gamma: 0.8, Omega: -0.3. These correlations can be interpreted as similarities in the characteristics of the five roles. The loadings can—with all due caution—be interpreted as amount of contribution a role makes to the group. This seems plausible as al roles do contribute to the group except the Omega role. Thus, it is prone to being excluded from the group, as the theory suggests. 228 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 4.4 Confirmatory Factor Analysis (CFA) Lastly, a confirmatory factor analysis (CFA) has been performed to confirm the validity of the model. There, the higher order model (with general factor g) and the regular model (without general factor g) were compared. Their quality characteristics were calculated as shown in table 3. Table 3: Quality characteristics of CFA CFI TLI RMSEA SRMR 𝜒𝜒2-Test Good Fit ≥ 0.97 ≥ 0.97 ≤ 0.05 ≤ 0.05 ≤ 2 Acceptable Fit ≥ 0.95 ≥ 0.95 ≤ 0.08 ≤ 0.10 ≤ 3 Regular Model 0.973 0.970 0.042 0.065 1,75 (without general factor) Higher Order Model 0.961 0.957 0.050 0.072 2.06 (with general factor) The regular model performed well on behalf of almost all quality criteria (Gäde 2020, p. 649). Only the SRMR value was 0.065 and, thus, can be considered as acceptable only. The higher order model performed acceptable on almost all quality criteria. Only the RMSEA value was 0.050 and, thus, is regarded as good fit. This means the higher order model did slightly worse than the regular model without the general factor g. This is traced back on the fact that the higher order equation must fulfil more conditions as the regular model. 5 Relevance for business teams To establish a good working culture, it first must be clarified what this really means. It is a common misunderstanding that the best working culture would be as cooperative as possible and total y free of conflicts. On closer inspection, this turns out to be untrue. Conflicts, in general, occur for the purpose to find the right decision about group related issues. For this can be done efficiently, an intact informal hierarchy is needed. Rank dynamic conflicts are necessary to constantly adjust the informal hierarchy and create a rank dynamic structure every member of group can live with. Avoiding these conflicts means stal ing the process and preventing the group from reaching a productive state. This means, conflicts must P. Vatter, A. Weisbeck: Empirical Verification of Dif erent Rank Dynamic Roles in Informal Hierarchies 229 be managed, so they can take an appropriate share in a balanced mix of a team’s social interactions. For a long-term success, leaders need to ask: 1. Is the informal hierarchy of the team balanced? Is there an adequate mixture of informal roles? 2. Is the relation of competition and cooperation balanced? Can group members have the conflicts they need to have and are those resolved quickly and do not consume unnecessary energy? 3. Am I accepted as the instance that sets the norms according to which conflicts are fought out? If the last point is given, the team leader steer the team by cultural norms. This means, that there is a common understanding, which kind of behavior is accepted and what wil not be tolerated. This has many advantageous over the directives as they are commonly applied. Those are circumvented by competitive or unsatisfied employees where possible. In the first case, cooperative employees are protected by cultural norms and can freely live out their creativity and compete in terms of the individual strengths. 6 Summary and Outlook The success of project teams, particularly when new liberal leadership approaches are applied, highly depend on a functioning group dynamic. Team members make specific contributions to the team, depending on their psychological capabilities, needs and values, competing for a specific rank in the group. In the present paper, Schindler’s set of roles has been slightly adjusted, so the match better to the circumstances of work environments. A self-assessment questionnaire has been designed and a survey was carried out. A factor analysis showed that the five factors could be determined which is a strong indication for the existence of the roles. Though this may be considered as a major step ahead, there are three main limitations to this work. First, the results are solely based on self-assessment and do not proof that other team members share the appraisal of the test subjects. Therefore, an alignment with a third person assessment wil be subject of subsequent work. Second, it is not clear if the assumed rank order can really withstand in real conflict situations and group resources are assigned accordingly. There might be 230 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY more influence factors which determine, who wins a confrontation in a real-world environment. This intensifies as a large proportion of test-persons were students with an unknown amount of work experience. Thus, the model needs more validation on behalf of the outcome of competitive experiments or field observations. The third limitation refers to the definition of the roles. Even though their existence seems without doubt there is a lack of detailed description and a full understanding of their motives, behavioral patterns, and interaction. Thus, a deeper qualitative investigation and further adjustment seem to be justified. Moreover, it is not excluded that there are further important roles which occur less frequently, and so they could not be detected with the given item set. References Bachmann, T. (2022). Functional Group Positions and Contact Behavior in Problem-Solving Groups. Gr Interakt Org, 53(1), 131–144. https://doi.org/10.1007/s11612-021-00613-6 Bergmann, F. (2019). New Work New Culture. Zero Books, Winchester, UK. Bühner, M. (2021). Einführung in die Test- und Fragebogenkonstruktion. 4th ed., Pearson Education, Halbergmoos, Germany. Busch, M. W. & Link, K. (2021). Was macht Agilität mit Macht? Eine Analyse des Machtphänomens und agilen Organisationen. Journal für Psychologie 29 (1): 9–38. https:// doi.org/10.30820/0942-2285-2021-1-9. Buss, D. (2015). Evolutionary psychology: The new science of the mind. 5. ed., Psychology Press, New York, USA. Diefenbach, T. & Sillince, J. A. A. (2011). Formal and Informal Hierarchy in Different Types of Organization. Organization Studies, 32(11), 1515–1537. https://doi.org/10.1177/017084 0611421254 Duhigg, C. (2016). What Google Learned From Its Quest to Build the Perfect Team. In: The New York Times Magazine, 25.02.2016 Gäde, J. C., Schermel eh-Engel, K. & Brandt, H. (2020). Konfirmatorische Faktorenanalyse (CFA). In: Moosbrugger, H. & Kelava A. (Edts.): Testtheorie und Fragebogenkonstruktion, 3. ed., Springer, Berlin, Heidelberg, Germany, 615–660. Holt, C. A., & Roth, A. E. (2004). The Nash equilibrium: A perspective. Proc. National Academy of Sciences, 101(12), 3999-4002. https://doi.org/10.1073/pnas.0308738101 Lorenz, K. (1980). Das sogenannte Böse: Zur Naturgeschichte der Aggression, 7. ed., dtv, München, Germany. McAndrew, F. T. (2002). New Evolutionary Perspectives on Altruism: Multilevel-Selection and Costly-Signaling Theories. Current Directions in Psychological Science, 11(2), 79–82. https://doi.org/10.1111/1467-8721.00173 McAndrew, F. T. (2018). Costly Signaling Theory. In T. Shackelford (Ed.), Encyclopedia of Evolutionary Psychological Science, 1–8. Springer, Cham, Germany. https://doi.org/ 10.1007/978-3-319-16999-6_3483-1 McGregor, D. (1960). Theory X and theory Y. Organization Theory, 358(374), 5. P. Vatter, A. Weisbeck: Empirical Verification of Dif erent Rank Dynamic Roles in Informal Hierarchies 231 Milgram, S. (1963). Behavioral Study of Obedience. Journal of Abnormal and Social Psychology. 67(4), 371-378 Ouchi, W. G. & Price, R. L. (1978). Hierarchies, clans, and theory Z: A new perspective on organization development. Organizational Dynamics, 7(2), 25-44. Schindler, R. (2016). Das lebendige Gefüge der Gruppe: Ausgewählte Schriften. Psychosozial-Verlag, Gießen, Germany, First publication: 1957 Schwaber, K. & Sutherland, J. (2020). The Scrum Guide. Scrum Alliance. Available online: https:// scrumguides.org/docs/scrumguide/v2020/2020-Scrum-Guide-US.pdf. Last Access: 02/03/2023 Spal er, C. (2018). Das lebendige Gefüge der Gruppe: Plädoyer für eine kritische Relektüre der Rangdynamik nach Raoul Schindler. Gr Interakt Org, 49(4), 405–412. https://doi.org/ 10.1007/s11612-018-0436-4 Vatter, P. & Kugler, M. (2022). Rangdynamische Konflikte – Dysfunktionales Verhalten in infor-mellen Hierarchien. zfo – Zeitschrift Führung + Organisation, 6, 359-363. Weisbeck, A. (2023). Informel e Hierarchie - Untersuchung des Zusammenhangs zwischen Persönlich-keitsstruktur und rangdynamischer Position, Master Thesis, FOM University of Applied Sciences, Nuremberg, Germany. 232 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY ENHANCING THE UNDERSTANDING OF E- COMMERCE REVIEWS THROUGH ASPECT EXTRACTION TECHNIQUES: A BERT-BASED APPROACH LALEH DAVOODI Åbo Akademi University, Faculty of Social Sciences, Business and Economics, and Law, Turku, Finland laleh.davoodi@abo.fi The growth of online customer reviews on e-commerce platforms has led to an overwhelming volume and variety of data, making manual analysis impractical for both consumers and managers. Consequently, machine learning techniques, such as Aspect-Based Sentiment Analysis (ABSA), have gained prominence for their ability to determine sentiment at the aspect level. This study aims to fine-tune natural language processing models for aspect extraction in e-commerce customer reviews. We manually annotated 2781 online user review sentences in English and employed different extensions of the BERT model to identify implicit and explicit aspects. This approach diverges from prior studies, as our dataset comprises real user reviews from five prominent e-commerce platforms. The findings demonstrate the models’ effectiveness in extracting aspects from diverse e-commerce user reviews, yielding a deeper K eywords: understanding of user-generated content and customer customer reviews, satisfaction trends, and providing valuable insights for e-commerce, managerial decision-making. This study contributes to the ABSA aspect literature and offers practical implications for e-commerce extraction, BERT, platforms aiming to improve their products and services based machine on customer feedback. learning DOI https://doi.org/10.18690/um.fov.4.2023.15 ISBN 978-961-286-751-5 234 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Currently, customers heavily rely on review content posted by other users on various e-commerce sites prior to making a purchase decision. However, with the ever-increasing volume and variety of online content, it has become chal enging for customers to manual y sift through the vast amounts of information available (Ansari et al., 2020). Moreover, managerial decision-making requires a constant flow of up-to-date information to provide insight into the trends and dynamics of customer satisfaction and manual classification of reviews has become increasingly difficult. Therefore, the popularity of machine learning techniques, such as opinion mining and aspect term detection, has increased in recent years due to the rise in online data volume and diversity. Aspect-Based Sentiment Analysis (ABSA) is a subfield of Sentiment Analysis (SA) that primarily focuses on determining the sentiment of products or services at the aspect level. To gain a deeper understanding of user-generated content, an essential task in ABSA is to extract both implicit and explicit aspects from various online reviews. This paper aims to fine-tune the Bidirectional Encoder Representations from Transformers (BERT) and two of its extensions for aspect extraction in user-generated content in the form of customer reviews of e-commerce platforms. The research question we set forth to answer is the following: How can aspect extraction techniques from manual y annotated customer reviews be used to enhance the understanding of customer opinions in e-commerce? This was accomplished by manually annotating 2781 online user review sentences in English. Unlike the majority of literature that assesses aspect extraction tasks using only the Laptops and Restaurants datasets from the SemEval 2014, 2015, and 2016 ABSA aspect extraction context (Xu et al., 2019, Pereget al., 2019; Chauhan et al., 2022; Venugopalan et al, 2022), this study employed BERT to categorize implicit and explicit aspects of our manual y annotated dataset into 14 distinct groups. The data used for this analysis was obtained from real user reviews of five prominent ecommerce platforms that are available online. The present study demonstrates the effectiveness of using BERT for aspect extraction from a diverse range of user reviews in the e-commerce domain. By utilizing the manually annotated dataset and categorizing implicit and explicit aspects into 14 distinct groups, this study has L. Davoodi: Enhancing the Understanding of E-commerce Reviews through Aspect Extraction Techniques: A BERT-Based Approach 235 provided a deeper understanding of user-generated content and customer satisfaction trends offering novel academic insights. Furthermore, the findings of this study can inform managerial decision-making and help e-commerce platforms to improve their products and services to meet the needs and expectations of their customers. The rest of the paper is structured as follows. Section 2 presents a literature review on academic contributions focusing on aspect detection in customer reviews. The methodology and data preparation are discussed in Section 3, with the main results presented and discussed in Section 4. Finally, some conclusions, limitations and future research directions are provided in Section 5. 2 Aspect detection and extraction Sentiment analysis (SA), also referred to as opinion mining, is the process that involves identifying, recognizing, and categorizing users' emotions or opinions on various services such as movies, products, events, or any attributes as positive, negative, or neutral. The data used in the analysis can be gathered from diverse sources such as review websites, forum discussions, blogs, Twitter, etc. SA is a powerful tool as it provides valuable information about people's preferences and can help companies gain a clear perspective regarding their product or service features (Mehta et al., 2020). By analyzing customer sentiments, companies can identify areas of improvement, determine customer satisfaction levels, and gauge market trends. SA has been widely applied in various industries, including hospitality, healthcare, and e-commerce, among others. Previous studies have general y categorized SA into document-level, sentence-level, and aspect-level SA. These levels aim to classify whether a whole document, a sentence (subjective or objective), and an aspect express a sentiment (Nazir et al., 2020). Target-based sentiment analysis (TBSA) or Aspect-based Sentiment Analysis (ABSA) is a sub-task of SA that provides a better understanding of the problem of SA at a fine-grained level (Liu 2012; Pontiki 2014) because it focuses directly on sentiments rather than on language structure. ABSA analyzes specific aspects or entities of a product, service, or topic and determines the sentiment associated with them. This technique provides a more in-depth and nuanced understanding of public opinion, making it useful for companies to develop marketing strategies, improve 236 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY product features, and enhance customer satisfaction. ABSA includes subtasks such as aspect/target term extraction (ATE), opinion term extraction (OTE), aspect/target term sentiment classification (ATC), and others (Peng et al, 2020). The objective of ATE is to identify and extract terms that represent aspects of a given sentence. For instance, in the sentence "Excel ent customer service and delivery", the aspect terms are “customer service” and “delivery”. ATE typical y involves two sub-tasks: (1) extracting al aspect terms (such as "delivery") from the text, and (2) grouping aspect terms with similar meanings into categories, where each category represents a single aspect (e.g., “delivery”, “shipping”, and “track number” wil be clustered into the shipping aspect). It should be noted that ABSA categorizes aspects into two types: explicit and implicit aspects. Explicit aspects are directly mentioned in a text. In contrast, implicit aspects are not explicitly indicated by any specific word or term (Alqaryouti et al., 2020). For instance, in the sentence “If you need something tomorrow go somewhere else”, “shipping” is an implicit aspect. Prior research on aspect extraction can be classified into four approaches: rule-based (Poria et al., 2014; Liu et al., 2015), supervised (Maitama et al., 2021; Poria et al., 2016), unsupervised (Chauhan et al., 2020; Luo et al., 2019), and semi-supervised (Ansari et al., 2020; Anand et al., 2016). According to He et al., (2017), rule-based methods typical y do not group extracted aspect terms into categories, while supervised learning requires annotated data and may face domain adaptation issues. Unsupervised methods are utilized to avoid reliance on labeled data. In addition to the methods such as statistical analysis, topic modeling, and dependency parsing used in prior unsupervised aspect extraction studies, supervised aspect extraction techniques such as Conditional Random Field (CRF) and long short-term memory have also been employed (Maitama et al, 2021). In recent years, deep learning has become one of the most effective approaches for natural language processing tasks, due to its supervised training process with large amounts of training data. However, acquiring a large amount of supervised data can be a difficult and time-consuming process, particularly for NLP tasks in low-resource languages. In these cases, transfer learning can offer a solution by al owing a model to be pre-trained on a large amount of unsupervised data before being fine-tuned for a specific task under supervised conditions. BERT is one of the latest popular algorithms that employ this transfer learning approach (Yanuar et al., 2021). L. Davoodi: Enhancing the Understanding of E-commerce Reviews through Aspect Extraction Techniques: A BERT-Based Approach 237 3 Methodology This section wil present the NLP-based aspect extraction and classification methodology. Our main goal is to construct a model that is able to detect the aspects of customer feedback automatically. As presented in Figure 1, we have applied five steps to achieve our research objective. In this section, we discuss data col ection, annotation, and preprocessing, and present the machine-learning models used for aspect extraction and classification. Figure 1: The stages of the research process 3.1 Data col ection and processing In order to study the user-generated content in e-commerce and evaluate their potential insights with ABSA, we collected 12,000 English reviews from Trustpilot, one of the biggest platforms in hosting online reviews, between 2013 and 2021. Next, we randomly selected 3,500 messages for aspect extraction and manual data annotation, excluding messages that did not meet certain criteria, such as not being written in English or not discussing any of the considered aspects. The final dataset consists of 2,782 data points consisting of the reviews from five online stores, namely Zalando, Wish, Sheinside, Boozt, and Nelly. Reviewers from 73 different 238 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY countries contributed to the reviews, with the UK (1166 reviews) and the US (778 reviews) having the highest number of reviews among all 73 countries. Prior to extracting our probable candidate aspects, we executed a series of seven text-cleaning steps: converting the text to lowercase, eliminating any non-English text, discarding all non-alphabetic characters, stripping out HTML tags and URLs, expanding contractions in the reviews, eliminating stop words, and performing lemmatization. Following text processing, the total count of distinct words was determined to be 4580. In preparation for manual data annotation, it was necessary to specify the aspects relevant to the domain. The final set of aspects, as presented in Table 1, is extracted from the literature and analysis of the reviews. Some further details can be found in Davoodi et al. (2022). L. Davoodi: Enhancing the Understanding of E-commerce Reviews through Aspect Extraction Techniques: A BERT-Based Approach 239 Table 1: Final aspects 1 Shipping The quality of the delivery can refer to various factors, such as the cost and timeline. 2 Trust The customers' overal sentiment towards the store. 3 Item quality The level of quality of the products. 4 Customer The quality of customers' direct interactions with the store's service representatives, such as their helpfulness, professionalism, and effectiveness 5 Pricing The prices offered by the store, as well as the availability of discounts and promotional campaigns. 6 Product The quality of the product images and size guides provided on features the website and implicit aspects such as described or expected 7 Refund The speed and quality of the refund process, as wel as the process effectiveness and efficiency of the store's handling of refund- related issues. 8 Return The speed, convenience, and cost associated with the return process process for customers. 9 App The experience that a user has while interacting with the experience website of the store. 10 Delivered The state of the products that have been delivered, such as if product status they are broken or have an odor 11 Information The presence and accuracy of the information, including the quality of advertisements and whether they may be deceptive. 12 Packaging The visual appeal and quality of the product’s packaging. 13 Payment The level of quality in the financial transaction process, including accuracy, security, efficiency, and different payment options. 14 Product The range of products or brands that are being offered. availability 240 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3.2 Data annotation Prior to annotating the data, we utilize Python libraries such as contextualSpellCheck and Caribe to conduct spel and grammar checks on the reviews, thereby converting them into standard English text. Subsequently, we established a set of rules to convert implicit aspects into explicit ones while endeavoring to maintain the original text as closely as possible. Following the text-cleaning process, two annotators manual y annotated the data separately. The sentences were divided into individual words, and labels were assigned to the nouns or noun phrases only if they conveyed a sentiment related to at least one of the aspects, subsequently disregarding aspects with a neutral sentiment. Previous studies have revealed that aspects are often represented by nouns or noun phrases, while opinions are typical y conveyed through adjectives or adjective phrases (Samah et al., 2014). Research on sentiment analysis has also demonstrated that certain tag combinations are instrumental in identifying aspects and opinions. In contrast to these prior investigations, the present study relied heavily on sentence parsing by considering a broader range of sentence components as potential aspects and/or opinions (Hu et al., 2004; Hu et al., 2004; Subrahmanian et al., 2008; Samha et al., 2014). It is important to note that the aspect terms were extracted solely based on their positive or negative sentiment in the given sentence without considering the preceding or subsequent sentence, i.e., the review as a whole. The ATE task is typical y approached as a sequence labeling problem, where each input word is assigned one of three labels: "B" for the beginning of an aspect term, "I" for the continuation of an aspect term, or "O" for non-aspect terms. The "I" label is necessary because some aspect terms can consist of multiple words, and the system must identify and extract all of them as aspects (Karimi et al., 2020). However, in this study, we provided 29 labels as follows: O, App_experience_B, App_experience_I, Trust_B, Trust_I, Customer_service_B, Customer_service_I, Delivered_product_status_B, Delivered_product_status_I, Information_B, Information_I, Item_quality_B, Item_quality_I, Packaging_B, Packaging_I, Payment_B, Payment_I, Pricing_B, Pricing_I, Product_availability_B, 'Product_availability_I, Product_features_B, Product_features_I, Refund_process_B, Refund_process_I, Return_process_B, Return_process_I, Shipping_B, and Shipping_I. L. Davoodi: Enhancing the Understanding of E-commerce Reviews through Aspect Extraction Techniques: A BERT-Based Approach 241 In the given example, " Love this site fantastic saving for quality stuf. ", the text is first transformed into a set of words, and then the annotators assign corresponding labels to each word. The label sequence for this example is as fol ows: ['O', 'O', 'Trust_B', 'O', 'Pricing_B', 'O', 'Item_quality_B', 'Item_quality_I', 'O']. Once the individual annotations were completed, disagreements were identified and discussed to arrive at a final agreement between the annotators. 3.3 Machine learning models for aspect extraction To evaluate our annotated model, we selected three recent transformer-based machine-learning models: BERT, RoBERTa, and BERT_Review. BERT (Devlin et al., 2018) is a language model that can evaluate the context of a word from both the left and right sides simultaneously, unlike traditional language models that process sentences from one direction. This is achieved using the masked language modeling (MLM) technique, which randomly masks a word in a sentence and replaces it with a [MASK] token. The model then attempts to predict the masked word based on the context from both sides of the masked word. In addition to MLM, BERT also includes a next-sentence prediction (NSP) task, which involves predicting whether a given sentence follows another sentence or not. This helps BERT to capture the relationship between sentences and improve its ability to perform tasks such as question answering and text classification. Overall, BERT's bidirectional feature and MLM technique make it a powerful language model that can provide more contextual features from a sentence compared to other models such as ELMO. RoBERTa (Liu et al., 2019) and BERT_Review (Xu et al., 2019) are variants of the BERT model, with the former showing promising results in various natural language processing (NLP) tasks. These variants are designed to enhance BERT's performance by incorporating different modifications. RoBERTa, for example, utilizes a transformer architecture like BERT but is trained differently and for a longer duration. Additionally, RoBERTa employs the entire sentence as input and eliminates the fol owing sentence prediction objective. 242 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY BERT_Review is designed to address Review Reading Comprehension (RRC). For this purpose, Xu et al. (2019) created a dataset, which is based on a wel -known benchmark used for aspect-based sentiment analysis. However, due to the limited training examples available for RRC and aspect-based sentiment analysis in their data, they subsequently explored a post-training method using BERT to improve the fine-tuning process for RRC. Additional y, they applied the proposed post-training approach to other review-based tasks, such as aspect extraction and aspect sentiment classification in aspect-based sentiment analysis, to demonstrate its universality. The results indicated that the proposed post-training approach is highly efficient (Xu et al., 2019). 3.4 Model building To evaluate our annotated model, we selected three recent transformer-based machine-learning models: BERT, RoBERTa, and BERT_Review. Given the limited size of our dataset, we opted to employ cross-validation as a means of achieving a more reliable estimation of our models' performance. Specifical y, we partitioned our manual y annotated dataset into distinct training and validation sets by means of a five-fold cross-validation methodology, involving ten steps of iteration per fold. Subsequently, to obtain an aggregate summary of model performance, we calculated the mean value across al folds for each performance metric. To transform the input text to numeric features, we utilized a transformer tokenizer for al models. We utilized loss and accuracy as the performance metric for al models, with the main model serving as the baseline. For training BERT, RoBERTa, and BERT_Review our network model includes the main model with 12 layers and 768 hidden dimensions as well as the initial learning rate of 1e-4. Moreover, the training batch size was defined as 8. The programming language used for model building and data analysis is Python 3.10.11. We evaluated the models' performance using two commonly used measures: accuracy and F1 score. 4 Results In this section, we wil present the outcomes of our experiments and demonstrate the efficacy of different machine-learning models in the domain of aspect extraction and classification. We will also compare our findings to prior academic research in L. Davoodi: Enhancing the Understanding of E-commerce Reviews through Aspect Extraction Techniques: A BERT-Based Approach 243 this area. Additionally, we will discuss several noteworthy observations that are relevant to managing expectations regarding performance. 4.1 Performance analysis As detailed in the preceding section, a total of three distinct models were built and trained. The outcomes of these models are presented in Table 2. The table displays the mean loss, accuracy, and F1 value of the validation set after performing 5-fold cross-validation. Additionally, the execution time required to build the models is reported. Table 2: Aspect extraction performance Method Loss. Accuracy- F1- Execution time (in validation validation validation seconds) set set set BERT_Review 0.104 0.972 0.841 3,077 BERT 0.113 0.969 0.829 3,075 RoBERTa 0.109 0.969 0.828 2,905 After conducting 5-fold cross-validation for three transformer-based models, we calculated the mean validation accuracy for each model. To test for statistical significance between the models, we performed a one-way ANOVA with post-hoc Tukey HSD tests. The results showed that BERT_Review had a significantly higher mean accuracy compared to the other two models. However, there was no significant difference in mean accuracy between BERT and RoBERTa. As demonstrated in Table 2, the BERT_Review model displayed marginal y superior performance when compared to two other models, achieving an F1 score of 84%. Given the fact that the study involved 29 distinct categories, this level of performance can be considered promising. After running the cross-validation, to analyze the performance of the best model we partitioned the dataset into the fixed train (70%), validation (15%), and test (15%) sets. We used the test set for the error analysis. The analysis results are presented in Table 3. 244 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 3: Aspect extraction performance of the best model Method Accuracy- F1- Accuracy- F1-test Number of validation validation test misclassifications (test) BERT_Review 0.975 0.86 0.972 0.85 169 In regard to the specific misclassifications generated by the BERT_Review model, a portion of the errors can be attributed to the identification of neutral aspects in the messages. For instance, the fol owing review was misclassified by the model as a return process. However, there should not be any aspect marked in this message, as the sentiment expressed towards the return process in this message is neutral: "I ordered around 300 dol ars worth of clothes for my kids and I need to return 100 dol ars worth of clothes." Secondly, there are certain aspects in the dataset that occur less frequently, such as payment. Consequently, the model did not have sufficient samples to accurately identify these aspects. Thirdly, in the majority of the misclassification cases, the model was able to extract the correct aspects. However, there were slight differences in the position of the true and predicted aspects, which we believe can be easily addressed by increasing the training sample. For instance, in the following example: "No more purchase from Wish, no support after delivery of a product." The true aspects are Company, Customer Service, and Shipping, and the model was able to correctly identify them. However, for the Shipping aspect, the model should have identified "delivery of a product" as the shipping aspect but instead marked each " delivery" and " product" word separately as shipping. Finally, consider the following example: "Order arrived realy fast and was wel wrapped." In this case, the aspects being referred to are shipping and packaging. However, the model used in the study only predicted the aspect of shipping, whereas the true label encompasses both shipping and packaging. As a result, both predicted aspects are technical y true, but the issue of accurately identifying multiple aspects needs to be addressed further in future research. Moreover, it is important to acknowledge that human annotation errors may occur during the manual annotation process, and such errors have the potential to negatively impact the performance of the machine learning models trained in this research. L. Davoodi: Enhancing the Understanding of E-commerce Reviews through Aspect Extraction Techniques: A BERT-Based Approach 245 4.2 Discussion The input for classification models and data annotation, which involves extracting relevant aspect terms from reviews, is crucial in natural language processing. Previous research in this area has predominantly utilized datasets from SemEval 2014, SemEval 2015, and SemEval 2016 (Wang et al, 2021; Dai et al., 2019) for their experimental studies. However, our study focused on e-commerce businesses that do not have physical shops when collecting data, aiming to identify and understand the most prevalent sources of satisfaction and dissatisfaction in this context. For this purpose, data were col ected from five distinguished and widely used e-commerce platforms that cater to a diverse range of customers. The comprehensiveness and broadness of this data make it applicable to research endeavors aimed at analyzing smal er online retailers and comparable enterprises. While automated annotations are less time-consuming and costly than manual annotations, they are general y less accurate. In this study, we opted for manual annotation to gain a comprehensive understanding of the review content and its relevance to the target companies. By presenting a manually annotated dataset that includes extracted aspect terms with negative and positive sentiments, companies can ensure that they concentrate on the correct elements by obtaining precise aspect detection results. Additional y, the combination of sentiment classification of the extracted aspects can enable automated and accurate identification of the sources of dissatisfaction. From our review of the literature, we have found that BERT and RoBERTa are among the most frequently employed models for aspect extraction in user reviews (Chauhan et al., 2020; Tian et al., 2020; Yanuar et al., 2020; Lopes et al., 2021) and the F1 score is achievable in the range of 0.738-0.85 depending on the domain and the language of the reviews. Our experimental results revealed that the BERT_Review model outperformed the other two models under consideration. Among all categories, the Trust_B aspect attained the highest F1 score of 0.92, while Payment_I had the lowest score of 0.4. This finding indicates that the inclusion of infrequent aspects may negatively impact the model's performance. One potential solution to address this issue is to automatically generate labeled data for infrequent aspects. 246 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 5 Conclusions This study involved the development of a novel manually annotated dataset for the aspect extraction task, which was evaluated by using three state-of-the-art transformer-based models. The evaluation results demonstrate that the models performed wel , but further improvements could be achieved by adding more samples to the dataset, particularly for infrequent aspects such as payment and packaging. A possible direction for future research could involve the use of this manual y annotated dataset to generate automated labeled data, thereby enhancing the performance of the models. For example, this could entail fine-tuning Large Language Models like GPT-2 (Radford et al., 2019) and GPT-3 (Brown et al., 2022), as explored by Veyseh et al. (2022) and Ding et al. (2022). Moreover, this approach could help to address the key chal enges associated with aspect extraction tasks, which is the scarcity of training data that limits the training of large neural networks. Large Language Models have proven remarkable few-shot capabilities in various NLP tasks (Min et al., 2021). Given this, a logical use case would be utilizing it for data annotation purposes. References Alqaryouti, O., Siyam, N., Abdel Monem, A., & Shaalan, K. (2020). Aspect-based sentiment analysis using smart government review data. Applied Computing and Informatics. Vol. ahead-of-print, 20 pages. Anand, D., & Naorem, D. (2016). Semi-supervised aspect-based sentiment analysis for movies using review filtering. Procedia Computer Science, 84, 86-93. Ansari, G., Saxena, C., Ahmad, T., & Doja, M. N. (2020). Aspect term extraction using graph-based semi-supervised learning. Procedia Computer Science, 167, 2080-2090. Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J. D., Dhariwal, P., . . & Amodei, D. (2020). Language models are few-shot learners. Advances in neural information processing systems, 33, 1877-1901. Chauhan, G. S., Meena, Y. K., Gopalani, D., & Nahta, R. (2020). A two-step hybrid unsupervised model with attention mechanism for aspect extraction. Expert systems with applications, 161, 113673. Chauhan, G. S., Meena, Y. K., Gopalani, D., & Nahta, R. (2022). A mixed unsupervised method for aspect extraction using BERT. Multimedia Tools and Applications, 81, 31881-31906. Dai, H., & Song, Y. (2019). Neural aspect and opinion term extraction with mined rules as weak supervision. arXiv preprint arXiv:1907.03750. Davoodi, L., & Mezei, J. (2022). A comparative study of machine learning models for sentiment analysis: Customer reviews of e-commerce platforms. In Proceedings of the 35th Bled eConference Digital Restructuring and Human {Re} (pp. 217-230). Devlin, J., Chang, M. W., Lee, K., & Toutanova, K. (2018). Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805. Ding, B., Qin, C., Liu, L., Bing, L., Joty, S., & Li, B. (2022). Is GPT-3 a Good Data Annotator? arXiv preprint arXiv:2212.10450. L. Davoodi: Enhancing the Understanding of E-commerce Reviews through Aspect Extraction Techniques: A BERT-Based Approach 247 He, R., Lee, W. S., Ng, H. T., & Dahlmeier, D. (2017, July). An unsupervised neural attention model for aspect extraction. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (pp. 388-397). Hu, M., & Liu, B. (2004, August). Mining and summarizing customer reviews. In Proceedings of the tenth ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 168-177). Hu, M., & Liu, B. (2004, July). Mining opinion features in customer reviews. In Proceedings of the 19th national conference on American Association for Artificial Intel igence (pp. 755-760). Im, J., Song, T., Lee, Y., & Kim, J. (2019). Confirmatory aspect-based opinion mining processes. arXiv preprint arXiv:1907.12850. Karimi, A., Rossi, L., & Prati, A. (2021, January). Adversarial training for aspect-based sentiment analysis with bert. In Proceedings of 2020 25th International Conference on Pattern Recognition (ICPR) (pp. 8797-8803). Liu, B. (2012). Sentiment Analysis and Opinion Mining. Synthesis Lectures on Human Language Technologies, 5, 1-167. Liu, Q., Gao, Z., Liu, B., & Zhang, Y. (2015, June). Automated rule selection for aspect extraction in opinion mining. In the Proceedings of the Twenty-Fourth international joint conference on artificial intelligence (pp. 1291-1297). Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., . . & Stoyanov, V. (2019). Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692. Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., . . & Stoyanov, V. (2019). Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692. Luo, L., Ao, X., Song, Y., Li, J., Yang, X., He, Q., & Yu, D. (2019, August). Unsupervised Neural Aspect Extraction with Sememes. In Proceedings of The Twenty-Eighth International Joint Conference on Artificial Intel igence {IJCAI-19} (pp. 5123-5129). Lopes, É., Correa, U., & Freitas, L. (2021, April). Exploring Bert for aspect extraction in Portuguese language. In Proceedings of The International FLAIRS Conference (pp. 1-4). Maitama, J. Z., Idris, N., Abdi, A., & Bimba, A. T. (2021, May). Aspect extraction in sentiment analysis based on emotional affect using supervised approach. In Proceedings of the 4th International Conference on Artificial Intel igence and Big Data (ICAIBD) (pp. 372-376). Mehta, P., & Pandya, S. (2020). A review on sentiment analysis methodologies, practices, and applications. International Journal of Scientific and Technology Research, 9, 601-609. Min, B., Ross, H., Sulem, E., Veyseh, A. P. B., Nguyen, T. H., Sainz, O., . . & Roth, D. (2021). Recent advances in natural language processing via large pre-trained language models: A survey. arXiv preprint arXiv:2111.01243. Nazir, A., Rao, Y., Wu, L., & Sun, L. (2020). Issues and chal enges of aspect-based sentiment analysis: A comprehensive survey. IEEE Transactions on Affective Computing, 13, 845-863. Peng, H., Xu, L., Bing, L., Huang, F., Lu, W., & Si, L. (2020, April). Knowing what, how and why: A near complete solution for aspect-based sentiment analysis. In Proceedings of the AAAI Conference on Artificial Intel igence (pp. 8600-8607). Pereg, O., Korat, D., Wasserblat, M., Mamou, J., & Dagan, I. (2019). ABSApp: a portable weakly-supervised aspect-based sentiment extraction system. arXiv preprint arXiv:1909.05608. Pontiki, M., Galanis, D., Papageorgiou, H., Androutsopoulos, I., Manandhar, S., AL-Smadi, M., . . & Eryiğit, G. (2016). Semeval-2016 task 5: Aspect-based sentiment analysis. In Proceedings of the ProWorkshop on Semantic Evaluation (SemEval-2016) (pp. 19-30). Poria, S., Cambria, E., & Gelbukh, A. (2016). Aspect extraction for opinion mining with a deep convolutional neural network. Knowledge-Based Systems, 108, 42-49. Poria, S., Cambria, E., Ku, L. W., Gui, C., & Gelbukh, A. (2014, August). A rule-based approach to aspect extraction from product reviews. In Proceedings of the second workshop on natural language processing for social media (SocialNLP) (pp. 28-37). Radford, A., Wu, J., Child, R., Luan, D., Amodei, D., & Sutskever, I. (2019). Language models are unsupervised multi task learners. OpenAI blog, 1, 9. 248 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Samha, A. K., Li, Y., & Zhang, J. (2014). Aspect-based opinion extraction from customer reviews. arXiv preprint arXiv:1404.1982. Subrahmanian, V. S., & Reforgiato, D. (2008). AVA: Adjective-verb-adverb combinations for sentiment analysis. IEEE Intel igent Systems, 23, 43-50. Tian, H., & White, M. (2020, July). A pipeline of aspect detection and sentiment analysis for Ecommerce customer reviews. In Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval {SIGIR} (pp. 1-9). Venugopalan, M., & Gupta, D. (2022). An enhanced guided LDA model augmented with BERT based semantic strength for aspect term extraction in sentiment analysis. Knowledge-based systems, 246, 108668. Veyseh, A. P. B., Dernoncourt, F., Min, B., & Nguyen, T. H. (2022, July). Generating Complement Data for Aspect Term Extraction with GPT-2. In Proceedings of the Third Workshop on Deep Learning for Low-Resource Natural Language Processing (pp. 203-213). Wang, Q., Wen, Z., Zhao, Q., Yang, M., & Xu, R. (2021, November). Progressive self-training with discriminator for aspect term extraction. In Proceedings of the 2021 conference on empirical methods in natural language processing (pp. 257-268). Xu, H., Liu, B., Shu, L., & Yu, P. S. (2019). BERT post-training for review reading comprehension and aspect-based sentiment analysis. arXiv preprint arXiv:1904.02232. Yanuar, M. R., & Shiramatsu, S. (2020, February). Aspect extraction for tourist spot review in Indonesian language using BERT. In Proceedings of the 2020 International Conference on Artificial Intelligence in Information and Communication (ICAIIC) (pp. 298-302). THE ROLE OF THE SERVICE DESK FOR EHEALTH SOLUTIONS IN THE DIGITAL TRANSFORMATION OF SLOVENIAN HEALTHCARE ŽIVA RANT,1 HAJDI KOSEDNAR,1 DALIBOR STANIMIROVIĆ2 1 National Institute of Public Health, Ljubljana, Slovenia ziva.rant@nijz.si laleh.davoodi@abo.fi, hajdi.kosednar@nijz.si 2 University of Ljubljana, Faculty of Public Administration, Ljubljana, Slovenia dalibor.stanimirovic@fu.uni-lj.si The Service Desk carries out three basic tasks: general support for al users of the eHealth system, support in connecting to the secure healthcare network (zNET), and making electronic appointments for medical services. This article analyses the operation of the Service Desk through time dynamics, the content of requests dealt with and services provided during the COVID-19 pandemic. The article is based on the case study research methodology. The data showed accelerated growth in Keywords: the use of the Service Desk, with a more than seven-fold increase eHealth, recorded in 2021 relative to 2020. The biggest proportion of the user help, Service Desk requests in 2021 related to the zVEM Patient Portal eHealth and zNET. The Service Desk is one of the most important Service mechanisms for the successful development and use of the Desk, digital eHealth system in Slovenia, and the pandemic only made this transformation of even more evident and distinct. healthcare DOI https://doi.org/10.18690/um.fov.4.2023.16 ISBN 978-961-286-751-5 250 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The effective and comprehensive digital transformation of the Slovenian healthcare system is one of the fundamental changes that should contribute to greater success in dealing with numerous challenges facing Slovenia’s healthcare sector. Experiences in developed countries indicate (Bokolo, 2021; Arcury et al., 2020) that successfully implemented projects for healthcare digitalisation have exceptional strategic importance for further development of the healthcare system and considerable implications for increased social wel -being and economic growth (European Commission, 2018). The project of Slovenian healthcare digitalisation (eZdravje or eHealth), which follows the national and European guidelines, was one of the key long-term goals of the public sector in Slovenia. The management of the digital solutions that were created as part of the eHealth project was taken over in 2015 by the National Institute of Public Health (NIJZ). Up until then, the eHealth project was managed by the Ministry of Health, and a large proportion of the start-up funds was provided from the European Social Fund. An exhaustive review of the field indicates that the difficulties encountered right from the outset in the eHealth project stem on the one hand from the technological characteristics of the existing and rather fragmented healthcare information systems (HIS), which are a consequence of the uncoordinated development in the area of health informatics in recent decades. On the other hand, the responsibility for the existing state of affairs can be attributed principal y to the decision-making entities that left the development of health informatics in this period up to individual or institutional initiatives, needs and interests, without unified strategic guidelines. Moreover, decision-makers in this period have not been able to effectively promote the eHealth initiative or ensure stronger political (financial, HR, organisational) support through the formulation of a modern and consistent strategy in this area. The consequences of these factors are reflected in specific chal enges identified by the NIJZ in its activities for the successful implementation and management of the eHealth system: • the unpreparedness (administrative, technological, organisational, procedural, and so forth) of certain healthcare providers for appropriate use of the eHealth solutions; Ž. Rant, H. Kosednar, D. Stanimirović: The Role of the Service Desk for eHealth Solutions in the Digital Transformation of Slovenian Healthcare 251 • procedural, organisational, security, and user problems at healthcare providers; • the lack of professional consensus regarding the substantive issues (e.g. the healthcare service code (VZS), authorizations for accessing the Central Registry of Patient Data (CRPD)); • the narrow focus of individual stakeholders on their own professional field without being aware of the interdependency of al stakeholders in the healthcare ecosystem; • a lack of competent IT experts at healthcare providers who could ensure adequate maintenance and operation of the eHealth system; • inadequate funds for digitalisation projects at healthcare providers and the NIJZ, which works to ensure the development and maintenance of central national eHealth system components. Despite these chal enges, the lack of unified strategic documents and inadequate investment in the area, great progress has been made in the development and implementation of eHealth solutions in the past four years. The national importance of some eHealth solutions was widely acknowledged, such as the zVEM Patient Portal, which received the two biggest IT prizes in Slovenia: the prize for current achievements in the information society for 2022, and the 2022 "eReward" for the zVEM mobile application. The zVEM+ portal is a version of the zVEM portal and is only intended for use by healthcare professionals. As expected, and as it actual y turned out during the management of the eHealth system, a large number of users often need help or advice for various reasons. Accordingly, a special organisational unit called the Service Desk was established and is managed by an external provider. The Service Desk is intended for al eHealth system users who wish to report disruptions in operation, who need assistance or require information regarding the functioning of the eHealth system. In the meantime, the Service Desk has proven to be one of the most important components of eHealth, as it has directly or indirectly helped thousands of users and significantly contributed to the successful development of eHealth solutions, their implementation and increasing use. In line with the points set out above, this article focuses on the research question of how the COVID-19 pandemic impacted the functionality and use of the Service Desk. Accordingly, the article analyses the 252 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY operation of the eHealth Service Desk during the COVID-19 pandemic through the dynamic and content of requests dealt with and services provided. 2 Methods The article presents an analysis of the functionality and use of the eHealth Service Desk during the COVID-19 pandemic. Our research sought to answer the question of how the COVID-19 pandemic impacted the development and use of the Service Desk. During the pandemic, new circumstances appeared that required rapid adaptation, so the Service Desk underwent an accelerated evolution. This is an example of the digital service that experienced fast-track expansion process during COVID-19, which indicates that in this respect the pandemic was a special opportunity for development. Our research was based on the case study research methodology (Yin, 2018; Kljajić, 2021), which included an in-depth analysis of the field and its critical review. We used statistics from the administrative and business intel igence modules in the analysis to compare the number of events, cal s received and time taken in minutes in 2020 and 2021. We also compared data on the support provided for various eHealth solutions. The analysis was conducted on the one hand based on a review of the literature in this field (Lee, 2022), as well as on the examination of project documentation and the technical specifications for the Service Desk. On the other hand the study was carried out on the basis of observations, experiences and the expert opinion of professionals at the NIJZ who are in charge of the eHealth system and the Service Desk. In addition we used the actual statistical data to support the findings obtained (NIJZ, 2021). The selection of research methods was based on the particular features of the research field and the fact that the entire area of healthcare digitalisation in Slovenia is still in a relatively early stage, so there is just a narrow circle of experts with appropriate knowledge and experience in this field. This methodological approach enabled not just an insight into the current theoretical and technological basis for these kinds of digital solution, but also an empirical overview of the use of the Service Desk in the Slovenian healthcare system. The participation of experts from the NIJZ in the research provided an insight into the technological and statistical aspects of the operation, and also enabled a critical and thorough insight into the user aspects of the Service Desk, since the participating experts from the NIJZ are very familiar with the user experiences of patients and health workers in Ž. Rant, H. Kosednar, D. Stanimirović: The Role of the Service Desk for eHealth Solutions in the Digital Transformation of Slovenian Healthcare 253 the field, and with their satisfaction with the eHealth solutions. The analysis of the functionality and use of the Service Desk was conducted in the first half of 2023. Structured interviews with the NIJZ experts and the acquisition of statistical data from business and administrator modules were carried out between January and April 2023. The article focuses on the Service Desk principal y due to its importance both for patients and healthcare workers. The synthesis of findings from the literature, user functionalities, statistical reports and the views of experts from the NIJZ enables us to formulate credible conclusions based on verifiable data regarding the highlighted research aims (Lindgren et al., 2020). The use of the above-defined methodological framework, including a combination of various approaches and techniques for data gathering, was vitally important for the credibility of the analysis (Sim & Waterfield, 2019). The comprehensive analysis of data obtained from a diverse array of sources and from interviews with experts from the NIJZ ensured an appropriate platform for interpreting the data and formulating consistent conclusions regarding the research objectives (Thomas, 2021), which address the functionality and use of the eHealth Service Desk. 3 Results The Service Desk for eHealth users was considered in the very first strategic documents relating to eHealth. Since the transfer of eHealth to the NIJZ at the end of 2015, its role has become increasingly significant. This was especially evident during the COVID-19 pandemic, when the usual methods of training users and dealing with requests were not possible, and the development and usage of the eHealth solutions took a major leap forward. The Service Desk provides assistance for nearly 30 eHealth features and has three basic tasks: • general support for al users of the eHealth system; • support in connecting to the zNET; • making electronic appointments for medical services. 254 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY General support for al users of the eHealth system is intended for patients, healthcare workers, administrative staff, IT specialists, software providers and all other users of the eHealth system who wish to report disruptions in operation, who need help or require information relating to the functioning of the eHealth system. Support in connecting to the zNET offers assistance to healthcare providers who wish to connect to the national eHealth infrastructure and become a part of the secure healthcare network. The eAppointments system for medical services helps patients to make appointments electronical y. Users can access help by means of the online form at the web page, via email messages or by telephone for al eHealth features, for assistance with eAppointments, and for connection to the zNET. The website also provides answers to frequently asked questions. During the COVID-19 pandemic the Service Desk responsibilities were expanded to provide support for a range of new eHealth solutions. New solutions for healthcare providers concerned assistance in screening tests for COVID-19, entering test results for COVID-19 in zVEM+, the zVEM+ feature for COVID-19 entry points and support for patient information. The zVEM+ portal enables the capture of data, its processing, and the issuing of various reports that healthcare providers must send out. It is intended for providers, which do not use their own information system for this kind of operation (Rant, Stanimirović & Janet, 2022a; Rant, Stanimirović & Janet, 2022b). New eHealth solutions for patients included the EU Digital COVID certificate (EU DCP) and registration for vaccination against COVID-19. In addition to these features, special support was provided to patients in making eAppointments, since the operation of the entire healthcare system was extremely restricted. Support is provided by a permanent, experienced team of advisers specialised in various fields and features of eHealth. The advisers undergo constant training – both general and for specific features – and attend lectures, trainings and workshops. Videos are available for review and training of new co-workers. The Service Desk has built a massive knowledge base, keeping documentation and working to provide answers to frequently asked questions. The Service Desk website publishes notices related to the operation of the eHealth system and promptly publishes information about detected malfunctions and problems on the national level. Ž. Rant, H. Kosednar, D. Stanimirović: The Role of the Service Desk for eHealth Solutions in the Digital Transformation of Slovenian Healthcare 255 In the course of this research we used the statistical data from the administrative and business intel igence modules and compared the number of events, cal s received and time taken in minutes in 2020 and 2021 by the Service Desk. Data reveals that the number of Service Desk requests grew considerably during the COVID-19 pandemic (Figure 1). Figure 1: Ratio of Service Desk requests in 2020 and 2021 Additionaly we compared the number of Service Desk requests in 2021 concerning the different eHealth solutions (Table 1). Table 1: Percentage of Service Desk requests for different eHealth solutions in 2021 Percentage of requests eHealth solution % zVEM Patient Portal 69.06 ePrescription 0.32 eAppointment 4.80 CRPD 5.08 Vaccination registry 1.27 zNET 12.50 Security scheme 1.06 Application for preventive treatment 0.15 Teleradiology 0.12 Telestroke 0.05 eTriage 0.07 Other information 5.49 256 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY The biggest proportion of the Service Desk requests in 2021 related to the zVEM Patient Portal (Figure 2). 7000 6000 5000 4000 3000 2000 1000 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 2: Number of Service Desk requests related to the zVEM Patient Portal in 2021 The data also show that the Service Desk requests for help with eAppointments followed the epidemiological situation, as the numbers grew the most in May 2021, after the first wave of the COVID-19 pandemic (Figure 3). Figure 3: Service Desk requests for help with eAppointments Since January 2021, the number of calls for eAppointments has been increasing, with a significant jump in May of that year, as a result of the introduction of the new digital solution for vaccination against COVID-19, which citizens mistakenly Ž. Rant, H. Kosednar, D. Stanimirović: The Role of the Service Desk for eHealth Solutions in the Digital Transformation of Slovenian Healthcare 257 expected from the eHealth Service Desk. The increased volume of immaterial and irrelevant cal s, which continued in June, was causing problems and preventing the Service Desk from reaching those patients who real y needed help with eAppointments. In an attempt to free up the phone lines, we removed the eAppointment phone number from the entry page of the zVEM Patient Portal, as it was seen by many as the entry point for booking vaccinations. We also added congestion alerts to the website and additional instructions for patients looking for answers to other questions. A significant increase in the number of cal s to the Service Desk was detected after the introduction of eHealth solutions designed to monitor rapid tests for COVID-19, the introduction of vaccinations against COVID-19 and the issuance of EU digital COVID certificates. The existing team in the Service Desk was burdened more than usual, which affected the resolution of other eHealth requests. The provider of the Service Desk points out a really large number of requests for connecting to the zNET, with which they are burdened on a daily basis. The rush to join was especially intense in April, because it was decided that from May 2021 eHealth solutions can only be accessed through the zNET. It was pointed out that it is necessary to provide the contact information of IT specialists for many new users of the zNET. It is also evident that a greater proportion of requests are resolved at the Service Desk. However, due to the heavy load on the agents, there are occasional errors such as request redirection and missing data. In order to reduce the workload and stress of the Service Desk agents, we redesigned the informational website and supplemented it with updated contents. To connect to the zNET, a special phone number and email address were introduced. We analysed the increased volume of requests related to COVID-19 and prepared strategies to improve response times in resolving requests. A workshop for resolving requests related to IT issues and the zNET was held in April. Discussions are still taking place regarding Service Desk agents' access to the zVEM+ portal for user support purposes. 258 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 4 Discussion The research findings showed that the COVID-19 pandemic had a profound and extensive impact on the functionality and use of the Service Desk. The COVID-19 pandemic showed how important the digitalisation of healthcare is (Stanimirović & Matetić, 2020), and in particular how the existence of a national infrastructure and central systems is vital for digitalisation. This was the case in developing the eHealth solutions in Slovenia. The COVID-19 pandemic circumstances and the inaccessibility of health services has accelerated the use of digital solutions, including eHealth solutions. This demanded the upgrades of numerous eHealth solutions, which needed to be developed and implemented in the shortest time possible. Some digitalisation objectives were achieved through existing solutions with some adaptations, while certain solutions needed to be reworked, and some needed to be developed completely from scratch. In the 2020-2021 period two solutions in particular made great progress – the zVEM Patient Portal and the CRPD (Rant et al., 2022a). On the zVEM Patient Portal the number of registered users grew nine-fold (925%) in 2021 relative to 2020, and the number of visits grew 12 times (1,273%). The number of documents in the CRPD grew eight-fold in 2021 relative to 2019 (849%) (Rant et al., 2022b). The rapid development of digital solutions and the exceptional y short time for implementation, which did not allow for the usual training of users and dealing with requests, generated unexpected pressure at the Service Desk. To a large extent, the pressure came from the users’ and general public's lack of information and high expectations of eHealth solutions created by the media and politics, as wel as the general public. Data on the number of requests received by the Service Desk reveal the exceptional growth in activity in the last two years. The number of events in 2021 grew more than seven times (766%), as did the number of calls received (742%), and the time taken in minutes grew five-fold (493%) (Figure 1). The highest growth is visible in May and June 2021 (Figure 2), which is also related to the acquisition of the EU Digital COVID certificates via the zVEM Patient Portal and the introduction of vaccinations and testing at healthcare providers. In May 2021 for instance, 80% to 90% of requests related to the zVEM Patient Portal or to questions regarding the EU DCP (Figure 2). This proportion declined later on. Ž. Rant, H. Kosednar, D. Stanimirović: The Role of the Service Desk for eHealth Solutions in the Digital Transformation of Slovenian Healthcare 259 One of the reasons for such demand for assistance with the zVEM Patient Portal is certainly help from the Service Desk agents in obtaining the EU DCP, which was required by the strict rules on movement in Slovenia during the COVID-19 pandemic. A significant advantage is that on the part of the agents, a real human answered the cal s and not an automated system. The agents could adapt to users, their knowledge, experience and level of digital literacy. In this way, users obtained assistance firstly in accessing the zVEM Patient Portal, and fol owing this, the agents would walk users step by step to finally obtaining the EU DCP. User satisfaction can be seen in the numerous expressions of gratitude they offered. In providing support, the agents encountered many chal enges and difficulties. Without a doubt, the greatest difficulty was the unpredictability of the daily and monthly workload, for which it was hard to prepare. There was a major increase in the number of cal s and requests during the COVID-19 pandemic, and the Service Desk workload increased radical y with the introduction of new features (e.g. the zVEM Patient Portal and the COVID certificate). Users needed help with a variety of digital solutions, including eAppointments for vaccination, aAppointments for examinations, obtaining referrals and prescriptions, access to results and discharge letters. All this was available to them on the zVEM Patient Portal, for which it was mandatory to obtain a qualified digital certificate, and later the SI-PASS. The Service Desk agents led users through what are at times the truly complex and time-consuming procedures of registration for various online services. One of the difficulties encountered by the Service Desk is the large number of cal s that are not related to the eHealth systems, but are principal y related to the COVID-19 pandemic. Such calls cause a notable burden at the Service Desk and the unavailability of the phone lines, which in turn causes dissatisfaction among those users who real y need help and who must frequently wait longer for assistance. During the pandemic the Service Desk received general questions about health, questions about COVID-19 treatment and medications, and about vaccinations, vaccines and side effects. Moreover, patients contacted the Service Desk with questions about the measures in place and restrictions on movement. Al these questions, requests and claims undermined the activities and main purpose of the Service Desk, which is to support users of the eHealth system, and not to provide 260 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY the public with answers to general questions about COVID-19 and related epidemiological measures. 4 Conclusion Since the completion of the project in 2015 the national eHealth system in Slovenia has undergone unimagined development. The whole national eHealth architecture underwent particularly accelerated development during the COVID-19 pandemic. In this period of such extensive development of new eHealth solutions and rapid growth in use (more than ten-fold), the Service Desk played a key role. Its proficient and unstinting assistance to users, be it citizens or health professionals, facilitated the effective use of the new and already existing eHealth solutions. The latter findings revealed the important role of the Service Desk in the field of healthcare digitalisation, especially in crisis situations and when introducing new digital solutions. On the other hand, the huge increase in requests received by the Service Desk during this period revealed that eHealth solutions were not wel promoted and that a large number of citizens and health professionals do not have the skil s to use digital solutions. These findings obligate managers of the national eHealth system to invest more in eHealth infrastructure, especial y in such important components as the Service Desk. At the same time, the findings imply that the efforts and resources al ocated to the promotion of eHealth solutions, education and training of citizens and professionals so far, have been decidedly insufficient and that the results of these activities have not yielded the desired results. It seems that the COVID-19 pandemic has done more to raise awareness and usage of eHealth solutions in a very short period of time than any other initiative before, be it of a political, legislative, administrative or financial character. Given this alarming fact, there should be a thorough examination and discussion of what activities are needed to address the exposed issues, otherwise we might largely undermine al efforts and achievements so far and significantly compromise the further development of healthcare digitalisation projects. Ž. Rant, H. Kosednar, D. Stanimirović: The Role of the Service Desk for eHealth Solutions in the Digital Transformation of Slovenian Healthcare 261 References Arcury, T. A., Sandberg, J. C., Melius, K. P., Quandt, S. A., Leng, X., Latulipe, C., . . & Bertoni, A. G. (2020). Older adult internet use and eHealth literacy. Journal of Applied Gerontology, 39(2), 141-150. Bokolo, A. J. (2021). Application of telemedicine and eHealth technology for clinical services in response to COVID‑19 pandemic. Health and Technology, 11(2), 359-366. European Commission. (2018). Communication from the Commission to the European Parliament, the European Council, the Council, the European Economic and Social Committee and the Committee of the Regions on enabling the digital transformation of health and care in the Digital Single Market; empowering citizens and building a healthier society. SWD (2018) 126 final. Brussels. Kljajić Borštnar, M. (2021). Raziskovanje informacijskih sistemov. . Fakulteta za organizacijske vede. https://estudij.um.si/pluginfile.php/676575/mod_resource/content/2/U%C4%8Dbenik% 20raziskovalna%20metodologija%20Kljaji%C4%87%20Bor%C5%A1tnar%20Mirjana%2020 21.pdf Lee, W. L., Lim, Z. J., Tang, L. Y., Yahya, N. A., Varathan, K. D., & Ludin, S. M. (2022). Patients' technology readiness and eHealth literacy: implications for adoption and deployment of eHealth in the COVID-19 era and beyond. CIN: Computers, Informatics, Nursing, 40(4), 244-250. Lindgren, B. M, Lundman, B., Graneheim, U. H. (2020). Abstraction and interpretation during the qualitative content analysis process. Int J Nurs Stud, 108: 103632. https://doi.org/10.1016/j.ijnurstu.2020.103632. Nacionalni inštitut za javno zdravje (2021). Statistika zahtevkov na Prvem nivoju podpore uporabnikom eZdravja. Nacionalni inštitut za javno zdravje, Ljubljana. Rant, Ž., Stanimirović, D., & Janet, J. (2022a). Functionalities and use of the zVEM patient portal and the central registry of patient data. In A. Pucihar, M. Kljajić Borštnar, R. Bons, A. Sheombar, G. Ongena, & D. Vidmar (Eds.), 35th Bled eConference Digital Restructuring and Human (Re)action: June 26 – 29, 2022, Bled, Slovenia (pp. 65–79). University of Maribor, University Press. doi:10.18690/um.fov.4.2022.4 Rant, Ž., Stanimirović, D., & Janet, J. (2022b). Razvoj portala zVEM in Centralnega registra podatkov o pacientu. In P. Šprajc, D. Maletič, N. Pavlović, I. Podbregar, A. Škraba, D. Tomić, U. Vincenzo, & A. Žnidaršič (Eds.), 41th International Conference on Organizational Science Development: societyʹs chal enges for organizational opportunities (pp. 873–884). University of Maribor, University Press. doi:10.18690/um.fov.3.2022.63 Sim, J, Waterfield, J. (2019). Focus group methodology: some ethical chal enges. Quality & Quantity. 2019 Jul;53(6):3003-3022. Stanimirovic, D, Matetic, V. (2020). Can the COVID-19 pandemic boost the global adoption and usage of eHealth solutions?. J Glob Health, 10(2): 0203101. https://doi.org/10.7189/jogh.10.0203101. Thomas G. (2021). How to do your case study. Thousand Oaks, CA: Sage Publications. Yin, R. K. (2018). Case study research and applications: design and methods (6th ed.). Sage. 262 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY ADAPTIVE LEARNING TECHNOLOGIES IN BLENDED LEARNING DESIGN: HOW DO STUDENTS AND TEACHERS USE THIS TECHNOLOGY IN PRACTICE? RICHARD BAKKERS,1 MARIOLA GREMMEN,2 ESTHER VAN DER STAPPEN1 1 Avans University of Applied Sciences, Research Group Digital Education, Centre of Applied Research Future-Proof Education, Breda, Netherland rwa.bakkers@avans.nl, ej.vanderstappen@avans.nl 2 HAN University of Applied Sciences, iXperium Centre of Expertise Learning with ICT, Nijmegen/Arnhem, Netherland mariola.gremmen@han.nl Blended learning is adopted as the teaching method in an increasing number of higher education institutes worldwide. Adaptive learning technology (ALT) can be incorporated in such learning designs, especial y to support students' a-synchronous, individual learning activities. In this empirical, mixed-method study, we investigated in what way teachers and students use the provided functionalities of a specific ALT to realise a blended course design. We interviewed four teachers delivering the same course using the ALT and we quantitatively analysed student trace data from the ALT log system. Our results show that teachers do recognize the added value of employing an ALT, but they do not realise its full potential by lack of usage of the dashboard, knowledge sharing among teachers and too Keywords: little attention for (meta)cognitive and social support of students. The blended trace data analysis shows that students display cramming behaviour (no learning, active repetition and/or spaced practice), they are selective in which learning learning, objectives they study and a majority chooses a suboptimal learning path. adaptive Based on our results, we conclude that, in the case we studied, the ful learning potential that the ALT offers is not realised to the benefit of students, technologies, case since both teachers and students show suboptimal behavior. We give study, recommendations for practice and future research based on our Bled conclusions. eConference DOI https://doi.org/10.18690/um.fov.4.2023.17 ISBN 978-961-286-751-5 264 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Blended education is high on the agendas of educational institutions, especially after the Covid pandemic (Bruggeman et al., 2022; Dziuban et al., 2018). By combining the strengths of different learning environments, students can have a richer learning experience. In order to provide students with a rich learning environment in which they actively learn independent of time and/or place, a-synchronous learning activities are often part of students’ learning arrangements. The deployment of adaptive learning technologies for these a-synchronous learning activities has grown rapidly and these technologies are increasingly used in educational settings. Real-time interaction is not needed in those learning activities, as students interact when they want and at their own pace with course content through, among others, screencasts, exercises, and discussion boards. As a consequence, in theory there is more time at school for students to interact about course contents with peers and teachers synchronously. In practice, we often see that students and teachers do not take enough advantage of the used adaptive technologies. Students do not always prepare enough to effectively use the time in the classroom for further understanding of al learning content. And teachers do not always (know how to) use al provided information in the dashboards about students’ progress to actively (change the) focus of the live interactions to the topics that students consider difficult. A main goal of adaptive technologies is supporting active learning. Therefore, the current study focuses on teachers’ possibilities to reach this goal by using adaptive learning technologies in a course design and aims to answer the question: “How do teachers and students use adaptive technologies in practice and what can they do to contribute to active learning by students?” 2 Theoretical framework 2.1 Blended and Active Learning Hrastinski analysed the many definitions of blended learning and came to the conclusion that blended learning has become “an umbrel a term” (Hrastinski, 2019). Therefore, he recommends explicitly stating what is understood by blended learning R. Bakkers, M. Gremmen, E. van der Stappen: Adaptive Learning Technologies In Blended Learning Design: How Do Students and Teachers Use This Technology in Practice? 265 in the context of specific research. In addition, research at the higher education institute in the Netherlands where this study took place has concluded that many of the definitions of blended learning (Kat-De Jong, 2021): (1) place too much emphasis on the distinction between physical and online education; and/or (2) place too much emphasis on the ICT aspect; and/or (3) insufficiently reason from the student’s perspective. Subsequently, within this institute blended learning has been defined as: “Providing a rich learning experience as a result of a deliberate, integrated, and harmonious combination of synchronous and a-synchronous learning activities in which students participate remotely, on campus, and in the workplace.” . Bernard et al. (2014) conclude from various comparative studies that a blended learning arrangement leads to better results for students (e.g., regarding motivation, engagement, and grades) than a ful y offline or a ful y online setup. In addition, Van der Stappen (2022) concludes that the added value of blended learning can be achieved precisely then when effectively combining the strengths of different learning environments, i.e., on campus, in the online learning environment, and in the workplace. Various international meta-studies have studied the added value of blended learning, which can be summarized as follows (Last & Prinsen, 2021): (1) Blended learning can increase students’ engagement and learning ef iciency; (2) Blended learning offers flexibility in place and (partly) time of learning, e.g. for (the growing group of) international and working students; (3) Blended learning can potential y maximize the benefits of several places while reducing their weaknesses by combining online and face-to-face learning activities; and (4) Blended learning can increase students' confidence in their own abilities (self-efficacy) and intrinsic motivation. Active learning is the central concept within the active (blended) learning ecosystem (Hedgepath, 2014; Last & Jongen, 2023). If wel -designed, blended learning motivates and activates students. This implies that students need to take ownership of their own learning process and, thus, need to engage in self-regulated learning (Jansen, 2021). However, self-regulated learning skills do not develop naturally, and research shows that the vast majority of students struggle to actually self-regulate their learning properly (Bol & Garner, 2011). In order for students to actual y achieve a state of active learning, al three aspects of the aforementioned ecosystem (i.e. 266 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY pedagogy, physical and virtual space, and technology) are to be considered integraly by teachers when designing and evaluating blended learning arrangements that support active learning. 2.2 Community of Inquiry Although not developed specifical y for blended learning, the Community of Inquiry framework is one of the most influential blended learning models (Hrastinski, 2019). It has been argued that it is useful for understanding and designing blended learning due to the generic nature of the framework, and the resonance with both face-to-face and online learning (Garrison & Vaughan, 2008). The Community of Inquiry framework represents a process of creating a deep and meaningful (collaborative-constructivist) learning experience through the development of three interdependent elements (Garrison et al., 2001): • Social presence: the ability of learners to project their personal characteristics into the Community of Inquiry, thereby presenting themselves as “real people”. Intervention categories (Arbaugh et al., 2008): affective expression, open communication, and group cohesion; • Cognitive presence: the extent to which the participants in any particular configuration of a Community of Inquiry are able to construct meaning through sustained communication. Intervention categories (Arbaugh et al., 2008): activating activities, exploration, integration, and completion; • Teaching presence: the design, facilitation, and direction of cognitive and social processes for the purpose of realizing personal y meaningful and educational worthwhile learning outcomes. Intervention categories (Arbaugh et al., 2008): design & organization, facilitating discourse, and direct instruction. The idea behind this framework is that teachers create a blended learning environment encompassing a strong (interaction between) social, cognitive and teaching presence. R. Bakkers, M. Gremmen, E. van der Stappen: Adaptive Learning Technologies In Blended Learning Design: How Do Students and Teachers Use This Technology in Practice? 267 2.3 Adaptive Learning Technologies With Covid and home-schooling as major accelerators, the deployment of adaptive learning technologies has increased significantly. A learning environment is adaptive “…to the degree that (a) its design is based on data about common learner chal enges in the target subject matter, (b) its pedagogical decision-making changes based on psychological measures of individual learners, and (c) it interactively responds to learner actions.” (Aleven et al., 2016). Although adaptive learning technologies have been considered an important trend in education, they still have not reached their full potential as of yet, possibly since there has not been enough attention for the role of these technologies in the design of rich learning experiences (Rivera Muñoz et al., 2022; Weber et al., 2019). In terms of blended learning, adaptive learning technologies are mainly used for a-synchronous learning activities. A-synchronous learning does not require real-time interaction. Students interact with the course content at their own pace, when and where they want, through e.g., learning modules, discussion boards, or pre-recorded videos. Assuming that students engage with the materials in line with the teacher’s blended course design, the idea is that students arrive wel -prepared at the synchronous, on campus learning activity (e.g., a classroom lecture or workshop). That would result in more time to interact about course contents with peers and teachers synchronously in order to deepen and/or broaden the students’ knowledge. The design of such a blended learning arrangement is frequently illustrated as a “blended wave”. A fictitious example is presented in Figure 1 (SURF, 2022). Figure 1: Example of a blended learning arrangement designed as a “blended wave” 268 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3 Methodology In this study, we want to understand better if and how a course – designed as a “blended wave” and supported by adaptive learning technology (ALT) – works in practice. Ultimately, we want to identify opportunities for (blended) educational improvements. Therefore, we answer the fol owing research question: “How do teachers and students use adaptive technologies in practice and what can they do to contribute to active learning by students? To answer this question, we performed a mixed-method single case study (Yin, 2018). Below, we describe the real-world educational context in which the study took place, as wel as the participants in the study, and the instruments and the methods that were used. 3.1 Context The study was performed at a large university of applied sciences in the Netherlands with a focus on the course Introduction to Business Economics in the fall of 2021. This course is offered in both the Human Resources Management (“HRM”) and Business Administration (“BA”) Bachelor programmes.,during the first quarter of the propaedeutic phase, hence the students in question are on average 17-18 years of age. The blended learning activities took place during seven consecutive teaching weeks. During these seven weeks, students engaged in the adaptive tool (a-synchronous) and had one (synchronous) lecture per week. Neither engaging in the adaptive tool nor participating in the lectures was compulsory. The eighth week was used for preparing (a-synchronously) for the written exam. This exam was taken in the ninth week. The ALT (software) in this study is provided by a Dutch company: ABC (fictious name). ABC's software supports setting up the aforementioned “blended wave” (see Figure 1). For a student, ABC's software includes among others a question module. For a teacher, ABC’s software includes among others a teacher dashboard. The course in scope of our research is constructed along twelve learning objectives (subjects/themes). These learning objectives have a suggested order (learning objective sequence 1-12 as captured in the trace data; see section 3.2). However, the R. Bakkers, M. Gremmen, E. van der Stappen: Adaptive Learning Technologies In Blended Learning Design: How Do Students and Teachers Use This Technology in Practice? 269 software does not prescribe a student to start with a specific learning objective in the software, nor which order to follow. Hence, a student can start with learning objective 11 and subsequently go to learning objective #5, from there to #7, etc. By answering questions in the question module, a student can increase the so-cal ed “ABC score”. This is a grade number (1-10) that reflects the students’ progress per learning objective and across al learning objectives. The ABC score can be increased or decreased, according to an algorithm, by answering questions in the module correctly or incorrectly. In maximal y four steps, the students receive feedback on their incorrect answer attempts to instantly help the student proceed; this is referred to as “scaffolding” (Rivera Muñoz et al., 2022). The ABC score is also disclosed to a teacher via the aforementioned teacher dashboard. Appendix A contains a further explanation about the ALT in this case study. 3.2 Research Design The research is designed as a mixed-method single case study, where the unit of analysis is the execution of a blended course design using ALT for a-synchronous activities. We applied triangulation (Cohen et al., 2018) in the data col ection in several ways: both qualitative and quantitative methods were used, and data was col ected from both teachers and students involved in the same course. Semi-structured interviews were conducted with four teachers involved in the course (Table 1). Each interview was audio-recorded, ful y transcribed and coded using open coding (Flick, 2018). The interview questions – reviewed by a second researcher – were targeted at teacher knowledge of adaptive learning and the associated technology, their views on the three aspects of the Community of Inquiry Framework, and the usage of the specific ALT applied in the course they were teaching. All teachers were informed of the purposes of our study, the data management and al gave recorded oral consent. Moreover, a quantitative analyses on the log trace data extracted from the ALT which logs most of the students’ online actions was performed. The original data file (Excel format) was retrieved anonymised from ABC. The trace data reflects the logging of all student activities in the question module in the period September 2021 to January 2022 for the course. Please refer to Appendix B for a further explanation about the 270 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY original trace data as well as the cleaning and augmentation actions performed by the researchers. The data analysis was limited to activities in week 1 up to and including week 9 (written exam). The activities of nine students were discarded from the trace data set, either because these students were only enrol ed to re-take the exam after failing the year before, or because they were not assigned to a class group which was supervised by a teacher. The final data set thus contains the data of 273 - 9 = 264 (mostly first year) students. No personal information (e.g., gender or age) was col ected, since it was deemed unnecessary to answer the research question in this study. The students gave consent for the use of their log data in the system for research and improvement purposes when enrolling in the ALT system. Table 1: Participant (interviewee) information Interviewee Teacher Used ALT In Scope Experience Before? A 5 years Yes, 4 years B 8 years Yes, 4 years C 9 years Yes, 3 years D 15 years No 4 Results 4.1 Interviews with teachers 4.1.1 Ambiguous definitions The interviewees mentioned various definitions of adaptive learning/adaptivity. The answers had in common that teachers were thinking about more student-tailored education. A few examples can be found in Table 2. R. Bakkers, M. Gremmen, E. van der Stappen: Adaptive Learning Technologies In Blended Learning Design: How Do Students and Teachers Use This Technology in Practice? 271 Table 2: Quotes on the definition of adaptive learning Interviewee Quote A "Thinking along, giving tips, supporting/promoting the learning process, customisation, individual and immediate insight into ‘right or wrong’ but especial y into why this was the case, adapting to level (of questions), different ways of learning (reading, videos), own pace, own moment." B “…tailoring your programme to your students. Customisation. And that means you underrstand how a particulat student learns and you take this as a starting point." C ".. that students can partly learn at their own pace. So working on something by themselves and using that to find out for which topic(s) there is a need for help...so that you can match what each individual student needs." B+D "...time-independent and location-independent, self-paced learning." 4.1.2 Motivations for deployment of the ALT The interviewees mentioned several reasons for deploying an ALT such as the one deployed in this study. These vary from provider’s sales push to the belief that the language and word usage in the ALT are much more in line with today's students. Most predominant is the belief that an ALT supports the concept “students in the lead": moving from consuming to demand-driven learning. Furthermore, direct, concrete feedback provided by the software means students do not have to wait (long) for a response from a teacher and can move on more quickly. Direct, concrete feedback can avoid misconceptions: “Suppose the answer is EUR 100,000, a student may start working towards this answer, but not in the correct way. He/she therefore learns an incorrect method at that point. The ALT prevents this. ” Another important motivation is referred to as "no ballast": “. .the ALT compresses the material to its essence; in the previously used textbook, about 2/3rds of the content is not covered in the course. ” 4.1.3 Teachers recognize added value, but also make critical comments The perceived benefits of using an ALT by the interviewees play an important role in supporting the deployment of an ALT. However, the interviewees also made some critical comments. Most importantly, the ALT is constructed along micro learning objectives. This leads to students “…continue to ‘think micro': students struggle to make connections between the individual learning objectives and/or to oversee the whole course with a helicopter view. Furthermore, students quickly get used to the way questions are formulated in the 272 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY ALT’s question module”. In the ALT, topics as wel as questions are structured per (micro) learning objective. However, during their written exam, students need to combine al information by solving an integrated case. As many students face difficulties with these kinds of cases, teachers need to explicitly address those while offering the blended learning arrangement. Otherwise the required constructive alignment is not in place (Biggs & Tang, 2011). To overcome this, as stated in the interviews, teachers started experimenting with so-cal ed “integration case studies”, which have been aligned with the ALT’s supplier. These case studies are discussed during the synchronous, on campus, lessons. The preliminary results, based on feedback about the deployment of these “integration case studies" from both teachers and students, are promising. 4.1.4 Teaching presence dominant The interviewees heavily focussed on interventions related to teaching presence. This is mainly due to the tangibility of and familiarity with the intervention categories related to teaching presence, i.e.: design & organisation, teaching facilitation and direct instruction (Arbaugh et al, 2008). For each of these aspects, without having to think for long, teachers named concrete examples of teaching presence related interventions in the context of this course. This applies to a much lesser extent to social and cognitive presence. When asked, these two concepts are not directly recognised and the teachers make only (very) limited use of conscious, explicit interventions to promote social and cognitive presence. The deployment of the aforementioned integration case studies is the main example provided by the interviewees concerning cognitive presence. Three of the interviewees explicitly indicated that no specific a-synchronous social presence interventions were organised during this series of lessons, neither to interact with peers nor with teachers. However, research shows that students “…value , above al , regular synchronous and a-synchronous interactions with peers, tutors…” (Armellini et al., 2021). R. Bakkers, M. Gremmen, E. van der Stappen: Adaptive Learning Technologies In Blended Learning Design: How Do Students and Teachers Use This Technology in Practice? 273 4.1.5 Teacher dashboard could be used more effectively The interviewees indicated that the capabilities of the teacher dashboard were only partial y used. Teachers reported using the dashboard mainly in the first two weeks. In doing so, some teachers chose to open the dashboard in class ("confrontation"), others to address students 1-on-1. After a few weeks, the use of the teacher dashboard decreased. Reason are that a certain " dashboard fatigue" had developed and, in addition, the effect on the students by (publicly) displaying the dashboard seemed to be had worn off by then. Within the teacher dashboard, the ABC score (grade; 1-10) was mainly looked at, much less at the number of activity and attempts. Therefore, the teachers missed relevant information such as: (1) The questions/learning objectives for which the students needed the most response attempts in order to reach a sufficient result; (2) Students who were ‘skipping’ learning objectives; and (3) Students who were going through the various learning objectives in an undesirable sequence. 4.1.6 Teachers did not sufficiently share or discuss As a group of teachers, it is important to share and discuss beliefs (or obstacles) considering the aforementioned elements (Gremmen, 2022). Ideal y, this is done on a regular basis before (designing), during (adjusting/fine-tuning), and after (evaluating) the series of lessons. The interviewees indicated that mutual discussions heavily focused on course content instead of on beliefs (or obstacles) about active and blended learning in general, adaptive learning (techniques), cognitive and social presence interventions, dashboard deployment, etc. 4.2 Student trace data analysis The trace data from the ALT was analysed from various perspectives such as number of activities (number of questions answered and number of response attempts per question), learning objectives, and ABC score. The results show a significant drop in students’ activities in the ALT after week 2, lasting until week 7 (see Figure 2). The written exam was in week 9; the higher activity in week 8 might indicate a- synchronous last-minute exam preparation. This cramming behaviour is not in line with recommendations from research on effective (blended) learning strategies, such as repetition and spaced practice (Yeung et al., 2021). 274 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Figure 2: Number of activities and attempts performed by students per week The ALT does not prescribe a student to start with a specific learning objective in the software, nor which order of learning objectives to follow, whereas the learning objectives do in fact build on the content of the previous one(s). Therefore, it was analysed which learning objective students chose to study in the system, and in which order. Here, visiting a learning objective (i.e., attempting at least one question related to this objective) is distinghuished from completing it (i.e., gaining a sufficient ABC score of at least 6.0). Just 136 out of 264 (51%) of the students visited all learning objectives and only 156 (60%) of the students started the final learning objective #12 (See Figure 3) Therefore, it was analysed which learning paths (=learning objective sequence) students chose; see Figure 3 and Figure 4. In Figure 4, al chosen learning paths are shown as line plots, with line thickness representing the number of students chosing that path. Most students stayed close to the default path, but some outliers exist, choosing for example the learning path (#5, #7, #1, #2, #9, #10, #11). 111 students chose the default path, i.e., #1 through #12. Nonetheless, there is a great variety in learning paths, e.g., 13 students never visited learning objective #1, one student started with #6, and one student visited #11 as his/her second objective. R. Bakkers, M. Gremmen, E. van der Stappen: Adaptive Learning Technologies In Blended Learning Design: How Do Students and Teachers Use This Technology in Practice? 275 Figure 3: Students visiting a given learning objective at a given rank Figure 4: Learning paths chosen by students 276 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Finaly, the maximum ABC scores that students achieved on each of the learning objectives to compare scores between students that chose the default path and those that chose different paths was analysed. The complete results are shown in Table 3. An unpaired, two-sided T-test was performed. The difference between the average ABC score for default paths (Mdefault=7.70, SDdefault=0.80) and the average ABC score for different paths (Mdifferent=6.28, SDdifferent=1.80) is statistically significant, t222.90485=8.65, p=1.00e-15, 95% CI [1.0964, 1.7431]. The results in Table 3 and the T-test point in the direction that students that chose the default learning path outperformed those that didn’t, in terms of ABC score. Table 3: Absolute and relative numbers of students scoring (in)sufficient on learning objectives 5 Conclusions and Discussion In this study, we aimed to answer the question: ‘ How do teachers use adaptive technologies and what can they do to contribute to active learning by students? ’ by interviewing teachers and analysing student trace data. Results show that both groups do not utilize the ALT in question as intended, i.e., to realize active learning through a blended wave in the course design. This is in line with other studies on ALTs, which were not specifically focused on blended learning designs (Harati et al., 2021). Our study adds to the knowledge base on ALT, since studies tend to focus more on technology than on learning (Rivera Muñoz et al., 2022). R. Bakkers, M. Gremmen, E. van der Stappen: Adaptive Learning Technologies In Blended Learning Design: How Do Students and Teachers Use This Technology in Practice? 277 Teachers show prominent teaching presence where cognitive and social presence are underrepresented. For an optimal blended learning experience, teachers should employ a balanced mix of these three presences, with cognitive presence being a strong indicator of students’ satisfaction (Giannousi & Kioumourtzoglou, 2016). Students display cramming behaviour (no repetition and/or spaced practice), they are selective in which learning objectives they study and a majority chooses a suboptimal learning path. Based on our research, we formulate some recommendations for educational practice. First, we advise teachers to actively integrate (insights from) the ALT in their synchronous activities, utilising the dashboard, and to use interventions directed at social and cognitive presence such as instruction on effective learning strategies or open communication aimed at community building (Biwer et al., 2020). These recommendations are in line with the recommendations from a recent study Müller et al. (2023). In addition, teachers could share their beliefs and strategies related to active, blended learning and ALTs when implementing a blended course with ALT to learn from each other. The ALT system could be extended to help students overcome drops in engagement by sending reminder notifications or employ (gamified) nudges. In the previous months, we conducted an exploratory study with students to assess the technical feasibility of such nudges based on machine learning analyses, and the preliminary results are promising. Our study has some limitations. It took place in a specific context, where characteristics of teachers, students and technology may have influenced the results. The student participants group was quite large (264), but a mere four teachers were interviewed. Moreover, quantitative data was not available on teacher behaviour, and we did not include qualitative student data . To conclude, we identify some interesting directions for future research. First, we would like to perform the same analyses in other educational contexts, in terms of study domain, teacher and students’ characteristics, and ALT used. Second, a more qualitative study directed at students’ perspective and behaviour could add to the insights presented in this paper. Finally, since our results show that the behaviour of both teachers and students was suboptimal, we think studies on interventions to help both groups to engage in more effective teaching and study behaviour would be very relevant to increase student success. 278 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Acknowledgements The authors would like to thank Lotte van Dijk (Avans Universitiy of Applied Sciences, Institiutional Research) for her contribution to the trace data analysis. The authors would also like to thank ABC for providing the log trace data. References Aleven, V., Mclaughlin, E., Glenn, A., & Koedinger, K. (2016). Instruction Based on Adaptive Learning Technologies. Arbaugh, J. B., Cleveland-Innes, M., Diaz, S. R., Garrison, D. R., Ice, P., Richardson, J. C., & Swan, K. P. (2008). Developing a community of inquiry instrument: Testing a measure of the Community of Inquiry framework using a multi-institutional sample. The Internet and Higher Education, 11(3–4), 133–136. https://doi.org/10.1016/j.iheduc.2008.06.003 Armel ini, A., Teixeira Antunes, V., & Howe, R. (2021). Student Perspectives on Learning Experiences in a Higher Education Active Blended Learning Context. TechTrends, 65(4), 433–443. https://doi.org/10.1007/s11528-021-00593-w Bernard, R. M., Borokhovski, E., Schmid, R. F., Tamim, R. M., & Abrami, P. C. (2014). A meta-analysis of blended learning and technology use in higher education: From the general to the applied. Journal of Computing in Higher Education, 26(1), 87–122. https://doi.org/10.1007/s12528-013-9077-3 Biggs, J. B., & Tang, C. S. (2011). Teaching for quality learning at university: What the student does (4th edition). McGraw-Hil , Society for Research into Higher Education & Open University Press. Biwer, F., Egbrink, M. G. A. oude, Aalten, P., & de Bruin, A. B. H. (2020). Fostering effective learning strategies in higher education—A mixed-methods study. Journal of Applied Research in Memory and Cognition, 9(2), 186–203. https://doi.org/10.1016/j.jarmac.2020.03.004 Bloom, B. S., Engelhart, M. D., Furst, E. J., Hil , W. H., & Krathwohl, D. R. (1956). Taxonomy of educational objectives: The classification of educational goals: handbook I: cognitive domain. New York, US: D. Mckay. Bol, L., & Garner, J. K. (2011). Challenges in supporting self-regulation in distance education environments. Journal of Computing in Higher Education, 23(2–3), 104–123. Bruggeman, B., Hidding, K., Struyven, K., Pynoo, B., Garone, A., & Tondeur, J. (2022). Negotiating teacher educators’ beliefs about blended learning: Using stimulated recal to explore design choices. Australasian Journal of Educational Technology, 100–114. https://doi.org/10.14742/ajet.7175 Cohen, L., Manion, L., & Morrison, K. (2018). Research methods in education (Eighth edition). Routledge. Dziuban, C., Graham, C. R., Moskal, P. D., Norberg, A., & Sicilia, N. (2018). Blended learning: The new normal and emerging technologies. International Journal of Educational Technology in Higher Education, 15(1), 3. https://doi.org/10.1186/s41239-017-0087-5 Flick, U. (2018). An introduction to qualitative research (6th edition). SAGE Publications. Garrison, D. R., Anderson, T., & Archer, W. (2001). Critical thinking, cognitive presence, and computer conferencing in distance education. American Journal of Distance Education, 15(1), 7–23. https://doi.org/10.1080/08923640109527071 Garrison, D. R., & Vaughan, N. D. (2008). Blended learning in higher education: Framework, principles, and guidelines (1st ed). Jossey-Bass. Giannousi, M., & Kioumourtzoglou, E. (2016). Cognitive, Social, and Teaching Presence as Predictors of Students’ Satisfaction in Distance Learning. Mediterranean Journal of Social Sciences. https://doi.org/10.5901/mjss.2016.v7n2s1p439 R. Bakkers, M. Gremmen, E. van der Stappen: Adaptive Learning Technologies In Blended Learning Design: How Do Students and Teachers Use This Technology in Practice? 279 Gremmen, M. (2022). What’s on teachers minds? Teachers’ helping and hindering beliefs about blended education. European Association for Practitioner Research on Instruction and Learning, Nijmegen. Harati, H., Sujo-Montes, L., Tu, C.-H., Armfield, S., & Yen, C.-J. (2021). Assessment and Learning in Knowledge Spaces (ALEKS) Adaptive System Impact on Students’ Perception and Self-Regulated Learning Skills. Education Sciences, 11(10), 603. https://doi.org/10.3390/educsci11100603 Hedgepath, P. (2014). Active Learning. https://www.slideshare.net/hedgepath/lets-flip-it-2014 Hrastinski, S. (2019). What Do We Mean by Blended Learning? TechTrends, 63(5), 564–569. Jansen, R. (2021, July 20). Zelfregulatie in blended onderwijs | Onderwijskennis. https://www.onderwijskennis.nl/kennisbank/zelfregulatie-blended-onderwijs Kat-De Jong, M. (2021). Reflectie op blended onderwijs. Verkenning, speelveld en implementatie van blended onderwijs binnen Avans Hogeschool. Breda: Avans Hogeschool. Last, B., & Jongen, S. (2023). Blended learning design: From theory to practice. Boom. Last, B., & Prinsen, F. (2021). Blended learning en onderwijsontwerp. https://www.onderwijskennis.nl/kennisbank/blended-learning-en-onderwijsontwerp Mül er, C., Mildenberger, T., & Steingruber, D. (2023). Learning effectiveness of a flexible learning study programme in a blended learning design: Why are some courses more effective than others? International Journal of Educational Technology in Higher Education, 20(1), 10. https://doi.org/10.1186/s41239-022-00379-x Rivera Muñoz, J., Berríos, H., & Arias-Gonzales, J. (2022). Systematic Review of Adaptive Learning Technology for Learning in Higher Education. Eurasian Journal of Educational Research (EJER), 98, 221–233. https://doi.org/10.14689/ejer.2022.98.014 SURF. (2022). De blended learning wave. https://communities.surf.nl/blended-learning/artikel/de-blended-learning-wave Van der Stappen, E. (2022). Keuzes maken in blended en flexibel onderwijs: Hoedan#. https://www.avans.nl/onderzoek/expertisecentra/stand-alone- lectoraten/lectoraten/digitale-didactiek/lectorale-rede Weber, N., Alexander, B., Ashford-Rowe, K., Barajas-Murphy, N., Dobbin, G., Knott, J., McCormack, M., Pomerantz, J., & Seilhamer, R. (2019). Educause Horizon report: 2019 Higher Education edition. EDUCAUSE. Yeung, K. L., Carpenter, S. K., & Corral, D. (2021). A Comprehensive Review of Educational Technology on Objective Learning Outcomes in Academic Contexts. Educational Psychology Review, 33(4), 1583–1630. https://doi.org/10.1007/s10648-020-09592-4 Yin, R. K. (2018). Case study research and applications: Design and methods (Sixth edition). SAGE. 280 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Appendix A - Further Explanation ALT The adaptive software is provided by a Dutch company: ABC (fictious name). ABC's software supports setting up the “blended wave” (see Figure A). ABC has developed several courses (modules) that correspond to topics within a curriculum. Each course contains several learning objectives (subjects/themes) that can be completed independently, but which have a logical structure. For example, the course in scope of our research – Introduction to Business Economics – consists of twelve learning objectives. These learning objectives have a suggested order (learning objective sequence 1-12 as captured in the trace data), however the software does not restrict a student to start with a specific learning objective in the tool. Hence, a student could start with learning objective #4 or even with the last one (#12). Figure A: The “blended wave” supported by the ALT This “blended wave” ensures – at least that is the philosophy behind this software – that the student individual y takes the first steps towards mastering a particular learning objective. The student takes these ‘first steps’ where and when he/she wishes (i.e. a-synchronously) by working digitally in the adaptive tool. The student practices with the learning materials per learning objective by reading texts, watching animations, and answering questions. With these 'first steps', the student eventually reaches Bloom's third level (“apply”) with respect to the learning objective (see Figure B) by him(her)self. Whether and how quickly the student reaches Bloom's third level (Bloom et al., 1956) is calculated by an algorithm that is included in ABC's software. R. Bakkers, M. Gremmen, E. van der Stappen: Adaptive Learning Technologies In Blended Learning Design: How Do Students and Teachers Use This Technology in Practice? 281 The outcome of the algorithm is influenced by the number of attempts the individual student needs to correctly answer questions as wel as the complexity of a given question. The teacher guides the student to (at least) the fourth Bloom level (“analysing”) during the synchronous learning activities (e.g. on campus lectures or workshops). To this end, various work formats are deployed, such as integration cases, discussion groups about current news items or group presentations. Figure B: Bloom's levels in relation to the “blended wave” and the deployment of the ALT For a student, ABC's software includes the following components: • Adaptive question module (arranged per learning objective); • Instant feedback on answers given in the question module; • Theory (also downloadable) related to a learning objective as wel as animations of and sample questions about that theory; • Insights into own level, answer attempts and activity; • Practical tests; • Asking questions to the teacher and giving feedback on the material in the module. 282 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY For a teacher, ABC's software includes the following components: • Creating and managing classes; • Gaining insight into level and progress of the (students in the) class through the teacher dashboard (see Figure C for a fictitious example); • Creating teacher materials, such as newspaper articles and self-developed in-depth questions or (integration) case histories; • Going through the lesson material in 'student mode'; • Creating, opening and analysing (practice) tests; • Responding to questions or feedback from students. Figure C: Snapshot (fictious) of the ALT's teacher dashboard The teacher dashboard includes the “ABC score”. This is a grade number (1-10) that shows a student’s (or class’s) progress per learning objective and across all learning objectives. The ABC score can be increased, according to the aforementioned algorithm, by answering questions in the module correctly (or decreased in case of an incorrect answer). In maximal y four steps, the student receives feedback on his wrong answer attempts to instantly help the student in the learning moment. The ABC score is also disclosed to a student. R. Bakkers, M. Gremmen, E. van der Stappen: Adaptive Learning Technologies In Blended Learning Design: How Do Students and Teachers Use This Technology in Practice? 283 Appendix B - Trace Data Explanation The original data file contains the following columns: • Student id (number; student’s full identity is known by ABC, not by the researchers); • Question id (number); • Learning objective1 id (number; xxx); • Learning objective sequence id (number; 1-12); • Type of question (e.g. multiple choice or table question); • Complexity of the specific question (number; 1-5); • Time stamp start of answering the specific question; • Time stamp end of answering the specific question; • Answer(s) provided by the student to the specific question (text; al answer attempts are captured in one cel in the original data file); • Question-answer id (number). Each row in the data file represents a specific student’s activity, i.e. answering a specific question in the question module. The student has a maximum of four (scaffolded) attempts to answer a question correctly. All answers attempts were logged in the same row. A number of adjustments were made to the original data file: • Split from one to four columns to capture the student’s answer attempts to the specific question (text); • Addition of the following columns: o Count (number; 1-4) of the answer attempts to the specific question; o Counter id (number) per row; o The day of the week (Monday - Friday) the specific student has a classroom lecture (synchronous learning activity); o Calendar week (number) of the student’s activity in the question module; 1 A learning objective represent a certain topic in the area of Business Economics, e.g. a company’s balance sheet. 284 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY o Teaching week (number) of the student’s activity in the question module; o Day of the week (Monday - Sunday) the student was active in the question module; o Time spent (in minutes) for answering the specific question; o ABC score (grade; 1-10) as a result of answering the specific question. The ABC score is presented per learning objective and overal ; o Result (grade; 1-10) of the written exam taken in the ninth week. In this course, 273 students were enrolled. Since our unit of analysis is the execution of a blended course design, it was decided to include in the scope of the data analysis only those students that were able to actual y undergo the blended wave of independent (online) preparation (a-synchronous) and classroom attendance (synchronous, usual y in a group of 28-32 students). We therefore filtered out students that were not assigned to a class group. TOWARD SOVEREIGN DATA EXCHANGE THROUGH A META-PLATFORM FOR DATA MARKETPLACES: A PRELIMINARY EVALUATION OF THE PERCEIVED EFFICACY OF CONTROL MECHANISMS ANTRAGAMA EWA ABBAS, WIRAWAN AGAHARI, HOSEA OFE, ANNEKE ZUIDERWIJK, MARK DE REUVER Delft University of Technology, Faculty of Technology, Policy and Management, Delft, Netherland a.e.abbas@tudelft.nl, w.agahari@tudelft.nl, ofehosea@gmail.com, a.m.g.zuiderwijkvaneijk@tudelft.nl, g.a.dereuver@tudelft.nl. The landscape of platform ecosystems is becoming increasingly complex, with new types of platforms emerging that glue together otherwise fragmented ecosystems. One recent case is meta- platforms that can contribute to the European Data Economy by interconnecting data marketplaces; however, meta-platforms may intensify data sovereignty concerns: the inability of data providers to own and control the exchanged data. While smart contracts and certification can general y enhance data sovereignty, it is unknown whether data providers perceive these control mechanisms as valuable in the complex meta-platform setting. This study aims to evaluate the perceived efficacy of the control mechanisms to ensure data sovereignty in meta-platforms. The findings from a survey study (n=93) indicate that respondents perceive high data sovereignty. One potential explanation is that smart contracts can Keywords: potential y enable providers to maintain ownership and control over data their exchanged data; meanwhile, certification may signal meta- economy, meta-platforms, platforms’ responsibility to deliver secure data exchange data infrastructure and assist providers in adhering to relevant marketplaces, regulations. This study contributes to advancing design knowledge business data for meta-platforms, showcasing that meta-platforms can be exchange, designed in a way to resolve fragmentation without neglecting data data sovereignty principles. sovereignty DOI https://doi.org/10.18690/um.fov.4.2023.18 ISBN 978-961-286-751-5 286 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The growing demand to unleash the ful potential of the Data Economy has led to the emergence of data marketplaces: multi-sided platforms that facilitate business data exchange among enterprises (Spiekermann, 2019). This phenomenon is particularly evident in the European context, where efforts to strengthen the European Data Economy have accelerated the proliferation of these marketplaces (European Commission, 2020). Due to the specialized nature of data as a commodity, data marketplaces often focus on specific industries, resulting in significant fragmentation and heterogeneity (Aaltonen et al., 2021). This fragmentation is expected to continue, causing lock-in effects and data discovery chal enges for data providers and consumers (Santiago & Laoutaris, 2022). The existing platform literature recognizes meta-platforms as potential measures to reduce fragmentation and achieve critical mass (Mosterd et al., 2021; Pitt et al., 2021). A meta-platform is built on top of two or more existing platforms to connect their ecosystems (Zhang & Williamson, 2021). A typical example of a meta-platform in the tourism industry is Trivago, which federates and coordinates other platforms (e.g., Expedia, Booking, and Airbnb). Other examples in the data marketplace context include the recently developed TRUSTS1 and i3-Market2. Nevertheless, while exchanging business data on a data marketplace is already difficult due to data sovereignty concerns (i.e., the inability of data providers to own and control the exchanged business data), these concerns wil likely intensify in a meta-platform setting because data may flow from one data marketplace to another (Zappa et al., 2022). Additional y, complying with data sovereignty principles has recently become a prerequisite for exchanging business data within the European context (European Commission, 2020). Smart contracts and certification are among the most-discussed control mechanisms to enhance data sovereignty in various data exchange settings (Lauf et al., 2022; Schmidt et al., 2022). For example, smart contracts have been extensively implemented in data marketplace cases to guarantee data sovereignty (Fruhwirth et al., 2020; Precht & Gómez, 2021). Likewise, certification has been implemented in 1 https://www.trusts-data.eu/, accessed on May 11, 2023 2 https://www.i3-market.eu/, accessed on May 11, 2023 A. E. Abbas, W. Agahari, H. Ofe, A. Zuiderwijk, M. de Reuver: Toward Sovereign Data Exchange Through a Meta-Platform for Data Marketplaces: A Preliminary Evaluation of the Perceived Ef icacy of 287 Control Mechanisms supply chains (Bastiaansen et al., 2020; Dalmolen et al., 2019) and data ecosystem settings (Azkan et al., 2020). Given the novel and intricate nature of meta-platforms (e.g., al owing data to flow from one data marketplace to another), it is unknown whether data providers view these control mechanisms as valuable. For example, data providers may argue that certification is less valuable due to the difficulty of tracing valid and non-expired certificates in the complex constel ations of data marketplaces. Against this backdrop, this study investigates the perceived efficacy of control mechanisms, namely smart contracts and certification, to enhance data sovereignty in data marketplace meta-platforms. This research advances design knowledge for meta-platforms to address fragmentation without compromising data sovereignty principles. 2 Research Background 2.1 Meta-platform offerings Multiple approaches can be adopted to interconnect digital platforms. Digital platforms can create direct application programming interfaces to connect with each other (Hodapp & Hanelt, 2022). Third parties can “bridge” two platforms by creating an application (e.g., the Philips Hue case) (Hilbolling et al., 2020) or “fork” a platform by exploiting its core resources (e.g., Amazon Fire) (Karhu et al., 2018). Meta-platforms, on the other hand, are built on top of two or more existing platforms, thereby connecting their respective ecosystems (Zhang & Williamson, 2021). Because meta-platforms may act as a coordinator to enable collective actions, legitimate governance, and transparency values (Pitt et al., 2021), meta-platforms are potential y suitable for tackling the fragmentation of data marketplaces in the European context. Trivago is an example of a meta-platform where Expedia, Booking, and Airbnb serve as platform participants. These platform participants have their provider- and consumer-side (we refer to them as end-users). Therefore, meta-platforms cannot exist independently (Lagutin et al., 2019). Meta-platforms have two key offerings to create value for end-users: information aggregation and a one-stop-shop portal. First, meta-platforms aggregate information from platform participants to create new services, for example, by creating a meta-search engine to manage information flow and 288 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY disseminate information (Lanza et al., 2016; Pitt et al., 2021). Aggregating information often aims to give recommendations to end-users (Floetgen et al., 2021; Yang & Wang, 2019). Second, meta-platforms provide a one-stop-shop portal to enable end-users to seamlessly interact with only a single user interface to perform necessary activities (Floetgen et al., 2021; Reinartz et al., 2019). This portal mediates interoperability between platform participants by providing service or technical integration (Ulrich & Alt, 2021). While participating platforms may be reluctant to standardization (Costabile et al., 2022), a meta-platform creates a common interface or protocol for interaction within its ecosystem without requiring participating platforms to modify their internal standards. 2.2 Data sovereignty dimensions and indicators Data sovereignty encompasses various dimensions. To maintain a focused analysis, we prioritize the dimensions most closely associated with data sovereignty: ownership, control, responsibility, and security (Hummel et al., 2021). We also investigate compliance as another dimension, given its recent legal prominence in contexts such as the European Data Governance Act (Duisberg, 2022). We define indicators for each dimension to provide observable measurements of data sovereignty as a multi- dimensional construct. Data ownership is the exclusive right and authority to make decisions regarding data assets (Hummel et al., 2021). Despite the ongoing debate on who should own data assets (e.g., an individual, an organization, or a platform) (Lee et al., 2017), we focus here on the organization as a unit of analysis because end-users of a meta-platform are organizations, not individuals. We define four indicators for data ownership: (1) data providers should be able to express the term of use of data exchange, (2) be involved in determining (monetary) incentives (Dalmolen et al., 2020), (3) define the data type to exchange (Lee et al., 2017), and (4) decide which data marketplace receives the meta-data description (Abbas et al., 2022). Control over exchanged data is among the most heavily recognized dimension of data sovereignty, referring to the ability of data providers to steer data exchange flows according to pre-defined agreements (Hummel et al., 2021). We define four indicators for data control. First, data providers can technical y enforce terms of use A. E. Abbas, W. Agahari, H. Ofe, A. Zuiderwijk, M. de Reuver: Toward Sovereign Data Exchange Through a Meta-Platform for Data Marketplaces: A Preliminary Evaluation of the Perceived Ef icacy of 289 Control Mechanisms of data exchange (Dalmolen et al., 2020). In doing so, data providers can track the data usage history (the second indicator). Third, data providers should be able to determine where they can store the shared (meta-) data (e.g., on the meta-platform, on its infrastructure, or the data consumer infrastructure) (Dalmolen et al., 2020). Final y, if something happens, data providers can withdraw their (meta-)data for a meta-platform and data marketplace participants (Lauf et al., 2022). Another critical data sovereignty dimension relates to compliance. As data exchange is subject to specific regulations, data providers should: (1) receive sufficient information to avoid violating laws and regulations, (2) obtain sufficient (technical) procedures to respond to those laws and regulations, and (3) utilize dispute mechanisms to handle conflicts (if any, with data consumers) (Hummel et al., 2021). One distinguishing dimension of data sovereignty due to the context novelty is the responsibility dimension, primarily because of the complex constel ations of data marketplaces via a meta-platform. As our previous study reveals (Abbas et al., 2022), it should be clear who is responsible for what to ensure sovereign data exchange. Hence, we propose the three indicators: meta-platforms should (1) responsibly select data marketplace participants that adhere to data exchange standards, (2) clearly divide responsibilities between the meta-platform and the data marketplace participants, and (3) take responsibility if the sensitive data is misused or stolen. Finally, we include an essential data sovereignty component: security. Building from the work of Hartono et al. (2014) and Hummel et al. (2021), we propose four indicators for security: (1) meta-platforms should prevent the disclosure of the exchanged data to unauthorized parties, (2) prevent the alteration of the exchanged data, (3) enable data providers to execute data-sharing transactions without system failures, and (4) implement up-to-date security features. In summary, we use these five data sovereignty dimensions to evaluate the perceived efficacy of control mechanisms in the meta-platform setting. 2.3 Control mechanisms: Smart contracts and certification Control theories explain how and why control mechanisms enacted by controllers can influence the behaviors of controlee (Saunders et al., 2020). Control mechanisms can be divided into formal (input, process, output) and informal (self and clan) control (Wiener et al., 2016). In this study, we want to examine how smart contracts 290 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY (as a process control) and certification (as an input control) can diminish data sovereignty concerns of data providers. We select the combination of these two mechanisms because they may influence al sovereignty dimensions. Smart contracts are “…any self-executing program running in the distributed ledger environment, and it is often meant to implement automated transactions agreed by the parties” (Governatori et al., 2018, p. 378). Regarding ownership, smart contracts offer a pre-fil ed template to define use cases of data exchange, which include the term of use, monetary incentives, and the data type to exchange (Moyano et al., 2021). Considering data control, smart contracts provide data provenance to enable transparency of data access and usage. At a more advanced level, smart contracts can automatical y monitor data compliance usage (Karger et al., 2021; Tuler De Oliveira et al., 2022). Furthermore, data providers can automatically revoke the license if consumers violate use and access rights (Jagals et al., 2021). We hypothesize that smart contracts enhance the data sovereignty dimensions of ownership and control (H1). Certification for sovereign data exchange “…defines a standardized level for security related to technical and organizational aspects” (Menz et al., 2019). Therefore, certification aims to confirm compliance with these pre-conditions (Biegel et al., 2020). Within the meta-platform context, applying certification means that a meta-platform decides pre-conditions that must be fulfil ed by data marketplace participants and their end-users (data providers and consumers). They need to, for instance, apply technical integration services such as application programming interfaces (or data ecosystem nodes) to join the meta-platform federation infrastructure. For example, compliance and security can be achieved by incorporating the International Data Space certification, demonstrating compliance with ISO/IEC 27001 (international standard for information security management) and IEC 62443 (cybersecurity for operational technology in automation and control systems). Related to the clarity of responsibility, certification can distinguish actors’ roles and responsibilities (Lansing et al., 2018). For example, a meta-platform and data marketplaces can only access the meta-data, not actual business data themselves; a meta-platform and an external third party wil act as evaluators to assess certification compliances. We hypothesize that certification can advance the A. E. Abbas, W. Agahari, H. Ofe, A. Zuiderwijk, M. de Reuver: Toward Sovereign Data Exchange Through a Meta-Platform for Data Marketplaces: A Preliminary Evaluation of the Perceived Ef icacy of 291 Control Mechanisms compliance, security, and responsibility dimensions of data sovereignty (H2). To sum up, Figure 1 summarizes the research model. Figure 1: The research model 3 Research approach This research is part of a larger design science project. We first explored state-of-the-art data marketplaces as federation objects and defined meta-platform boundary conditions. With a focus on data sovereignty concerns, we established metrics to evaluate artifact efficacy. We then examined control mechanisms in various business data exchange contexts, prioritizing smart contracts and certification. We developed a prototype to address data sovereignty concerns in collaboration with an EU project consortium. This paper concentrates explicitly on the final step, evaluating the perceived efficacy of control mechanisms in addressing data sovereignty concerns. We define efficacy as “the ability to produce a desired or intended result.” 3 While objective measures of data sovereignty (e.g., Firdausy et al., 2022), smart contracts (e.g., Hai & Liu, 2022), and certifications (e.g., Menz et al., 2019) technical y exist, they do not always reflect the subjective experience of data providers interacting with these control mechanisms. The level of control that smart contracts offer may not always match the perceived level of control due to various factors, including the complexity of the smart contracts or the ability to interpret how smart contracts work. Obtaining feedback on the perception of control mechanisms can help 3 http://www.oxforddictionaries.com/definition/english/efficacy, accessed on 06 April 2023. 292 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY identify gaps beyond the technical aspects, which can inform the design of control mechanisms. We conducted a survey study to achieve our objective, recruiting 93 participants residing in Europe through the Prolific platform (47 female, 46 male). The sample size was determined using G*Power statistical calculations. The majority of participants were young to middle-aged adults (31-45 years old, 51%), followed by young adults (17-30 years old, 40%) and older adults (9%). Educational backgrounds were diverse, with 46% holding a Master’s degree and 33% possessing a Bachelor’s degree. The target participant profile included employees with management experience and leadership responsibilities. A significant proportion of participants (82%) had planned or conducted business-sensitive data exchanges, and 75% self-reported being knowledgeable about data marketplaces. The online survey via Qualtrics consisted of three elements: a video explanation, a prototype, and a questionnaire. The video explained a hypothetical scenario where users play the role of a data provider, a telecommunication company so-cal ed TELCO. 4 Data providers will exchange their business data about Cal Detail Records via a meta-platform. Next, participants engaged with the prototype by completing a series of pre-defined tasks. 5 Task 1 consisted of simple sub-tasks designed to familiarize participants with the prototype. Task 2 involved describing meta-data associated with the platform, while Task 3 focused on creating and managing contracts. Lastly, Task 4 allowed participants to exercise the control capabilities of the meta-platform. After exercising the prototype, participants filled out a questionnaire. Because the measurement of data sovereignty does not yet exist, we mostly self-developed the indicators of each dimension, as elaborated in Section 2.2. For example, in the data control dimension, we asked the following question for the first indicator (DC_1): If I would share sensitive data, I believe the meta-platform of ers me technical means to enforce data usage policies. We also employed generic indicators such as (DS_G) “I believe the meta-platform enables sovereignty for the sensitive data that I would share.” These generic indicators were utilized as an enabler to check the convergent validity of the data sovereignty dimensions; for a detailed view of these indicators, please 4 The video can be accessed here: https://www.youtube.com/watch?v=9b7iKM3BiMs. 5 The prototype can be accessed here: https://www.figma.com/proto/KJUcfObwTZp8GaOrTyVhNi/TRUSTS-meta-platform?page-id=2506%3A47793&node-id=2506-50925&viewport=-444%2C-564%2C0.19&scaling=min-zoom&starting-point-node-id=2506%3A50925. A. E. Abbas, W. Agahari, H. Ofe, A. Zuiderwijk, M. de Reuver: Toward Sovereign Data Exchange Through a Meta-Platform for Data Marketplaces: A Preliminary Evaluation of the Perceived Ef icacy of 293 Control Mechanisms refer to the online supplementary material. 6 Participants answered the questions on a 5-point Likert scale. Because we have five dimensions contributing to data sovereignty, our model can be seen as a Hierarchical Component Model (HCM). HCM offers several benefits, for example, minimizing the quantity of path model connections, overcoming the bandwidth-fidelity dilemma, and decreasing collinearity among dimensions (Sarstedt et al., 2019). We employed a standard approach to validate the measurement of the HCM in SmartPLS 4: a joint two-stage approach (Ringle et al., 2012). In the first stage, we evaluated all indicators regarding indicator reliability, internal consistency reliability, convergent validity, and discriminant validity. In the second stage, we formed a latent composite score of each dimension and evaluated their convergent validity, collinearity issues, and relevance (Hair et al., 2021). Following this, we conducted a one-sample t-test in SPSS to assess the extent to which the perceived efficacy of control mechanisms, as reported by participants, is significantly better than the midpoint of our Likert scale. 4 Results The reliability of each indicator is confirmed, as the outer loading (λ) for all indicators is within the range of 0.6 and 0.9 (Hair et al., 2021). The internal consistency reliability for each aspect is also established, as indicated by the composite reliability (rho_a) score for each greater than 0.7. Convergent validity is likewise confirmed, as the Average Variance Extracted for al aspects surpasses 0.5. Consequently, we opted against removing any indicators. As for discriminant validity, the Heterotraitmonotrait ratio (HTMT) for al dimensions is below the recommended threshold of 0.9, except for Security (S) and Responsibility (R). Thus, we examine cross-loadings and remove one item (S_4), establishing discriminant validity. Our final model comprises five data sovereignty dimensions. Specifical y, the dimension of data ownership is represented through four indicators (DO_1, DO_2, DO_3, DO_4), data control through four indicators (DC_1, DC_2, DC_3, DC_4), compliance through four indicators (C_1, C_2, C_3, C_4), responsibility through three indicators (R_1, R_2, R_3), and finally, security through three indicators (S_1, S_2, 6 The online supplementary material can be accessed here: https://doi.org/10.4121/e4cacfac-31f0-4523-81f4-35383ba958a8. 294 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY S_3). We also confirm the validity of a generic data sovereignty construct measured by six indicators. Next, we calculated the Latent Variable (LV) score for each data sovereignty dimension from the Hierarchical Component Model (HCM) of data sovereignty (Figure 2). The convergent validity is established as (β = 0.713, p = 0.00) and R2 > 0.5, indicating data these dimensions well represent data sovereignty. The HCM exhibits no collinearity issue, as all dimensions have a Variance Inflation Factor (VIF) less than 5. Although Outer Weight (OW) testing shows significance only for the responsibility (OW = 0.38, p = 0.01) and security dimensions (OW = 0.38, p = 0.00), we retain the other dimensions since their outer loadings are greater than 0.5, as suggested by Hair et al. (2021). Figure 2: Hierarchical component model of data sovereignty After validating the HCM measurement model, we conducted a one-sample t-test to compare the mean score of al data sovereignty dimensions against the benchmark value of three. Respondents perceive high data sovereignty when faced with our control mechanisms, as the mean of ownership [t(92) = 16.1, p < 0.01], control [t(92) = 16.48, p < 0.01], compliance [t(92) = 12.41, p < 0.01], security [t(92) = 9.89, p < 0.01], and responsibility [t(92) = 9.06, p < 0.01] are all significantly greater than the benchmark value of three (see Table 1). Detailed elaboration, including raw data, survey indicators, and the complete analytical statistic, is available in the online supplementary material. A. E. Abbas, W. Agahari, H. Ofe, A. Zuiderwijk, M. de Reuver: Toward Sovereign Data Exchange Through a Meta-Platform for Data Marketplaces: A Preliminary Evaluation of the Perceived Ef icacy of 295 Control Mechanisms Table 1: One-sample t-test calculation (n = 93, test value = 3) Descriptive statistic One-sample t-test Dimension Mean SD Mean dif. t value p Ownership 4.07 0.64 1.07 16.10 <0.01 Control 4.15 0.67 1.15 16.48 <0.01 Compliance 4.04 0.80 1.04 12.41 <0.01 Security 3.78 0.76 0.78 9.89 <0.01 Responsibility 3.74 0.79 0.74 9.06 <0.01 5 Discussions and conclusions This research aims to evaluate the perceived efficacy of control mechanisms, namely smart contracts and certification, in enhancing data sovereignty within the context of a meta-platform for data marketplaces. Our findings indicate that the meta-platform prototype containing the control mechanisms is evaluated positively on al data sovereignty dimensions. One possible explanation can be reflected in the proposed hypotheses: smart contracts may play a role in impacting data ownership and control dimensions (H1); meanwhile, certification may influence compliance, security, and responsibility dimensions (H2). We primarily contribute to two streams of literature: 1) multiplatform constellations (or platform ecologies) and 2) data exchange. Specifical y, we do so by advancing design knowledge for meta-platforms by providing an initial assessment of the perceived efficacy of control mechanisms in addressing data sovereignty concerns. We are among the first that showcase data sovereignty is evaluated positively in meta-platforms, meaning that, even though meta-platforms exponential y amplify the risks of data exchange, they may stil be designed in ways that do not harm data sovereignty. Considering design knowledge in the problem space, measuring data sovereignty efficacy remains ambiguous and complex. Existing research often equates data sovereignty with control (e.g., Jarke et al., 2019; Otto & Jarke, 2019) while overlooking other dimensions (cf. Hummel et al., 2021). Our findings resolve this tension by offering an alternative approach that captures the multifaceted nature of 296 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY data sovereignty. To do so, we advance Hummel et al.’s (2021) work in three key aspects: first, by incorporating an added dimension of responsibility due to the unique context of meta-platforms with increased governance complexity among data marketplace participants; second, by offering empirical evidence on the col ective influence of the five dimensions (ownership, control, security, compliance, and responsibility) on data sovereignty; and third, by enhancing granularity through the introduction of data sovereignty measurement models employed as survey instruments. Regarding design knowledge in the solution space, our study offers valuable insights into how data providers perceive control mechanisms as valuable for ensuring data sovereignty within the unique context of meta-platforms. Our findings align with existing literature on data exchange, which suggests that smart contracts can technically enable ownership and control (e.g., Saini et al., 2020; Zhang et al., 2018), and certifications can enhance compliance, security, and responsibility (Lansing et al., 2018). However, our contribution extends beyond this general understanding by opening up future discussions for their applicability in distinctive meta-platform characteristics. For instance, smart contracts in meta-platforms are distinct from those in supply chains due to the requirement for seamless interoperability among multiple interconnected marketplaces. This interoperability demands the development of smart contracts that automatical y enforce data usage policies and agreements between individual marketplaces and across varied legal and regulatory environments. As a result, smart contracts that leverage interoperable “side chains” emerge as a potential solution to explore (Singh et al., 2020). Alternatively, meta-platforms can offer shared services for data marketplace participants, serving as a backbone infrastructure to facilitate smart contract deployment and alleviate interoperability chal enges. This study has some limitations. First, this study employed a one-sample t-test, thus constraining our comparison to the Likert scale’s mid-point. To further improve the validity of the finding, we will continue this research by conducting a between-subject 2x2 factorial experiment. In doing so, we can compare the effect of the presence of these control mechanisms and identify potential interaction effects to confirm the proposed H1 and H2. Second, while our study focuses on the most critical dimension of data sovereignty, we are aware of the potential significance of A. E. Abbas, W. Agahari, H. Ofe, A. Zuiderwijk, M. de Reuver: Toward Sovereign Data Exchange Through a Meta-Platform for Data Marketplaces: A Preliminary Evaluation of the Perceived Ef icacy of 297 Control Mechanisms other dimensions (e.g., justice). To account for this, we considered the justice dimension as a control variable in the prototype development by suggesting appropriate data pricing to ensure fair revenue distributions. Finally, the technical aspects of smart contracts and certification are beyond the scope of our work. To what extent these two control mechanisms can be implemented in a large-scale setting needs further research. This paper has important policy implications as it suggests ways forward to a single European Data Market while allowing specialized data marketplaces (or data spaces) to exist. This resolves tensions in the European policy to promote a single market for data and interoperable data sharing (e.g., in EU Data strategy, Data Act) and promote verticals/sector-specific data platforms (e.g., the eight verticals in the Digital Europe program), while at the same time, adhere to data sovereignty principles. Acknowledgments The research leading to these results has received funding from the European Union’s Horizon 2020 Research and Innovation Programme, under Grant Agreement no 871481–Trusted Secure Data Sharing Space (TRUSTS) and No 825225–Safe Data-Enabled Economic Development (Safe-DEED). References Aaltonen, A., Alaimo, C., & Kallinikos, J. (2021). The Making of Data Commodities: Data Analytics as an Embedded Process. Journal of management information systems, 38(2), 401-429. https://doi.org/10.1080/07421222.2021.1912928 Abbas, A. E., Ofe, H., Zuiderwijk, A., & de Reuver, M. (2022). Preparing Future Business Data Sharing via a Meta-Platform for Data Marketplaces: Exploring Antecedents and Consequences of Data Sovereignty. 35th Bled eConference - Digital Restructuring and Human (Re-Action), Bled, Slovenia. Azkan, C., Möl er, F., Meisel, L., & Otto, B. (2020). Service dominant Logic Perspective on Data Ecosystems-a Case Study based Morphology. ECIS 2020 Research Papers, Marrakesh, Morocco. Bastiaansen, H., Dalmolen, S., Kol enstart, M., & van Engers, T. M. (2020). User-Centric Network-Model for Data Control with Interoperable Legal Data Sharing Artefacts. PACIS 2020 Proceedings, Dubai, the United Arab Emirates. Biegel, F., Bongers, A., Chidambaram, R., Feld, T., Garloff, K., & Ingenrieth, F. (2020). GAIA-X: Driver of digital innovation in Europe. Germany’s Federal Ministry for Economic Affairs and Energy (BMWi). 298 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Costabile, C., Iden, J., & Bygstad, B. (2022). Building digital platform ecosystems through standardization: an institutional work approach. Electronic Markets, 32(4), 1877-1889. https://doi.org/10.1007/s12525-022-00552-0 Dalmolen, S., Bastiaansen, H., Kol enstart, M., & Punter, M. (2020). Infrastructural sovereignty over agreement and transaction data (‘metadata’) in an open network-model for multilateral sharing of sensitive data. 40th International Conference on Information Systems, ICIS 2019, Munich, Germany. Dalmolen, S., Bastiaansen, H., Somers, E., Djafari, S., Kol enstart, M., & Punter, M. (2019). Maintaining control over sensitive data in the Physical Internet: Towards an open, service oriented, network-model for infrastructural data sovereignty. 6th International Physical Internet Conference (IPIC), London, the United Kingdom. Duisberg, A. (2022). Legal Aspects of IDS: Data Sovereignty—What Does It Imply? Designing Data Spaces, 61. European Commission. (2020). A European Strategy for Data. https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52020DC0066&from=EN Firdausy, D. R., De Alencar Silva, P., Van Sinderen, M., & Iacob, M.-E. (2022). Towards a Reference Enterprise Architecture to enforce Digital Sovereignty in International Data Spaces. 2022 IEEE 24th Conference on Business Informatics (CBI), Amsterdam, the Netherlands. Floetgen, R. J., Strauss, J., Weking, J., Hein, A., Urmetzer, F., Böhm, M., & Krcmar, H. (2021). Introducing platform ecosystem resilience: leveraging mobility platforms and their ecosystems for the new normal during COVID-19. European Journal of Information Systems, 1-18. https://doi.org/10.1080/0960085x.2021.1884009 Fruhwirth, M., Rachinger, M., & Prlja, E. (2020). Discovering Business Models of Data Marketplaces. Proceedings of the 53rd Hawai International Conference on System Sciences, Hawai , the United States. Governatori, G., Idelberger, F., Milosevic, Z., Riveret, R., Sartor, G., & Xu, X. (2018). On legal contracts, imperative and declarative smart contracts, and blockchain systems. Artificial Intelligence and Law, 26(4), 377-409. https://doi.org/10.1007/s10506-018-9223-3 Hai, X., & Liu, J. (2022). PPDS: Privacy Preserving Data Sharing for AI applications Based on Smart Contracts. 2022 IEEE 46th Annual Computers, Software, and Applications Conference (COMPSAC), Los Alamitos, the United States. Hair, J. F., Hult, G. T. M., Ringle, C. M., & Sarstedt, M. (2021). A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM). SAGE Publications. Hartono, E., Holsapple, C. W., Kim, K.-Y., Na, K.-S., & Simpson, J. T. (2014). Measuring perceived security in B2C electronic commerce website usage: A respecification and validation. Decision Support Systems, 62, 11-21. Hilbol ing, S., Berends, H., Deken, F., & Tuertscher, P. (2020). Complementors as connectors: managing open innovation around digital product platforms. R&D Management, 50(1), 18-30. https://doi.org/10.1111/radm.12371 Hodapp, D., & Hanelt, A. (2022). Interoperability in the era of digital innovation: An information systems research agenda. Journal of Information Technology, 0(0), 1-21. https://doi.org/10.1177/02683962211064304 Hummel, P., Braun, M., Tretter, M., & Dabrock, P. (2021). Data sovereignty: A review. Big Data & Society, 8(1), 1–17. https://doi.org/10.1177/2053951720982012 Jagals, M., Karger, E., Ahlemann, F., & Brée, T. (2021). Enhancing Inter-Organizational Data Governance via Blockchain–Shaping Scopes and Research Avenues. Forty-Second International Conference on Information Systems, Texas, the United States. Jarke, M., Otto, B., & Ram, S. (2019). Data sovereignty and data space ecosystems. Business & Information Systems Engineering, 61(5), 549-550. Karger, E., Jagals, M., & Ahlemann, F. (2021). Blockchain for AI Data–State of the Art and Open Research. Forty-Second International Conference on Information Systems, Texas, the United States. A. E. Abbas, W. Agahari, H. Ofe, A. Zuiderwijk, M. de Reuver: Toward Sovereign Data Exchange Through a Meta-Platform for Data Marketplaces: A Preliminary Evaluation of the Perceived Ef icacy of 299 Control Mechanisms Karhu, K., Gustafsson, R., & Lyytinen, K. (2018). Exploiting and defending open digital platforms with boundary resources: Android’s five platform forks. Information Systems Research, 29(2), 479-497. Lagutin, D., Bel esini, F., Bragatto, T., Cavadenti, A., Croce, V., Kortesniemi, Y., Leligou, H. C., Oikonomidis, Y., Polyzos, G. C., Raveduto, G., Santori, F., Trakadas, P., & Verber, M. (2019). Secure Open Federation of IoT Platforms Through Interledger Technologies - The SOFIE Approach. 2019 European Conference on Networks and Communications (EuCNC), Valencia, Spain. Lansing, J., Benlian, A., & Sunyaev, A. (2018). “Unblackboxing” Decision Makers’ Interpretations of IS Certifications in the Context of Cloud Service Certifications. Journal of the Association for Information Systems, 19(11), 3. Lanza, J., Sanchez, L., Gomez, D., Elsaleh, T., Steinke, R., & Cirillo, F. (2016). A Proof-of-Concept for Semantically Interoperable Federation of IoT Experimentation Facilities. Sensors, 16(7), 1006. https://doi.org/10.3390/s16071006 Lauf, F., Scheider, S., Bartsch, J., Herrmann, P., Radic, M., Rebbert, M., Nemat, A. T., Schlueter Langdon, C., Konrad, R., & Sunyaev, A. (2022). Linking Data Sovereignty and Data Economy: Arising Areas of Tension. Wirtschaftsinformatik 2022 Proceedings, Nuremberg, Germany. Lee, S. U., Zhu, L., & Jeffery, R. (2017). Data governance for platform ecosystems: Critical factors and the state of practice. PACIS 2017 Proceedings, Langkawi Island, Malaysia. Menz, N., Resetko, A., & Winkel, J. (2019). IDS Certification explained. Mosterd, L., Sobota, V. C. M., Van De Kaa, G., Ding, A. Y., & De Reuver, M. (2021). Context dependent trade-offs around platform-to-platform openness: The case of the Internet of Things. Technovation, 108, 102331. https://doi.org/10.1016/j.technovation.2021.102331 Moyano, J. P., Avital, M., Bühler, M., & Schmedders, K. (2021). Fostering Peer-to-Peer Blockchain-based Data Markets. The 25th Pacific Asia Conference on Information Systems, Dubai, the United Arab Emirates. Otto, B., & Jarke, M. (2019). Designing a multi-sided data platform: findings from the International Data Spaces case. Electronic Markets, 29(4), 561-580. https://doi.org/10.1007/s12525-019-00362-x Pitt, S., van Meelis Lacey, M., Scaife, E., & Pitt, J. (2021). No App is an Island: Collective Action and Sustainable Development Goal-Sensitive Design. International Journal of Interactive Multimedia & Artificial Intelligence, 6(5). Precht, H., & Gómez, J. M. (2021). Towards GDPR Enforcing Blockchain Systems. Wirtschaftsinformatik 2021 Proceedings, Essen, Germany. Reinartz, W., Wiegand, N., & Imschloss, M. (2019). The impact of digital transformation on the retailing value chain. International Journal of Research in Marketing, 36(3), 350-366. Ringle, C. M., Sarstedt, M., & Straub, D. W. (2012). Editor’s comments: a critical look at the use of PLS-SEM in” MIS Quarterly”. MIS quarterly, ii -xiv. Saini, A., Zhu, Q., Singh, N., Xiang, Y., Gao, L., & Zhang, Y. (2020). A smart-contract-based access control framework for cloud smart healthcare system. IEEE Internet of Things Journal, 8(7), 5914-5925. Santiago, & Laoutaris, N. (2022). A Survey of Data Marketplaces and Their Business Models. SIGMOD Record, 51(3), 18-29. Sarstedt, M., Hair Jr, J. F., Cheah, J.-H., Becker, J.-M., & Ringle, C. M. (2019). How to specify, estimate, and validate higher-order constructs in PLS-SEM. Australasian marketing journal, 27(3), 197-211. Saunders, C., Benlian, A., Henfridsson, O., & Wiener, M. (2020). MIS Quarterly Research Curation: IS Control & Governance. MIS quarterly. Schmidt, K., Munil a Garrido, G., Mühle, A., & Meinel, C. (2022). Mitigating Sovereign Data Exchange Chal enges: A Mapping to Apply Privacy-and Authenticity-Enhancing Technologies. International Conference on Trust and Privacy in Digital Business, Vienna, Austria. 300 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Singh, A., Click, K., Parizi, R. M., Zhang, Q., Dehghantanha, A., & Choo, K.-K. R. (2020). Sidechain technologies in blockchain networks: An examination and state-of-the-art review. Journal of Network and Computer Applications, 149, 102471. Spiekermann, M. (2019). Data Marketplaces: Trends and Monetisation of Data Goods. Intereconomics, 54(4), 208-216. https://doi.org/10.1007/s10272-019-0826-z Tuler De Oliveira, M., Reis, L. H. A., Verginadis, Y., Mattos, D. M. F., & Olabarriaga, S. D. (2022). SmartAccess: Attribute-Based Access Control System for Medical Records Based on Smart Contracts. IEEE Access, 10, 117836-117854. https://doi.org/10.1109/access.2022.3217201 Ulrich, D., & Alt, R. (2021). Social networking platforms to close the gender gap: an analysis of female doctoral students in information systems. ECIS 2021 Research Papers, Timișoara, Romania. Wiener, M., Mähring, M., Remus, U., & Saunders, C. (2016). Control configuration and control enactment in information systems projects: Review and expanded theoretical framework. MIS quarterly, 40(3), 741-774. Yang, F., & Wang, Y. (2019). Research Status and Trend Analysis of Learning Resource Aggregation in the Era of Big Data. Moscow, Russia. Zappa, A., Le, C.-H., Serrano, M., & Curry, E. (2022). Connecting Data Spaces and Data Marketplaces and the Progress Toward the European Single Digital Market with Open-Source Software. In (pp. 131-146). Springer International Publishing. https://doi.org/10.1007/978-3-030-98636-0_7 Zhang, M. Y., & Wil iamson, P. (2021). The emergence of multiplatform ecosystems: insights from China’s mobile payments system in overcoming bottlenecks to reach the mass market. Technological Forecasting and Social Change, 173, 121128. https://doi.org/10.1016/j.techfore.2021.121128 Zhang, Y., Kasahara, S., Shen, Y., Jiang, X., & Wan, J. (2018). Smart contract-based access control for the internet of things. IEEE Internet of Things Journal, 6(2), 1594-1605. AN ETHICAL PERSPECTIVE ON LOOT BOX PURCHASING - EXAMINING PSYCHOSOCIAL ANTECEDENTS AND THE ASSOCIATION WITH INDEBTEDNESS JUSSI NYRHINEN,1 ANU SIROLA,1 LAURI FRANK,2 JULIA NUCKOLS,1 TERHI-ANNA WILSKA1 1 University of Jyväskylä, Faculty of Humanities and Social Sciences, Jyväskylä, Finland jussi.nyrinen@jyu.fi ,anu.r.s.sirola@jyu.fi, julia.a.nuckols@jyu.fi, terhi-anna.wilska@jyu.fi 2 University of Jyväskylä, Faculty of Information Technology, Jyväskylä, Finland lauri.frank@jyu.fi Loot boxes are popular random reward mechanisms in digital games, attracting players to invest real money to enhance their gaming experiences. Loot boxes share striking similarities to gambling and might contribute to one’s economic strain, but more research is needed on the underlying vulnerabilities and motivational traits in loot box purchasing. This paper examines associations with self-reported increase in loot box purchasing and debt problems during the first year of the COVID-19 pandemic. International survey data were col ected in 2021, Keywords: loot consisting of Finnish, Swedish, and British respondents (N = box, 2,991) aged 18 to 75. Partial least squares modeling was used as indebtedness, an analytical technique. The findings bring valuable insight into resilience, loneliness, the underlying psychosocial and motivational factors in loot box social purchasing and its association with indebtedness. relationships DOI https://doi.org/10.18690/um.fov.4.2023.19 ISBN 978-961-286-751-5 302 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Loot boxes are an example of gambling-gaming convergence, referring to random- reward in-game purchase opportunities present in modern video games (Drummond & Sauer, 2018; Spicer et al., 2022; Zendle et al. 2019). Loot boxes can be defined as chance-based “mystery boxes” containing a selection of game-related items (e.g., weapons, cosmetic features), and they are typical y purchasable with real money (Zendle et al., 2020). The chance-based mechanism of loot boxes is often juxtaposed with gambling (Delfabbro & King, 2020; Spicer et al., 2022), and loot box opening even provokes similar physiological and psychological reactions as gambling does (Brady & Prentice, 2021). Since gambling is highly addictive, the gambling-like nature of loot boxes make them an addictive characteristic of video games (Király et al., 2023). From an ethical perspective, this raises concerns particularly regarding vulnerable players such as young people or individuals with psychosocial or financial problems. Even though loot boxes have gained scholars’ attention particularly in relation to disordered gambling, there is a cal for research to explore underlying factors and individual characteristics which might contribute to excessive loot box spending and consequent problems (Yokomitsu et al., 2021). There is some evidence that loot box expenditure is associated with financial harm (Carey et al., 2022), but more detailed studies are needed. Additional y, adolescent players with psychosocial problems such as depression are more susceptible to unplanned loot box purchasing (Irie et al., 2022), but more research is needed to gain better insight on psychosocial risks and protective factors. Most studies on loot box purchasing have focused on adolescent players (e.g., Hing et al., 2022; Irie et al., 2022; Kristiansen & Severin, 2020), which is important given the popularity of loot boxes in games that children play (Zendle et al., 2020). However, research on adult gamers is needed as well. Frank, Salo, & Toivakka (2015) conclude that digital gamers’ purchasing decisions are dominantly based on hedonic motivations, but loot box purchasing might have unique motivational traits due to their gambling-like nature. Despite the similarities with gambling, ethical issues regarding personal, social, and financial vulnerabilities in loot box purchasing, have to our knowledge, not been researched before. J. Nyrhinen, A. Sirola, L. Frank, J. Nuckols, T.-Anna Wilska: An Ethical Perspective on Loot Box Purchasing - Examining Psychosocial Antecedents and the Association with Indebtedness 303 Additionaly, the impact of the COVID-19 pandemic on problematic gambling and gaming behaviors is still topical. Due to shutdowns of societies and excessive social isolation, opportunities for daily recreational activities and enjoyment were highly limited particularly during the first year of the pandemic, making the role of online technologies more emphasized. In the early onset of the pandemic, the World Health Organization recommended digital gaming as a safe activity to spend time and connect with friends and family (King et al., 2020). Some individuals have been more vulnerable to develop harmful habits over the course of the pandemic. For example, problematic online behaviors such as excessive engagement in online gaming or gambling have acted as a response to a crisis and related mental distress for some individuals (Savolainen et al., 2022). Developing harmful habits such as increased purchase behaviors as a response to the pandemic might have long-lasting effects on one’s financial wel being such as the tendency for indebtedness. This paper investigates the (self-reported) increase in loot box purchasing and its association with indebtedness during the first year of the COVID-19 pandemic. We approach increased loot box purchasing as a form of problematic behavior because of its potential to harm one’s finances (Carey et al., 2022) and its striking resemblance with gambling mechanisms (Delfabbro & King, 2020; Király et al., 2023; Spicer et al., 2022). The aim of this paper is to first examine if loneliness, social support, and COVID-19 worry are drivers for loot box purchasing and if psychological resilience protects consumers from excessive loot box purchasing. Second, we study if loot box purchasing predicts tendencies for indebtedness. 2 Theoretical Framework Because the chance-based mechanism of loot boxes holds an analogy for gambling (Delfabbro & King, 2020, Király et al., 2023; Spicer et al., 2022), we build our framework on the work of Sirola et al. (2023), who examined the psychosocial drivers on gambling during the COVID-19 pandemic. Sirola et al. (2023) discovered an indirect association between COVID-19 worry and problem gambling through loneliness (H1). Therefore, in a similar manner, we investigate whether there is direct or indirect link between COVID-19 worry and loot box purchasing (H2). In line with prior studies (Sirola et al., 2023), we suggest that social support is negatively associated with loneliness (H3). However, according to a review by Nordmyr & Forsman (2020) there is contradictory evidence on the role of social support in 304 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY problem gambling. For example, some studies indicate that social support and socializing with peers might be positively associated with problem gambling, and social support might be particularly problematic if it promotes gambling activities (Dowling et al., 2017; Räsänen et al., 2016; Sirola et al., 2023; Yücel et al., 2015). Additionally, social identification and peer-pressure in team-based digital gaming might influence in-game purchase behaviors (Sirola et al., 2021). This makes it worth testing whether social support has a positive association with loot box purchasing (H4). Prior studies have shown social support to have positive associations with psychological resilience (H5) (Cohen & Wills, 1985; Mancini & Bonanno, 2009; Zhang et al., 2022). Existing research has also found positive associations between loneliness and higher problem gambling severity (Khazaal et al., 2017; Sirola et al., 2019; 2023; Vuorinen et al., 2021) and therefore we expect loneliness to have similar association with loot box purchasing (H6). Lussier et al. (2007) suggest that psychological resilience is a protective factor to gambling particularly among youth, while more recent studies have provided contradictory evidence on the role of resilience among adult gamblers (Mishra et al., 2019; Oei & Goh, 2015; Scholes-Balog et al., 2015; Sirola et al., 2019). Thus, it is also worth testing whether psychological resilience is negatively associated with loot box purchasing (H7). We also expand the framework by Sirola et al. (2023) by examining the consequences of loot box purchasing. Because harmful habits or compulsive behaviors such as excessive purchase behaviors and overspending might have long-lasting effects on one’s financial wellbeing (Achtziger, 2022), we postulate that loot box purchasing is positively associated with the tendency for indebtedness (H8). We control ed the research model for the effects of a respondent’s age, gender, and level of education. The research hypotheses are listed in Table 1 and the summary of our theoretical framework is illustrated in Figure 1. J. Nyrhinen, A. Sirola, L. Frank, J. Nuckols, T.-Anna Wilska: An Ethical Perspective on Loot Box Purchasing - Examining Psychosocial Antecedents and the Association with Indebtedness 305 Table 1: List of Research Hypotheses Hypotheses H1 Loneliness is positively associated with COVID-19 worry. H2 COVID-19 worry is positively associated with increased loot box purchasing. H3 Social support is negatively associated with perceived loneliness. H4 Social support is positively associated with increased loot box purchasing. H5 Social support is positively associated with psychological resilience. H6 Loneliness is positively associated with increased loot box purchasing. H7 Psychological resilience is negatively associated with increased loot box purchasing. H8 Loot box purchasing is positively associated with indebtedness. H1+ Loneliness H3- COVID-19 worry H6+ H2+ Loot box H8+ Indebtedness purchasing H4+ Social support H7- Controls: H5+ Psychological Age resilience Gender Education Figure 1: Theoretical framework and hypotheses 3 Methods and Materials 3.1 Sample and Data Col ection The participants in this study's cross-national dataset (N = 2,991) ranged in age from 18 to 75 and originated from Finland (n = 995; mean age = 44.60; 50.4% female; SD = 15.67); Sweden (n = 998; mean age = 43.84; SD = 15.78; 50.3% female); and the UK (n = 998; mean age = 43.56; SD = 15.76; 50.8% female). During the COVID-19 pandemic in April 2021, information was simultaneously collected from 306 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY these three nations utilizing an anonymous online survey. The questionnaire was created to examine how the pandemic has affected people's general lives, including their online activity, consumption, and wellbeing. We selected Finland, Sweden, and the UK for our study, because these are three technological y advanced and cultural y relatively similar European nations. In Finnish, Swedish, and English, the survey's format and metrics were all similar. A data-provider company recruited respondents from an internet panel (CINT) using a random sample in each nation. Contacts with the panelists were made in a random order. The web panelists are volunteers who select to participate in surveys based on their preferences and areas of interest. Also, the panelists receive rewards and pay in exchange for their time and labor as provided by the research company. 3.2 Measures The measurement scales comprised of 15 items that involved 4 latent constructs and 2 single items. The scales used to measure loneliness, psychological resilience, and social support were drawn from prior literature. A three-item loneliness scale that was created for survey research and is a condensed but comparable version of the complete UCLA loneliness scale (Hughes et al., 2004) was used to measure loneliness during the pandemic. A condensed version of the Connor-Davidson resilience scale (CD-RISC) (Campbel -Sills & Stein, 2007; Connor & Davidson, 2003), which evaluates psychological resources to deal with chal enging circumstances like crises, was used to measure psychological resilience. Brief Form of the Perceived Social Support Questionnaire (F-SozU K-6) was used to measure social support (Kliem et al., 2015; Lin et al., 2019). COVID-19 worry was measured using a scale that was modified from previous studies that assessed pandemic-related worry (e.g., Broos et al., 2022; Mónaco et al., 2022; Okruszek et al., 2020; Ranta et al., 2020). ‘Loot box purchasing’ (Mean = 2.03, Standard Deviation = 1.13) and indebtedness (Mean = 1.17, Standard Deviation = 1.11) were measured using single items. ‘Loot box purchasing’ was measured with the following item “How have your online consumer habits changed during the coronavirus pandemic regarding the following services in comparison to your previous habits: Loot box purchases in digital games”, with a 5-point response scale (1 = I have not purchased at al 2 = I have purchased less 3 = the same amount 4 = more to some degree 5 = considerably J. Nyrhinen, A. Sirola, L. Frank, J. Nuckols, T.-Anna Wilska: An Ethical Perspective on Loot Box Purchasing - Examining Psychosocial Antecedents and the Association with Indebtedness 307 more)“. The measure for indebtedness was adopted from Wang and Xiao (2009) and it used a 5-point semantic differential scale: “Which of the following statements best describe your indebtedness? Choose the option which best suits your situation: 1 = The payment of bil s, payments, and/or payment instalments is not troublesome, and I am able to save money in the process; 2 = The payment of bil s, payments, and/or payment instalments is not troublesome, however I am unable to save money in the process; 3 = The payment of bil s, payments, and/or payment instalments is continuously chal enging for me; 4 = I have received payment notices and have been subject to paying tardiness interest, as I have not had sufficient funds to complete payments when the needed payments have been due; and 5 = I have a compromised credit score/have been subject to reposition actions.” 3.3 Nonresponse Bias and Common Method Bias Nonresponse bias was assessed by comparing the national samples to the structure of the populations aged 18–75 years in Finland, Sweden, and UK. Except for a few categories that are less than 8% over- or underrepresented in the data, the distributions of age, gender, and region are general y within the margin of error when compared to the overal population (see Appendix 1). Therefore, nonresponse bias is unlikely an issue in our data set. Finland had a response rate of 26%, the UK had a response rate of 73%, and Sweden had a response rate of 27%. The standard deviation is 3.1% across all nations. Less than 3.7% of the values for each item in this study were missing. The fol owing steps were taken to reduce common method bias (CMB) (see Podsakoff et al., 2012). In order to prevent questions measuring the same dimension from being subsequent, the order of the questionnaire items was first mixed. Second, we alternated the scales' answer formats. These two methods were used to lessen the possibility that respondents would rely their following responses on their previous responses. Thirdly, we reduced the ambiguity in the phrases used to describe the items. This method was chosen because, if the questions are not straightforward and clear, respondents may find it difficult to comprehend the meaning of confusing statements and instead rely on systematic answer patterns. The Harman single factor a common latent factor (CLF) test (Harman, 1976; Podsakoff et al., 2003) was run to reveal any potential CMB interference with the results. Because the total variance extracted by one component was 27.34% and was below the 50% criterion in our 308 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY data set, the findings of Harman's single factor CLF test indicate that CMB was likely not present. 3.4 Analysis Strategy For the following reasons, we used partial least squares structural equation modeling (PLS-SEM) with SmartPLS 3.2.7 to test our hypotheses: 1) Our work was exploratory in the sense that it examined the multiple new relationships in a complicated model, and many of the variables were not normal y distributed; and 2) Hair et al. (2017) emphasize that factor indeterminacy makes covariance-based SEM unsuitable for prediction purposes. Under these conditions, Hair et al. (2017) suggest PLS-SEM as the method to be used. 4 Analysis and Results 4.1 Assessment of Measurement Model Because the factor loadings were al >.78 (see Table 2), the composite reliabilities were al .75, and the Cronbach's alphas were al >.85 (the suggested cut-off value is .70) (see Table 3), the construct measures demonstrated good reliability and validity (Hair et al. 2017). Also, all Variance Inflation Factors (VIF) were below the threshold value of 5, which indicates that there were no collinearity problems (Hair, Ringle & Sarstedt, 2011). In order to formal y attain discriminant validity, the Fornel -Larcker criterion was applied (i.e., a latent variable should better explain the variance of its own indicators than the variance of other latent variables) (See Table 3). The Heterotrait-Monotrait (HTMT) Ratio was also examined (Appendix 2). All HTMT ratios were below the threshold value of.90 (Appendix 2), and the square root of the AVE of each latent variable outperformed the correlations with al the other latent variables (Table 3). J. Nyrhinen, A. Sirola, L. Frank, J. Nuckols, T.-Anna Wilska: An Ethical Perspective on Loot Box Purchasing - Examining Psychosocial Antecedents and the Association with Indebtedness 309 Table 2: Constructs and measures Construct Items M SD FL VIF Loneliness Thinking about the past year, how often have 1.86 .72 .85 1.82 (1 = hardly ever, you felt that you lack companionship? 2 = some of the time, Thinking about the past year, how often have 1.77 .72 .89 2.03 3 = often) you felt left out? Thinking about the past year, how often have 1.86 .73 .84 1.80 you felt isolated from others? COVID-19 worry Concern on the health of your loved ones. 3.69 1.16 .69 1.66 (1 = not worried at Concern on your own mental wel being. 3.04 1.28 .93 2.04 all; 5 = extremely worried) Concern on the mental wel being of your loved 3.31 1.21 .91 2.59 ones. Psychological I can deal with whatever comes my way. 2.55 .92 .78 1.63 resilience I believe I can achieve my goals, even if there 2.50 .97 .84 1.82 (0 = not true at al ; are obstacles. 4 = true nearly all of Under pressure, I stay focused and think 2.41 1.01 .80 1.71 the time) clearly. I think of myself as a strong person when 2.48 1.03 .82 1.76 dealing with life’s challenges and difficulties. Social support I receive a lot of understanding and security 3.69 1.16 .80 1.36 (1 = not true at al ; from others. 5 = very true) If I need to, I can borrow something from 3.04 1.28 .80 1.56 friends or neighbors without any problems. I know several people with whom I like to do 3.31 1.21 .84 1.63 things. Notes: M = Mean, SD = Standard Deviation, FL = Factor Loading, VIF = Variance Inflation Factor 310 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 3: Construct reliability and validity and discriminant validity CA CR AVE 1 2 3 4 5 6 7 8 9 1 n/a n/a n/a n/a 2 .82 .88 .72 -.20 .85 3 n/a n/a n/a -.07 .03 n/a 4 n/a n/a n/a -.17 .17 .05 n/a 5 n/a n/a n/a -.12 .11 -.16 .05 n/a 6 .83 .90 .74 -.32 .33 -.01 .13 .23 .86 7 n/a n/a n/a -.34 .11 .01 -.11 .17 .20 n/a 8 .82 .88 .65 .14 -.12 .13 -.03 -.18 -.29 -.06 .81 9 .74 .85 .66 .01 .07 .10 .05 -.17 -.28 .03 .40 .81 Notes: CA = Cronbach’s Alpha, CR = Composite Reliability, AVE = Average Variance Extracted; n/a = not applicable; 1 = Age, 2 = COVID-19 Worry, 3 = EDUC = Education, 4 = Gender, 5 = Indebtedness, 6 = Loneliness, 7 = Loot box purchasing, 8 = Psychological resilience, 9 = Social support; Square roots of AVES on diagonal Henseler (2014) suggested that the standardized root mean square residual (SRMR) should be used to assess the model fit in PLS-SEM because it does not give model fit statistics as covariance-based SEM does. In common factor models, values below .08 are regarded as a good fit. In our investigation, the SRMR value was only .05, and thus much below the threshold value. 4.1 Assessment of Structural Model To test our hypotheses we first analyzed the direct associations. We used the bootstrapping technique with 5,000 re-samples to determine the relevance of the pathways. Table 4 displays the outcomes of the testing of our hypothesis. J. Nyrhinen, A. Sirola, L. Frank, J. Nuckols, T.-Anna Wilska: An Ethical Perspective on Loot Box Purchasing - Examining Psychosocial Antecedents and the Association with Indebtedness 311 Table 4: Results of hypotheses testing Hyp DV IV β f2 R2 H1 Loneliness COVID-19 worry .35*** .15 .20 H2 Loot box purchasing COVID-19 worry ns H3 Loneliness Social support -.30*** .11 H4 Loot box purchasing Social support .10*** .01 .05 H5 Psychological resilience Social support .40*** .19 .16 H6 Loot box purchasing Loneliness .21*** .04 H7 Loot box purchasing Psychological resilience -.04** .00 H8 Indebtedness Loot box purchasing .16*** .02 .07 Cntl Indebtedness Age -.07*** .00 Indebtedness Gender .06*** .00 Indebtedness Education -.17*** .03 Notes: ns=not significant, **=p < .01, ***=p < .001; Hyp = Hypotheses, Cntrl = Control variables, DV = Dependent variable, IV = Independent variable As Table 4 presents, the conceptual model for 20% of the variance in 'Loneliness', 5% of the variance in 'Loot box purchasing', 16% of the variance in 'Psychological resilience', and 7% of the variance in 'Indebtedness'. Al of our hypotheses with the exception of H2 were supported (Table 4). Thus, only the hypothesised association between 'COVID-19 worry' and 'Loot box purchasing' was found not statistical y significant. However, we also tested for mediation and discovered that there was a positive total effect between 'COVID-19 worry' and 'Loot box purchasing' ( β = .21, p < .001), and a positive specific indirect effect from 'COVID-19 worry' to 'Loot box purchasing' through 'Loneliness' ( β = .07, p < .001). Therefore, the association between 'COVID-19 worry' and 'Loot box purchasing' seems to be indirect and mediated through 'Loneliness'. With respect to H1, 'COVID-19 worry' was positively associated with 'Loneliness' ( β = .35, p < .001). 'Social support' had a negative association with 'Loneliness', supporting H3 ( β = -.30, p < .001). With respect to H4 and H5, 'Social support' had also positive associations with both, 'Loot box purchasing' ( β = .10, p < .001) and 'Psychological resilience' ( β = .40, p < .001). 'Loneliness' was positively associated with 'Loot box purchasing' ( β = .21, p < .001), supporting H6. Contrary, ' Psychological resilience' had a negative association with 'Loot box purchasing' ( β = -.04, p < .01), confirming H7. As hypothesized in H8, 'Loot box purchasing' was positivetively associated with 'Indebtedness' ( β = .16, p < .001). The control variables also had significant associations to 'Indebtedness' which 312 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY is the outcome variable of our model. However, the effec sizes of age and gender were smal er than the effect of 'Loot box purchasing' on 'Indebtedness'. Education had sligthly bigger effect on 'Indebtedness' than 'Loot box purchasing', but it did not efface the effect of 'Loot box purchasing'. 5 Conclusion and Discussion This paper investigated loot box purchase behavior, indebtedness, and associated psychosocial factors during the first year of the COVID-19 pandemic. The findings supported most of the proposed hypotheses, and bring valuable insight into psychosocial and financial vulnerabilities in loot box buying behavior. Thus, the findings also serve as a ground for digital ethics issues related to the phenomena. Psychological resilience was negatively associated with loot box purchasing, indicating that psychological resources can protect from the development of problem behaviors particuarly during crises situations (Brailovskaia & Margraf, 2022; Fletcher & Sarkar, 2013). COVID-19 worry, however, was not associated with increased loot box purchasing. Even though worrying about coronavirus was a common experience during the first year of the pandemic, with excessive worry making people vulnerable to problems such as addictive behaviors (Avena et al., 2021), our results did not support this idea in relation to increased loot box buying. The findings regarding the role of loneliness and its association with loot box purchasing indicates that loneliness is a major risk to problematic monetary behaviors. Indeed, loneliness is associated with a myriad of harms and problems such as online addictive behaviors (Kuss et al., 2014) and problematic gambling (Sirola et al., 2023). However, given that social support was positively associated with loot box purchasing, it is important to understand the sources and quality of support. Even though meaningful social relationships are crucial for wellbeing, social support is not always beneficial and can promote or help to maintain harmful behaviors (Savolainen et al., 2022; Sirola et al., 2021). Given that social motives are a central aspect for team-based digital gaming and in-game purchases (Sirola et al, 2021), it is possible that social relationships and peer pressure inside the game might lead to excessive loot box purchasing. Individuals who are already social y excluded from meaningful in-person relationships, and thus prone to loneliness, may seek out meaningful social contacts via video games and video game communities. J. Nyrhinen, A. Sirola, L. Frank, J. Nuckols, T.-Anna Wilska: An Ethical Perspective on Loot Box Purchasing - Examining Psychosocial Antecedents and the Association with Indebtedness 313 The findings also revealed an association between loot box purchasing and indebtedness. However, it is likely that loot box purchasing is not the only reason for debt problems. For example, excessive loot box purchasing often co-occurs with problematic gambling (Delfabbro & King, 2020), and thus, underlying gambling problems might at least partial y explain the association between loot box purchasing and debt problems. Nevertheless, loot box purchasing can add to one's financial strain particularly when problematic (Hing et al., 2022) and thus reinforce financial problems and vulnerabilities. Loot boxes are typical y made attractive for gamers, and their mechanisms encourage players to make hasty purchase decisions, thus posing risks for vulnerable individuals in particular. These kinds of monetary mechanisms and their resemblance with mechanisms of gambling in digital games are an essential contributor to develop addictive or disordered gaming habits and consequent problems (Király et al., 2023). Our findings higlight that ethical concerns that are related to gambling apply also to loot boxes. Policy makers and game companies should consider these ethical aspects when fostering responsible gaming, similarly to how legislation, regulation and the industry's own self-regulation is applied to gambling. These regulation practices would be crucial to protect particularly vulnerable players such as minors and individuals with psychosocial problems. This study is not without limitations. First, data were cross-sectional and thus, causal interpretations between the studied variables are theoretical. Second, the self-reported nature of the survey items might be prone to biases. Third, the amount of money used in loot box purchasing, motives for buying loot boxes, or the particular games played were not asked, and these should be targeted by future studies. Additionally, the source of social support was not specified. Given that social support from other gamers or online networks is likely to work differently compared to support from one’s family members, for example, future studies should also investigate different sources of social support in more detail. It would be important to study how other (excessive) purchase behaviors, such as excessive gambling, overlap with loot box spending and contribute to debt problems. Our study focused on three culturally relatively similar European countries; thus, more cross-national studies would provide insight on potential cultural differences in loot box purchasing and associated factors. Finally, longitudinal studies are needed to gain evidence on the causal mechanisms in loot box purchasing. 314 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY References Achtziger, A. (2022). Overspending, debt, and poverty. Current opinion in psychology, 101342. Avena, N. M., Simkus, J., Lewandowski, A., Gold, M. S., & Potenza, M. N. (2021). Substance use disorders and behavioral addictions during the COVID-19 pandemic and COVID-19-related restrictions. Frontiers in Psychiatry, 12, 653674. Brady, A., & Prentice, G. (2021). Are loot boxes addictive? Analyzing participant’s physiological arousal while opening a loot box. Games and Culture, 16(4), 419-433. Brailovskaia, J., & Margraf, J. (2022). Addictive social media use during Covid-19 outbreak: Validation of the Bergen Social Media Addiction Scale (BSMAS) and investigation of protective factors in nine countries. Current Psychology, 1-19. Broos, H. C., Llabre, M. M., Saab, P. G., Leite, R. O., Port, J. H., & Timpano, K. R. (2023). The relationship between health worry, work distress, and affective symptoms during the COVID-19 pandemic: The mediating role of hopelessness and helplessness. British journal of clinical psychology, 62(1), 10-27. Campbell-Sills, L., & Stein, M. B. (2007). Psychometric analysis and refinement of the Connor– Davidson resilience scale (CD‐RISC): validation of a 10‐item measure of resilience. Journal of Traumatic Stress: Official Publication of The International Society for Traumatic Stress Studies, 20(6), 1019–1028. Carey, P. A. K., Delfabbro, P., & King, D. (2022). An evaluation of gaming-related harms in relation to gaming disorder and loot box involvement. International Journal of Mental Health and Addiction, 20(5), 2906-2921. Cohen, S., & Wil s, T. A. (1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 98(2), 310. Connor, K. M., & Davidson, J. R. (2003). Development of a new resilience scale: the Connor–Davidson resilience scale (CD-RISC). Depression and Anxiety, 18(2), 76–82. Delfabbro, P., & King, D. L. (2020). Gaming-gambling convergence: Evaluating evidence for the ‘gateway’hypothesis. International Gambling Studies, 20(3), 380-392. Dowling, N. A., Merkouris, S. S., Greenwood, C. J., Oldenhof, E., Toumbourou, J. W., & Youssef, G. J. (2017). Early risk and protective factors for problem gambling: a systematic review and meta-analysis of longitudinal studies. Clinical Psychology Review, 51, 109–124. Drummond, A., & Sauer, J. D. (2018). Video game loot boxes are psychological y akin to gambling. Nature human behaviour, 2(8), 530-532. Fletcher, D., & Sarkar, M. (2013). Psychological resilience. European Psychologist, 18(1). Frank, L., Salo, M.; & Toivakka, A. (2015). Why Buy Virtual Helmets and Weapons? Introducing a Typology of Gamers. BLED 2015 Proceedings. 25. https://aisel.aisnet.org/bled2015/25 Hing, N., Rockloff, M., Russel , A. M., Browne, M., Newal , P., Greer, N., . . & Thorne, H. (2022). Loot box purchasing is linked to problem gambling in adolescents when control ing for monetary gambling participation. Journal of Behavioral Addictions. Hughes, M. E., Waite, L. J., Hawkley, L. C., & Cacioppo, J. T. (2004). A short scale for measuring loneliness in large surveys: results from two population-based studies. Research on Aging, 26(6), 655–672. Irie, T., Shinkawa, H., Tanaka, M., & Yokomitsu, K. (2022). Online-gaming and mental health: Loot boxes and in-game purchases are related to problematic online gaming and depression in adolescents. Current Psychology, 1-12. Khazaal, Y., Chatton, A., Achab, S., Monney, G., Thorens, G., Dufour, M., & Rothen, S. (2017). Internet gamblers differ on social variables: a latent class analysis. Journal of Gambling Studies, 33(3), 881–897. King, D. L., Delfabbro, P. H., Bil ieux, J., & Potenza, M. N. (2020). Problematic online gaming and the COVID-19 pandemic. Journal of Behavioral Addictions, 9(2), 184-186. Király, O., Koncz, P., Griffiths, M. D., & Demetrovics, Z. (2023). Gaming disorder: A summary of its characteristics and aetiology. Comprehensive Psychiatry, 152376. J. Nyrhinen, A. Sirola, L. Frank, J. Nuckols, T.-Anna Wilska: An Ethical Perspective on Loot Box Purchasing - Examining Psychosocial Antecedents and the Association with Indebtedness 315 Kliem, S., Mößle, T., Rehbein, F., Helmann, D. F., Zenger, M., & Brähler, E. (2015). A brief form of the Perceived Social Support Questionnaire (F-SozU) was developed, validated, and standardized. Journal of Clinical Epidemiology, 68(5), 551–562. Kuss, J. D., Griffiths, M. D., Karila, L., & Bil ieux, J. (2014). Internet addiction: A systematic review of epidemiological research for the last decade. Current pharmaceutical design, 20(25), 4026-4052. Lin, M., Hirschfeld, G., & Margraf, J. (2019). Brief form of the Perceived Social Support Questionnaire (F-SozU K-6): validation, norms, and cross-cultural measurement invariance in the USA, Germany, Russia, and China. Psychological Assessment, 31(5), 609. Lussier, I., Derevensky, J. L., Gupta, R., Bergevin, T., & El enbogen, S. (2007). Youth gambling behaviors: an examination of the role of resilience. Psychology of Addictive Behaviors, 21(2), 165. Mancini, A. D., & Bonanno, G. A. (2009). Predictors and parameters of resilience to loss: toward an individual differences model. Journal of Personality, 77(6), 1805–1832. Mishra, S., Beshai, S., Wuth, A., & Refaie, N. (2019). Risk and protective factors in problem gambling: an examination of psychological resilience. International Gambling Studies, 19(2), 241–264. Mónaco, E., Schoeps, K., Valero-Moreno, S., Castro-Calvo, J., Montoya-Castil a, I., Del Rosario, C., & Esparza, N. A. A. (2022). Cross-cultural validation of the worries about COVID-19 and its consequences scale (W-COV) in adolescents and young people. Archives of Psychiatric Nursing, 40, 158–166. Nordmyr, J., & Forsman, A. K. (2020). A systematic review of psychosocial risks for gambling and problem gambling in the nordic countries. Health Risk & Society, 22(3–4), 266–290. Oei, T. P., & Goh, Z. (2015). Interactions between risk and protective factors on problem gambling in Asia. Journal of Gambling Studies, 31(2), 557–572. Okruszek, Ł., Aniszewska-Stańczuk, A., Piejka, A., Wiśniewska, M., & Żurek, K. (2020). Safe but lonely? Loneliness, anxiety, and depression symptoms and COVID-19. Frontiers in Psychology, 11, 579181. Ranta, M., Silinskas, G., & Wilska, T. A. (2020). Young adults’ personal concerns during the COVID-19 pandemic in Finland: an issue for social concern. International Journal of Sociology and Social Policy, 40, 1201–1219. Räsänen, T., Lintonen, T., Tolvanen, A., & Konu, A. (2016). Social support as a mediator between problem behaviour and gambling: a cross-sectional study among 14–16-year-old finnish adolescents. BMJ Open, 6(12), e012468. Savolainen, I., Sirola, A., Vuorinen, I., Mantere, E., & Oksanen, A. (2022). Online communities and gambling behaviors—A systematic review. Current Addiction Reports, 9(4), 400-409. Scholes-Balog, K. E., Hemphil , S. A., Toumbourou, J. W., & Dowling, N. A. (2015). Problem gambling and internalising symptoms: a longitudinal analysis of common and specific social environmental protective factors. Addictive Behaviors, 46, 86–93. Sirola, A., Nyrhinen, J., & Wilska, T. A. (2023). Psychosocial Perspective on Problem Gambling: The role of Social Relationships, Resilience, and COVID-19 Worry. Journal of gambling studies, 1-19. Sirola, A., Kaakinen, M., Savolainen, I., & Oksanen, A. (2019). Loneliness and online gambling-community participation of young social media users. Computers in Human Behavior, 95, 136– 145. Sirola, A., Savela, N., Savolainen, I., Kaakinen, M., & Oksanen, A. (2021). The role of virtual communities in gambling and gaming behaviors: A systematic review. Journal of Gambling Studies, 37(1), 165-187. Spicer, S. G., Nicklin, L. L., Uther, M., Lloyd, J., Lloyd, H., & Close, J. (2022). Loot boxes, problem gambling and problem video gaming: A systematic review and meta-synthesis. New Media & Society, 24(4), 1001-1022. Vuorinen, I., Oksanen, A., Savolainen, I., Sirola, A., Kaakinen, M., Paek, H. J., & Zych, I. (2021). The mediating role of psychological distress in excessive gambling among young people: a four- 316 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY country study. International Journal of Environmental Research and Public Health, 18(13), 6973. Yokomitsu, K., Irie, T., Shinkawa, H., & Tanaka, M. (2021). Characteristics of gamers who purchase loot box: a systematic literature review. Current Addiction Reports, 8, 481-493. Yücel, M., Whittle, S., Youssef, G. J., Kashyap, H., Simmons, J. G., Schwartz, O., & Al en, N. B. (2015). The influence of sex, temperament, risk-taking and mental health on the emergence of gambling: a longitudinal study of young people. International Gambling Studies, 15(1), 108– 123. Zendle, D., Meyer, R., & Over, H. (2019). Adolescents and loot boxes: Links with problem gambling and motivations for purchase. Royal Society Open Science, 6(6), 190049. Zhang, N., Yang, S., & Jia, P. (2022). Cultivating resilience during the COVID-19 pandemic: A socioecological perspective. Annual Review of Psychology, 73. J. Nyrhinen, A. Sirola, L. Frank, J. Nuckols, T.-Anna Wilska: An Ethical Perspective on Loot Box Purchasing - Examining Psychosocial Antecedents and the Association with Indebtedness 317 Appendix 1 Sample Characteristics FINLAND SWEDEN GREAT BRITAIN Gend. Smp Pop. Gend. Smp Pop. Gend. Smp Pop. Man 50 % 50 % Man 50 % 50 % Man 49 % 49 % Woman 50 % 50 % Woman 50 % 50 % Woman 51 % 51 % Age Smp Pop. Age Smp Pop. Age Smp Age 18–22 8 % 8 % 18–22 8 % 8 % 18–22 6 % 8 % 23–35 26 % 23 % 23–35 24 % 25 % 23–35 30 % 24 % 36–55 38 % 34 % 36–55 40 % 36 % 36–55 38 % 37 % 56–75 29 % 36 % 56–75 28 % 31 % 56–75 26 % 30 % Region Smp Pop. Region Smp Pop Region Smp Pop S.Fin. 46 % 52 % Mid-Nrdlnd 4 % 4 % East England 9 % 11 % E.Fin. 11 % 11 % NCentral 9 % 8 % London 14 % 15 % Sweden W.Fin 31 % 25 % Småland islands 9 % 8 % Midlands 16 % 10 % incl. N.Fin 12 % 12 % Stockholm 21 % 23 % Yorkshire and 12 % 9 % Humber S. Swe 15 % 15 % Northwestern 11 % 12 % W.Swe 19 % 20 % N.Ireland 3 % 3 % E.Central 17 % 17 % Scotland 8 % 9 % Sweden Uppr. Norland 6 % 5 % SE.England 15 % 16 % SW.England 8 % 9 % Wales 5 % 5 % 318 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Appendix 2 The Heterotrait-Monotrait Ratio Age (1) 2 3 4 5 6 7 8 COVID-19 Worry (2) .19 Education (3) .07 .04 Gender (4) .17 .18 .04 Indebtedness (5) .12 .10 .16 .05 Loneliness (6) .35 .35 .01 .15 .25 Loot box purchasing (7) .34 .10 .01 .11 .17 .22 Resilience (8) .16 .13 .14 .03 .20 .36 .07 Social support (9) .03 .13 .12 .06 .19 .35 .04 .51 HOW DO QR CODES ENHANCE CUSTOMER EXPERIENCE? OMNICHANNEL CUSTOMER EXPERIENCES IN A BRICK-AND-MORTAR FASHION STORE MATILDA HOLKKOLA,1 TIINA PAANANEN,1 LAURI FRANK,1 TIINA KEMPPAINEN,2 MARKUS MAKKONEN1 1 University of Jyvaskyla, Faculty of Information Technology, Jyvaskyla, Finland matilda.i.holkkola@jyu.fi, tiina.e.paananen@jyu.fi, lauri.frank@jyu.fi, markus.v.makkonen@jyu.fi 2 University of Jyvaskyla, Jyvaskyla University School of Business and Economics, Jyvaskyla, Finland ti na.j.kemppainen@jyu.fi In this qualitative study, we investigate how customers perceive QR codes enhancing their customer experience (CX) in the brick-and-mortar (B&M) store environment. Introducing QR codes to B&M stores is a part of the larger phenomenon of retailers creating omnichannel CXs by utilizing digital and cyber- physical elements in B&M stores. To col ect data from real customers’ CXs, we displayed QR codes in a Finnish fashion brand’s B&M store and interviewed customers who had authentically visited the store. The QR codes displayed in the Keywords: store were linked to a product information page on the brand’s QR codes, online store, a brand ambassador’s Instagram posts, and an brick-and-mortar, online survey. The data consist of 15 individual semi-structured omnichannel, customer thematic interviews of customers aged under 50 years. By experience, conducting qualitative content analysis, we found six main store advantages (i.e., interestingness, informativeness, usability, environment, cyber-physical, attractiveness, interaction, and innovativeness) and 14 subgroups B&M, of QR codes’ advantages in enhancing the CX. retail DOI https://doi.org/10.18690/um.fov.4.2023.20 ISBN 978-961-286-751-5 320 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The advancements in information and communication technologies, such as smartphones and Quick Response (QR) codes, enable today’s consumers to gain additional information on products during their brick-and-mortar (B&M) store visits. According to Briedis et al. (2020), customers prefer omnichannel customer experiences, where they can seamlessly utilize physical and digital channels (Rigby, 2011). During their B&M store visits, it is more common for customers to use the competing retailers’ online channels than the online channels of that same B&M retailer (Spaid et al., 2019). For B&M retailers, this means losing potential customers due to showrooming (i.e., gathering information offline but comparing options and purchasing online) as wel as online shopping. Thus, B&M retailers have begun to include digital and cyber-physical elements in store environments to offer unique customer experiences in B&M stores (Lemon & Verhoef, 2016). For example, QR codes (Yoon & Kim, 2014), robot assistants, and smart mirrors (Sheth, 2021) have been implemented in B&M store environments. Sheth (2021) presents that “B&M retailers can not only survive but also thrive by fusion of physical and digital [worlds]”. QR codes have been shown to enhance customer experience (CX) in online shopping (Hossain et al., 2018) and shopping in general (Albăstroiu & Felea, 2015), but there is a research gap in how QR codes can enhance CX in B&M fashion stores. The topic is important from both theoretical and practical perspectives because enhancing the CX in B&M stores with omnichannel elements is recommended (Sheth, 2021), but the research on QR codes’ potential in offering additional information, product reviews, and relevant social media content for B&M fashion store customers is stil scarce (Albăstroiu & Felea, 2015). Also, research on QR codes’ technological attributes (Kim & Yoon, 2014), consumer acceptance (Kim & Woo, 2016; Kim et al., 2014), and efficiency in providing product information (Rotsios et al., 2022) has been called for. According to Albăstroiu and Felea (2015), QR codes have usual y been studied from a technology use perspective, such as the design and functionality of digital technologies in B&M stores (Zimmermann et al., 2022). However, there is also a research gap in studying how also the mere presence of QR codes can enhance the CX as a physical in-store stimulus (Bitner, 1992). M. Holkkola, T. Paananen, L. Frank, T. Kemppainen, M. Makkonen: How Do QR Codes Enhance Customer Experience? Omnichannel Customer Experiences in a Brick-and-Mortar Fashion Store 321 Therefore, in this qualitative study, we investigate how customers perceive QR codes enhancing their CX in a B&M store. The data collection was conducted by displaying QR codes within a Finnish B&M fashion store and, after that, interviewing the customers who had visited the store. In the next section, we present the relevant research on CX in omnichannel B&M store environments and on QR codes. In the third section, we introduce our QR code intervention case, research data, and methods used. In the fourth section, we present the findings and discussion of this study. Finally, in the fifth section, we conclude with the conclusion, theoretical and practical implications, limitations of the study, and our suggestions for future research. 2 Theoretical Background 2.1 Omnichannel CX in B&M Store Environments CX consists of customers’ subjective responses, feelings, and interpretations (Gentile et al., 2007). It has been researched from many overlapping dimensions (Holkkola et al., 2022a), such as cognitive and physical dimensions (Verhoef et al., 2009). The cognitive dimension includes customers’ curiosity (Schmitt, 1999) and experiences of flow (Novak et al., 2000). The physical dimension is traditionally associated with offline elements, such as temperature, odor, furnishing, and signage (Bitner, 1992). There are plenty of studies on the physical dimension of CX in B&M stores, but as retailers have introduced digital and cyber-physical elements into their B&M stores, a new wave of CX research focuses on these novel elements (Zimmermann et al., 2022). To stand out in this omnichannel era, Parise et al. (2016) advise that new digital and cyber-physical elements should be introduced to B&M stores to create immersive, distinctive (Briedis et al., 2020), and customized CXs (Zimmermann et al., 2022). For instance, virtual mirrors are described as “something the customers wil remember” (Parise et al., 2016). These kinds of elements can enhance CX by immersion and flow among customer touchpoints, which wil positively affect customer behavior and attitudes (Parise et al., 2016). For example, an augmented reality (AR) assistive shopping mobile app was found to increase informativeness and usefulness in CX (Zimmermann et al., 2022). Informativeness consists of good and relevant product information (Hausman and Siekpe, 2009), such as product availability and price comparison information in a 322 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY store (Parise et al., 2016). Usefulness consists of perceived shopping performance, productivity, and effectiveness (Hausman & Siekpe, 2009). In addition, digital and cyber-physical elements, such as remote experts and digital assistants, can create highly immersive engagement with products as they can build a narrative behind the product simultaneously during the B&M store visit. This will result in a closer connection between the consumer and the brand, such as return visits and loyalty to the brand (Parise et al., 2016). Digital and cyber-physical elements can also enhance CX with personalization and interactivity (Parise et al., 2016). Interactivity gives consumers a sense of control as they can have an active part in engaging with the brand (Klein, 2003). 2.2 QR Codes QR codes are two-dimensional barcodes that can hold more information than one- dimensional Universal Product Codes (UPC) and can be scanned with a device that has a camera, a QR code reader software, and an Internet connection (Crompton et al., 2012). QR codes were first used in the auto industry (Denso Wave, 2022) and have since become common in many areas. QR code usage increased by 98% from 2018 to 2020 and is predicted to cover 80% of check-out services, orders, and payments by 2024 (Scanova Blog, 2022). In prior QR code research, customers have shown positive and negative attitudes toward and experiences with QR codes. Negative experiences include, for example, privacy concerns towards QR codes (Okazaki et al., 2012) and irritation from mobile tools usage in a B&M store (Zimmermann et al., 2022). However, many positive customer perceptions, attitudes, and effects of QR codes have also been found in prior research. Consumers have perceived QR codes to enhance their CX in multiple contexts (Albăstroiu & Felea, 2015) and in an imaginary B&M store scenario (Kjeldsen et al., 2023). In a smart shopping context, Kim and Yoon (2014) studied QR code virtual grocery and fashion stores that could be visited by scanning a QR code on the screen in South Korean subway stations. They identified the technological attributes of QR codes, including ubiquity, instant interactivity, digital signage, and wireless infrastructure, which were valued by fashion shoppers. Also, B&M grocery store customers found food traceability QR codes informative, useful, easy to use, and a positive invention (Kim & Woo, 2016). Yoon and Kim (2014) found QR codes important for B&M retailers’ multichannel strategies in integrating online channels M. Holkkola, T. Paananen, L. Frank, T. Kemppainen, M. Makkonen: How Do QR Codes Enhance Customer Experience? Omnichannel Customer Experiences in a Brick-and-Mortar Fashion Store 323 into B&M stores, as QR codes can offer customers additional and abundant information and channel integration. In the advertising context, young consumers were suggested to value the informativeness of QR codes as this perceived benefit strongly predicted their intention to scan them (Jung et al., 2012). In educational work, in turn, teachers perceived QR codes as being fun, interesting, engaging, attractive, enjoyable, and a new way of learning (Ali et al., 2017). Similarly, younger consumers found QR codes attractive and useful (Ertekin & Pelton, 2015). According to Dou and Li (2008), QR codes’ advantages for customers include being less intrusive and more engaging communication tools than shortcodes and image codes. QR codes have also proven to be a good tool from a cognitive perspective, and in brands’ communication, they have been suggested to arise customers’ curiosity (Srinounpan et al., 2020). Curiosity was also reported as one of the main reasons to scan QR codes for Polish consumers (Nogieć, 2017). 3 Methodology To col ect data from real customers in an authentic QR code setting, we displayed QR codes inside a B&M factory store of a wel -known Finnish fashion brand. The QR codes were printed with suitable texts and pictures in A4 and smaller papers, laminated, and displayed inside the store from May 2022 to September 2022. Three different sets of QR codes were used and placed in the store. The first set of QR codes (cf. Appendix 2) was placed with a product, anorak jackets, and their landing page was the product information page on the brand’s online store. The second set of QR codes directed to a brand ambassador’s Instagram posts, where she promoted the brand’s selected clothes, also displayed on a separate clothes rack in the B&M store. Thus, both the online and social media channels of the brand were utilized in providing additional information and inspiration for customers. The third set of QR codes directed customers to our online survey where one could also volunteer for individual interviews and be rewarded with a prize (a canvas bag). This set of QR codes was also printed and handed with one’s purchases by store personnel. During the time the QR codes were on display, we recruited participants for individual interviews with the third set of QR codes as well as pen-and-paper surveys in the B&M store. Due to our target group, B&M store customers under 50 years of age were eligible to participate in the interviews. This was because using mobile devices for searching information in B&M stores is common for consumers aged 324 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY under 50 years and significantly less common for older consumers (Holkkola et al., 2022b) who are suggested to have lower self-efficacy in mobile and omnichannel shopping (Makkonen et al., 2022). Similarly, Ertekin and Pelton (2015) excluded consumers born in the 1960s or earlier from their QR code study. For a data collection method, semi-structured individual thematic interviews were chosen since we wanted a profound understanding of CXs. We used semi-structured interviews as they al ow more free discussion and enable additional questions (Myers & Newman, 2007). Before the interviews, we conducted a test interview to try out our interview framework. The interviews were conducted in September 2022. Due to our recruiting technique, the participants were the brand’s real customers who had visited the B&M store while the QR codes were on display. The number of interviews was determined by saturation (Fusch & Ness, 2015). The participants (referred to as Participants 1–15: P1–P15) consisted of twelve females and three males who were 23–42 years old (for details, see Appendix 1). Instead of statistically representative results, the qualitative approach of this study aims to gain an in-depth understanding of the customers’ perceptions and was chosen because it is suitable for studying experiences and phenomena with little prior research. The interviews were held remotely via Microsoft Teams. The average duration of the interviews was 58 minutes. The interview themes covered omnichannel behavior, previous experiences with QR codes, the QR codes in the case B&M store, and customer engagement with the case brand. During the theme of QR codes in the case B&M store, we showed the participants pictures of the B&M store with QR codes (Appendix 2) to make it easier for them to share their experiences. The interviews were recorded and transcribed. In the analysis, we used qualitative content analysis in a data-driven manner because of the experimental nature of this study. As tools, the ATLAS.ti software and Microsoft Word were utilized for coding the data inductively with two coding cycles (Gioia et al., 2013). This resulted in 14 codes which were used as 14 subgroups of the findings, respectively. These 14 subgroups could ultimately be classified under six broader themes that form the main findings and are presented in the next section. The six themes and the 14 subgroups are also displayed in Figure 1 in the fifth section. M. Holkkola, T. Paananen, L. Frank, T. Kemppainen, M. Makkonen: How Do QR Codes Enhance Customer Experience? Omnichannel Customer Experiences in a Brick-and-Mortar Fashion Store 325 4 Findings and Discussion In the participants’ CXs in the B&M environment, we found six main advantages of QR codes: interestingness, informativeness, usability, attractiveness, interaction, and innovativeness. Next, we present these in relation to earlier research. The citations have been translated from Finnish to English. Interestingness. Firstly, the displayed QR codes were found interesting and aroused the participants’ curiosity for multiple reasons. Some perceived them as new technology, and others found them curious in the B&M environment. The participants were interested in “where the QR code leads and what it brings me” (P14). This is consistent with the findings of Srinounpan et al. (2020) and Nogieć (2017) and supports Ali et al.’s (2017) school context findings of QR codes being perceived as interesting and engaging. P12 described how QR codes are like “mysteries” because one cannot see their contents from the outside. All in al , QR codes attracted attention and got some participants to scan them for fun. “I think QR codes in the store bring something new and interesting to you, like ’hey, what’s this about’. Probably because they are stil quite a new thing for me that I haven’t quite got utilizing yet.” – Female, 29 (P5) “Even though the anoraks were not relevant to me, I did notice the QR codes when I passed by and checked what the trick here is.” – Female, 32 (P14) For some, specifical y the B&M environment enabled scanning QR codes out of pure curiosity since it was perceived as a safe environment for digital experimentation. This was because information security of a B&M store’s QR codes was perceived as more reliable compared to, for example, renting electric scooters P14 mentioned as an example. This finding is in line with Okazaki et al.’s (2012) QR code privacy concern findings in urban outdoor spaces. Thus, the participants felt safe scanning the QR codes in the B&M environment and perceived this as fun and entertaining, consistent with Zimmermann et al.’s (2022) findings on digital elements creating more entertaining CXs. P14 describes her experience as fol ows: “I thought it was fun to be able to search on their pages how the product looks on a model. That was quite a nice idea”. 326 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Informativeness. In addition to their interestingness, QR codes were scanned because the participants valued information. The participants found QR codes with additional information useful and thought that they contribute positively to their CX in B&M, which is consistent with Zimmermann et al.’s (2022) digital B&M store element findings. The participants wanted to get information before the purchase decision, for example, on the materials, caretaking instructions, available colors and sizes, and sustainability. This finding is also consistent with studies on other product groups, such as food traceability QR codes that customers also found informative (Kim & Woo, 2016). For example, P14 describes her wishes for more information as follows: “While shopping, I thought that these QR codes could be on al the products, so that you could search available colors of those shorts, for example, as they all weren’t on display.” Additional information was valued, and it also mattered where this information was obtained from. It was remarkable that the participants valued that the QR codes led to the brand’s online store and the integration of the online and offline channels. The participants perceived linking the brand’s online store to the B&M store as “smart” (P2) and “fun” (P14), and for some, it was a nice surprise. “I thought that the QR code would just lead to some additional information page, I somehow didn’t expect it to go on the brand’s website. I expected maybe some kind of an additional page with a menu, but it was very nice that it led to the website.” – Female, 32 (P14) Like P14, the participants described their positive thoughts about this online store linkage. Indeed, improving the flow among customer touchpoints is suggested to be important in the omnichannel era (Parise et al., 2016). The participants perceived connecting the online store to the B&M store with QR codes as a good idea to advertise and remind customers of this channel’s existence. “I don’t think the [brand’s] webpage has otherwise been very advertised. There could be something or someone near the cash desk mentioning the webpage, but I’d say they don’t advertise it in the stores.” – Female, 42 (P9) Usability. QR codes were found usable and useful since they were perceived as easy to use and quick and enabled independent shopping. Perceived ease of use is consistent with the findings of Kim and Woo (2016) and was explained by many reasons. Using a smartphone in the B&M environment “felt easy because you’re so used to fiddling with your phone” (P2) and using the phone’s QR code reader felt M. Holkkola, T. Paananen, L. Frank, T. Kemppainen, M. Makkonen: How Do QR Codes Enhance Customer Experience? Omnichannel Customer Experiences in a Brick-and-Mortar Fashion Store 327 easier than going to the online store by typing the web address to the web browser. The ease of use of QR codes is identified also by Kim and Woo (2016). QR codes in the B&M environment were considered “a simplified way to find somewhere” (P8). Even though the participants did not primarily associate QR codes with B&M stores, many were already familiar with QR codes. “It’s very easy because via your camera, the QR code leads you directly to the page, and you don’t need to type any addresses. So, it’s handy in that sense.” – Female, 24 (P13) QR codes were also perceived as quick. Similarly, in Hausman and Siekpe’s (2009) model, shopping effectiveness is one aspect of digital shopping elements’ usability. Many participants emphasized the quickness of the direct linkage that QR codes enabled. The participants felt that the brand is striving to make information search quick and easy, contributing positively to their CX. “If the QR code leads me directly to the product page to get more information, it’s no doubt faster than if I started googling that additional information on the Internet.” – Female, 30 (P2) The usability of QR codes was also praised because of the enabled independence. Some participants preferred to not socialize with salespersons or send messages to their friends when shopping. Instead, their smartphone was their shopping assistant, which they even used to “signal the salesperson that you want to be in peace” (P2). In contrast to 50-year-old and older consumers who are unlikely to utilize mobile phones while in a B&M store (Holkkola et al., 2022b), using a mobile phone in B&M stores felt natural for these independent shoppers. “I know that the salespersons are happy to help, but if you’re like me, who wants to look for information yourself before asking, then QR codes are an easy way to do that.” – Female, 32 (P14) Attractiveness. The participants perceived that QR codes contribute to the B&M environment by making it look attractive. This was due to QR codes being visually compact solutions as digital signage (Kim & Yoon, 2014). Similarly, under-50-year-old consumers (Ertekin & Pelton, 2015) and teachers (Ali et al., 2017) found QR codes attractive. In addition to their compactness, the participants perceived QR codes as a natural addition to a B&M store and, similarly to Dou and Li’s (2008) findings, seeing them around was not found intrusive for most of the participants. 328 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY The participants perceived QR codes as compact tools for tastefuly sharing additional information. They did not want the store environment to be visual y crammed with product information, nor did they want the QR codes to be “the main point” (P10) in the store environment. Instead, the participants valued “a clear overall appearance” (P14). For example, P7 found QR codes “pleasant”, “restrained”, “simple”, and “straightforward”. QR codes as a space solution were seen to also respect those who did not need additional information. “Compared to, say, big posters with a lot of information, using QR codes is a simpler and more tasteful way, so that the information is available for those who want it.” – Female, 33 (P12) “QR codes make my store experience more comfortable because the store is not flooded from every corner with information, but it’s offered in a restrained way.” – Male, 24 (P7) In addition to compactness and respectfulness, the participants considered QR codes to be a natural element in the B&M environment. Many said that QR codes fit wel into the store environment and are displayed nicely. P12 describes how “it somehow seemed natural that ‘ah, there are QR codes here’, so if someone needs, they can use them”. Although many had not seen QR codes in B&M stores before, the participants mostly accepted this technology in the B&M environment. “The whole store had been renovated since my last visit, it was overal much fresher than before. So, I just thought that those [QR] codes are now part of it.” – Female, 30 (P2) Interaction. QR codes, especial y the ones that led to our online survey, were associated with the brand actively wanting to interact with its customers. P2 describes how she seized the opportunity to interact with the brand in the B&M store: “I thought that since I like the brand, I want to give feedback and then try to influence it.” Similarly, Klein (2003) found that interactivity gives consumers a sense of control as they can have an active part in engaging with the brand. According to Parise et al. (2016), digital and cyber-physical elements can also enhance CX in B&M stores with their interactivity. Indeed, QR codes were considered a good tool to interact with customers, since “many certainly have a phone with them” (P14) when visiting a B&M store. QR codes are found as engaging communication tools also in Dou and Li’s (2008) study. “Even though not being a super customer from [the brand’s] perspective”, P8 described the positive atmosphere the QR codes created for her during the B&M store visit: M. Holkkola, T. Paananen, L. Frank, T. Kemppainen, M. Makkonen: How Do QR Codes Enhance Customer Experience? Omnichannel Customer Experiences in a Brick-and-Mortar Fashion Store 329 “The store environment looks a little different when those QR codes are there. It arises the thought that the place is not only for buying but that the brand also wants to know something about the customers’ thoughts … The atmosphere of being interested in the customers’ thoughts is created there.” – Female, 33 (P8) Innovativeness. QR codes were perceived to enrich the CX in a B&M store and talk about the brand’s development. The participants valued that, besides traditional B&M shopping behaviors, they could get to know the products in new and enriched ways. Similarly, Ali et al. (2017) found that teachers perceived QR codes as a new way of learning, and Parise et al. (2016) found that rememberable digital elements improve CX. As P6 put it, “the brand bringing new ways of presenting information” was appreciated. For example, the participants found it creative that they could see how the products look on a model by scanning the QR code. These additional cyber-physical elements were perceived as “quite a good addition” (P2) and to have a positive influence on the whole brand. “Just by moving your thumb, you could get the same thing done as by flipping through big clothes racks and going around the store, so [with the QR codes] you got a quick overview of what they have here.” – Male, 24 (P7) This kind of digital enrichment was perceived as a sign of the brand’s development. For the participants, QR codes symbolized modernity, development, pioneering, and alternativeness. Similarly, QR codes are found to be perceived as a positive invention (Kim & Woo, 2016). Bringing QR codes to the B&M environment was associated with the B&M store and the whole brand being developed, in order “to keep up with the times” (P2). QR codes positively affected the store environment, which, in turn, affected the image of the whole brand. “Those QR codes as a final touch to the new look of the renovated store, it seems that the whole store has general y moved into a new era with a more modern touch. It certainly is a good thing.” – Female, 24 (P13) “The QR codes affect the brand image. They tell of the desire to develop something. Like [the brand] is not stuck in place but is constantly trying to reform their business model.” – Female, 25 (P10) 330 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 5 Conclusion, Limitations, and Future Research In this qualitative study, we investigated how QR codes can enhance CX in B&M stores and we found six main advantages: interestingness, informativeness, usability, attractiveness, interaction, and innovativeness. The findings are summarized in detail in Figure 1 below. Figure 1: Summary of the findings on how QR codes enhance CX in B&M stores In conclusion, this study addresses the research gap of the multiple ways how QR codes can enhance CX in fashion B&M environments. As a novel finding, we suggest that independent engagement with the products enabled by QR codes enhances some B&M customers’ CX. This study also shows that channel integration provided by QR codes enhances CX because, in line with Jung et al. (2012) and Kjeldsen et al. (2023), customers value additional information. In line with Kim and Woo (2016), engaging with the products and the brand is perceived as easy and quick with QR codes, making them usable tools that enable customer-company interaction (Klein, 2003). This results in enriched and yet effortless CX in B&M. QR codes also enhance the cognitive dimension of CX, as they arise curiosity among B&M store customers, which is in line with Srinounpan et al. (2020). Consistent with Nogieć (2017), it seems that the interestingness of QR codes stems partly from their novelty in B&M store environments. Additional y, the curiosity about QR codes is not likely to disappear soon due to their appearance which contains hidden information. In addition, bringing QR codes to the B&M environment resulted in positive associations with the whole brand, such as perceived development and perceived openness towards customer feedback enabled by QR codes as easy interaction possibilities (Kim & Yoon, 2014). These brand-related positive associations during M. Holkkola, T. Paananen, L. Frank, T. Kemppainen, M. Makkonen: How Do QR Codes Enhance Customer Experience? Omnichannel Customer Experiences in a Brick-and-Mortar Fashion Store 331 the B&M store visit contribute to CX (Parise et al., 2016) and, ultimately, may deepen the brand relationship (Paananen et al., 2022a; 2022b). In sum, we suggest that the presence of QR codes expands the physical B&M store environment practically into an omnichannel environment preferred by customers (Briedis et al., 2020). As a practical implication, we recommend B&M retailers to offer additional information with QR codes because of the identified six advantages. Integrating online channels with QR codes into B&M stores (Yoon & Kim, 2014) may also reduce the loss of today’s and tomorrow’s customers to competing retailers’ online channels (Spaid et al., 2019). Finally, this study does not come without limitations. Firstly, studying experiences is always situational, so the generalizability of the findings to other B&M stores and other cultural contexts needs future research with different product groups and in different countries. Secondly, this study focuses on how QR codes enhance CX, so research on QR codes’ possible negative contributions to CX is also needed. In addition, regardless of whether QR codes are used, it seems that their mere presence in the B&M store environment affects CX, which needs more research. For quantitative research on QR codes, interestingness, informativeness, usability, attractiveness, interaction, and innovativeness could be operationalized as empirical y measurable constructs. We also cal for more holistic omnichannel research on CX that recognizes digital and cyber-physical elements’ tangible and intangible interfaces and consequences in B&M environments. References Albăstroiu, I., Felea, M. (2015). Enhancing the shopping experience through QR codes: the perspective of the Romanian users. Amfiteatru Economic, 17(39), 553–566 Ali, N., Santos, I., Areepattamannil, S. (2017). Pre-service Teachers’ Perception of Quick Response (QR) Code integration in Classroom Activities. TOJET the Turkish online journal of educational technology, 16(1). Bitner, M. J. (1992). Servicescapes: The Impact of Physical Surroundings on Customers and Employees. Journal of Marketing, 56(2), 57–71. Briedis, H., Kronschnabl, A., Rodriguez, A. and Ungerman, K. (2020). Adapting to the Next Normal in Retail: the Customer Experience Imperative. McKinsey & Company. Crompton, H., LaFrance, J., & van't Hooft, M. (2012). QR Codes 101. Learning and leading with technology, 39(8), 22. Denso Wave (2022). QR development story. https://www.denso-wave.com/en/technology/vol1.html (2022), Accessed 3rd March 2023. Dou, X., Li., H. (2008). Creative Use of Qr Codes in Consumer Communication. International Journal of Mobile Marketing, 3(2), 61–67. 332 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Ertekin, S., Pelton, L.E. (2015). An exploratory study of consumer attitudes towards QR code reader applications. In Kubacki, K. (Ed.), Ideas in Marketing: Finding the New and Polishing the Old. Springer, Cham, 185–191. Fusch, P. I., Ness, L. R. (2015). Are We There Yet? Data Saturation in Qualitative Research. The Qualitative Report, 20(9), How To Article 1, 1408–1416. Gentile, C., Spil er, N., Noci, G. (2007). How to Sustain the Customer Experience: An Overview of Experience Components that Co-create Value With the Customer. European Management Journal, 25, 395–410. Gioia, D. A., Corley, K. G., Hamilton, A. L. (2013). Seeking Qualitative Rigor in Inductive Research: Notes on the Gioia Methodology. Organizational Research Methods, 16(1), 15–31. Hausman, A.V., Siekpe, J.S. (2009). The effect of web interface features on consumer online purchase intentions. Journal of Business Research, 62(1), 5–13. Holkkola, M., Frank, L., Kemppainen, T., Paananen, T., Luhtanen, V. (2022a). The role of sustainability in online customer experiences : a qualitative study on female fashion shoppers. In MCIS 2022 : Proceedings of the 14th Mediterranean Conference on Information Systems. Association for Information Systems. Holkkola, M., Nyrhinen, J., Makkonen, M., Frank, L., Karjaluoto, H., Wilska, T.-A. (2022b). Who are the showroomers? Socio-demographic factors behind the showrooming behavior on mobile devices. In A. Pucihar, M. Kljajić Borštnar, R. Bons, A. Sheombar, G. Ongena, & D. Vidmar (Eds.), Proceedings of the 35th Bled eConference: Digital restructuring and human (re)action. 113–128. University of Maribor Press. Hossain M. S., Zhou X, Rahman M. F. (2018). Examining the impact of QR codes on purchase intention and customer satisfaction on the basis of perceived flow. International Journal of Engineering Business Management 10. Jung, J., Somerstein, R., Kwon, E. (2012). Should I scan or should I go?: Young consumers' motivations for scanning QR code advertising. International Journal of Mobile Marketing, 7(3). Kim, E. Y., Yoon, N. (2014). Perceived QR code technological attributes in the smart shopping context. Journal of global fashion marketing, 5(4), 297–307. Kim, Y. G., Woo, E. (2016). Consumer acceptance of a quick response (QR) code for the food traceability system: Application of an extended technology acceptance model (TAM). Food Research International, 85, 266–272. Kjeldsen, K., Nodeland, M., Fagerstrøm, A., Pawar, S. (2023). The relative impact of QR codes on omnichannel customer experience and purchase intention. Procedia Computer Science 219, 1049–1056. Klein, L. (2003). Creating virtual product experiences: The role of telepresence Journal of Interactive Marketing, 17(1), 41–55. Lemon, K.N., Verhoef, P.C. (2016). Understanding customer experience throughout the customer journey. Journal of Marketing, 80(6), 69–96. Makkonen, M., Nyrhinen, J., Frank, L., Karjaluoto, H. (2022). The Effects of General and Mobile Online Shopping Skilfulness and Multichannel Self-Efficacy on Consumer Showrooming Behaviour . In A. Pucihar, M. Kljajić Borštnar, R. Bons, A. Sheombar, G. Ongena, & D. Vidmar (Eds.), 35th Bled eConference : Digital Restructuring and Human (Re)action, 479–494. University of Maribor. Myers, M. D., Newman, M. (2007). The qualitative interview in IS research: Examining the craft. Information and Organization, 17(1), 2–26. Nogieć, J. (2017). Sposoby postrzegania i wykorzystywania fotokodów przez klientów indywidualnych. Marketing i Zarządzanie, 47(1), 281–290. Novak, T. P., Hoffman, D. L., Yung, Y. F. (2000). Measuring the Customer Experience in Online Environments: A Structural Modeling Approach. Marketing Science 19(1), 22–42. Okazaki, S., Li, H., Hirose, M. (2012). Benchmarking the Use of QR Code in Mobile Promotion. Journal of Advertising Research, 52(1), 102–117. Paananen, T., Frank, L. & Kemppainen, T. (2022a). Customer-brand relationships in the context of digital brands. In A. Pucihar, M. Kljajić Borštnar, R. Bons, A. Sheombar, G. Ongena, & D. M. Holkkola, T. Paananen, L. Frank, T. Kemppainen, M. Makkonen: How Do QR Codes Enhance Customer Experience? Omnichannel Customer Experiences in a Brick-and-Mortar Fashion Store 333 Vidmar (Eds.), 35th Bled eConference : Digital Restructuring and Human (Re)action (pp. 495-510). Maribor: University of Maribor Press. Paananen, T., Kemppainen, T., Frank, L., Holkkola, M., Mali, E. (2022b). Reinforcement of brand relationships in an omnichannel environment: A qualitative study on clothing shopping. In Proceedings of the 14th Mediterranean Conference on Information Systems (MCIS). Parise, S., Guinan, P., Kafka, R. (2016). Solving the crisis of immediacy: How digital technology can transform the customer experience. Business Horizons 59(4), 411–420. Rigby, D. (2011). The Future of Shopping. Harvard Business Review, 89(12), 64–75. Rotsios, K., Konstantoglou, A., Folinas, D., Fotiadis, T., Hatzithomas, L., Boutsouki, C. (2022). Evaluating the Use of QR Codes on Food Products. Sustainability (Basel, Switzerland), 14(8), 4437. Scanova Blog. (2023). QR Code Statistics 2023: Up-To-Date Numbers On Global QR Code Usage Retrieved March 8, 2023 from https://scanova.io/blog/qr-code-statistics/#: :text=Also%2C%20according%20to%20a%20recent,9.76%20Million%20scans%20in%2020 18 Schmitt, B.H. (1999). Experiential Marketing. Journal of Marketing Management 15(1/3), 53–67. Sheth, J. (2021) Future of brick and mortar retailing: how will it survive and thrive? Journal of Strategic Marketing, 29(7), 598–607. Spaid, B., O’Neil , B., Ow, T. (2019). The upside of showrooming: How online information creates positive spill-over for the brick-and-mortar retailer. Journal of Organizational Computing and Electronic Commerce, 29(4), 294–315. Srinounpan, B., Srinounpan, C., Sumethokul, P., Patwary, A. K. (2020). The Application of QR Code Technology to Create the Value-Added Products for The Baan Klong Peek Neur Beehive Community Enterprise Groupat Tambon Suankhan, Nakhon Si Thammarat Province. Systematic Reviews in Pharmacy, 11(7), 519–528. Verhoef, P. C., Lemon, K. N., Parasuraman, A., Roggeveen, A., Tsiros, M., Schlesinger, L. A. (2009). Customer Experience Creation: Determinants, Dynamics and Management Strategies. Journal of Retailing 85(1), 31–41. Yoon, N., Kim, E. Y. (2014). An Exploratory Study of QR Code Utilization for Retailers' Multichannel Strategy. Fashion & Textile Research Journal, 16 (5), 730–744. Zimmermann, R., Mora, D., Cirqueira, D., Helfert, M., Bezbradica, M., Werth, D., Weitzl, W.J., Riedl, R., Auinger, A. (2022). Enhancing brick-and-mortar store shopping experience with an augmented reality shopping assistant application using personalized recommendations and explainable artificial intelligence. Journal of Research in Interactive Marketing. 334 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Appendix 1: Information on the participants Interview Gender Age Status duration (min) P1 Female 37 Employee 43 P2 Female 30 Student, employee 53 P3 Female 23 Student 64 P4 Female 40 Employee 84 P5 Female 29 Employee 47 P6 Male 29 Student, employee 65 P7 Male 24 Student, employee 52 P8 Female 33 Employee 58 P9 Female 42 Employee 49 P10 Female 25 Student 40 P11 Male 38 Employee 69 P12 Female 33 Employee 70 P13 Female 24 Student 51 P14 Female 32 Employee 65 P15 Female 29 Employee 64 Appendix 2: QR codes on display in the B&M store WHAT PREVENTS CONSUMERS FROM MAKING RESPONSIBLE ONLINE PURCHASES? TIINA KEMPPAINEN,1 MARKUS MAKKONEN,2 LAURI FRANK2 1 University of Jyväskylä, School of Business and Economics, Jyväskylä, Finland tikemppa@jyu.fi 2 University of Jyväskylä, Faculty of Information Technology, Jyväskylä, Finland markus.v.makkonen@jyu.fi, lauri.frank@jyu.fi The objective of this qualitative study is to examine and describe what prevents consumers from making responsible online purchases. The data were col ected during February and March 2023 with an online survey. The data comprises the respondents' (N=245) free-form written responses, in which they articulated and explained the factors that prevent their responsible online purchases. The findings indicate that external and internal factors can prevent consumers' responsible online purchases. While Keywords: external factors – online stores – are mainly blamed for responsible promoting irresponsible buying behavior, internal factors – online purchasing, consumers' individual characteristics – are also recognized to sustainability, have an important role in irresponsible purchasing behavior. The online external factors identified in this study include 1) pricing, 2) shopping, e-commerce, information, and 3) availability. The internal factors include 1) Bled self-indulgence and 2) trust. eConference DOI https://doi.org/10.18690/um.fov.4.2023.21 ISBN 978-961-286-751-5 336 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Online purchasing has increased significantly during the last decade. This growth can largely be attributed to the increased use of the internet and benefits that ecommerce provides to consumers, such as the ability to search for low prices, access an extensive product range, and the convenience of shopping from home. Online purchasing has also been encouraged by the COVID-19 pandemic. Due to the restrictions during the pandemic, many consumers discovered novel ways of utilizing online retailers and service providers. This shift in consumer behavior has led to a boom in e-commerce, with many businesses having to rapidly adapt to meet the increased demand for online shopping. While online purchasing has become increasingly vital and provides various advantages for consumers, it has also led to unfavorable ecological outcomes such as increased waste, carbon emissions, and energy consumption. The ecological implications of e-ecommerce have been extensively discussed, with adverse effects arising from packaging materials, delivery methods, unsold products, and product returns (Tiwari & Singh, 2011), to name a few. In addition, studies show that the significance of environmental considerations in consumer shopping choices is on the rise (e.g., De Canio et al., 2021). As a result, in recent years, the heightened environmental concerns and competitive pressure have spurred a greater emphasis on sustainability issues by both researchers and practitioners in the e-commerce industry. However, despite the discussion surrounding the environmental and ethical impacts of online purchasing, there remains a lack of comprehensive understanding of the topic. Studies have only partially investigated sustainable e-commerce, focusing on retailer-related issues such as packaging materials, logistic management aimed at reducing carbon emissions, and minimizing adverse environmental effects (Zhang, 2023). There is a particular need for further investigation on the consumer perspectives. Although there is a prevailing trend towards pro-green attitudes, it is common for consumers to opt for non-green alternatives when making purchases (e.g., Park & Lin, 2020; Su et al., 2022). Thus, studies should investigate what hinders responsible shopping and the ways in which retailers can encourage their customers to adopt more responsible behaviors (Wiese et al., 2015). T. Kemppainen, M. Makkonen, L. Frank: What Prevents Consumers from Making Responsible Online Purchases? 337 The objective of this qualitative study is to examine and describe what prevents consumers from making responsible online purchases. By comprehending these factors, it becomes possible to better understand the attitude-behavior gap identified in responsible consumption (e.g., Park & Lin, 2020; Su et al., 2022). The findings can be utilized to pinpoint the specific areas where consumers require assistance in making more responsible purchasing decisions and where online businesses should take measures to encourage sustainable shopping practices. Next, Section 2 discusses previous studies on responsible online purchasing. Section 3 describes the methods used for data collection and analysis. Section 4 presents the results of the study, and Section 5 discusses the contributions of the study, its implications for management, and recommendations for future research. 2 Responsible online purchasing and its obstacles The topic of responsible purchasing has been explored using various concepts, with sustainability and responsibility being among the prevalent ones. Sustainability and responsibility are interdependent concepts that reinforce and complement each other. Sustainability refers to the ability to maintain or preserve resources for future generations. Sustainability encompasses a range of issues, including environmental protection, social equity, and economic viability. (Armstrong et al., 2019) Responsibility, on the other hand, refers to the obligation or duty to act in a way that is ethical and accountable. It involves being aware of the impact of one's actions on others and the environment and taking steps to minimize harm. Responsible behavior is often necessary to achieve sustainability goals, such as reducing waste, conserving resources, and mitigating environmental impacts. The concept of responsible purchasing is centered around engaging in activities that aim to minimize the impact of purchased goods or services on the environment (Follows & Jobber, 2000), and opting for social y and ethical y responsible purchases (Jain et al., 2022). Previous research has predominantly explored responsible consumption in a general sense, rather than delving into the specifics of purchasing channels (online/offline). The purchase of responsible products is influenced by factors such as engagement in sustainable consumption, the degree of environmental and health consciousness, social influences (Carter et al., 2021), and the reputation of a company's environmental performance (Grimmer & Bingham, 2013). As evidenced by studies, 338 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY several obstacles can prevent consumers from making responsible purchasing decisions. It has been shown, for example, that low availability hinders the consumption of sustainable products and social pressure increases purchase intentions (Vermeir & Verbeke, 2006). Responsible consumption is barriered by consumers’ materialism, and thus may be facilitated by an increase in consumers’ environmental concern (Kilbourne & Pickett, 2008). Sustainable consumption decisions are directly hindered by a lack of opportunity, and indirectly by lack of motivation to consume sustainably (Tong et al., 2023). Choosing to purchase responsible products often involves immediate costs for individuals, such as financial expenses or the effort required to change their behavior (Demarque et al., 2015). Despite holding pro-environmental attitudes, even environmentally conscious consumers often face a trade-off between sustainability and other attributes such as price, quality, and performance. As a result, non-environmental y friendly alternatives are chosen. (Olson, 2013) In prior research, only a few studies have concentrated on responsible online shopping, with most of these studies examining it only in quite limited product or service contexts, such as fashion retailing (e.g., Kemppainen et al., 2021, 2022). In contrast, considerably more studies have been done on sustainable online shopping. For example, Yang et al. (2018) examined the adoption of sustainable online shopping in the context of the China’s Double-11 shopping festival and found sustainable online shopping intention to be positively affected by the attitude, subjective norm, and perceived behavioral control related to sustainable online shopping and negatively affected by the atmospheric factors related to the shopping festival itself. In contrast, Song et al. (2020) found the lack of policy support, insufficient knowledge of sustainable consumption, and the lack of awareness of sustainable consumption to act as the main barriers to the adoption of sustainable online shopping, whereas Schumacher et al. (2022) found technostress to act as an additional barrier by weakening the link between the general preference for sustainable products and the actual choice of sustainable products in an online shopping situation. In turn, Kanay et al. (2021) examined the potential of goal setting to promote sustainable online shopping, finding that both the goal setting and the feedback concerning the carbon footprint of a shopping basket together promote more sustainable online consumption. Finally, Demarque et al. (2015), Antonides and Welvaarts (2020), Berger et al. (2020), Gossen et al. (2022), Hollaus and Schantl (2022), and Michels et al. (2022) all examined the potential of different kinds of T. Kemppainen, M. Makkonen, L. Frank: What Prevents Consumers from Making Responsible Online Purchases? 339 nudging techniques (cf. Thaler & Sunstein, 2008; Lehner et al., 2016) to promote sustainable online shopping, finding many of them to be highly effective for this purpose. 3 Data col ection and analysis The data for this study were col ected during February and March 2023 with an online survey that was conducted by using the LimeSurvey service. The survey respondents were recruited by promoting the survey on social media and via the communication channels of Finnish universities and student associations. To promote the response rate, al the respondents who completed the survey were able to participate in a prize drawing of ten gift boxes worth about 25 € each. The survey questionnaire consisted of three sections, of which the first section concentrated on the general background information of the respondents, the second section on their general online shopping behaviour, and the third section more specifical y on their responsible online shopping behaviour. This last section also contained the three open-ended questions concerning responsible online shopping behaviour. Answers to the question n:o 3 “What kind of factors promote you to be or prohibit you from being responsible when making online purchases? Why?” were utilized as the data of this study. Respondents who provided written responses that clearly identified a factor hindering their responsible purchasing were eligible for inclusion in the final analysis. In total, 245 respondents provided explanations regarding the prohibiting factors. Their average response time for the whole survey was about 20 minutes. The sample statistics in terms of the gender, age, yearly taxable income, socioeconomic status, and average online shopping frequency of the respondents are reported in Table 1. As can be seen, most of the respondents were women, students, and relatively young. Their age varied between 19 and 66 years, with a mean of 27.8 years and a standard deviation of 8.1 years. However, most of the respondents were relatively active online shoppers who shopped online at least monthly on average. Al of them also had at least some experience in online shopping. 340 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 1: Sample statistics (N = 245) N % N % Gender Socioeconomic status Man 44 18.0 Student 190 77.6 Woman 189 77.1 Employee or self- employed 53 21.6 Other 12 4.9 Unemployed or unable to work 6 2.4 Age Pensioner 1 0.4 Under 25 years 105 42.9 Other 2 0.8 25–49 years 133 54.3 Online shopping frequency 50 years or over 7 2.9 At least weekly 10 4.1 Yearly personal taxable income At least monthly 141 57.6 Under 15,000 € 160 65.3 At least yearly 88 35.9 15,000–29,999 € 33 13.5 Less frequently than yearly 5 2.0 30,000 € or over 39 15.9 Has never shopped online 0 0.0 No response 13 5.3 No response 1 0.4 The data analysis was conducted using NVivo 12 Pro qualitative analysis software. First, factors that represented the chal enges of responsible online purchasing were extracted from each respondent and coded based on their content. Second, after identifying the initial codes describing the obstacles, they were grouped and label ed based on the common themes identified within the codes. Consequently, the study identified five primary themes that describe what prevents consumers from making responsible online purchases. 4 Findings The findings indicate that external and internal factors can prevent consumers' responsible online purchases. While external factors – online stores – are mainly blamed for promoting irresponsible buying behavior, internal factors – consumers' individual characteristics – are also recognized to have an important role in irresponsible purchasing behavior. The external factors identified in this study include 1) pricing, 2) information, and 3) availability. The internal factors include 1) self-indulgence and 2) trust. The following sections wil discuss these themes and their corresponding content. T. Kemppainen, M. Makkonen, L. Frank: What Prevents Consumers from Making Responsible Online Purchases? 341 4.1 External online store related factors Pricing. The primary obstacle to responsible online purchases was the pricing of responsible products (157 references). The respondents perceived responsible products to be costly, or more expensive than their conventional counterparts. Some participants noted that domestic products and brands were particularly expensive compared to their foreign counterparts. The experience of the price difference between responsible and irresponsible products discouraged respondents from choosing the responsible option: the perceived financial sacrifice was often considered too high. The difference in price is noteworthy particularly when the perceived quality of the responsible product does not outweigh the drawbacks of choosing a less responsible alternative. Conversely, if the difference in price is perceived as negligible, the responsible option is chosen. If the responsible choice is only slightly more expensive than the other options, I'm usually willing to pay for it. – Female, 23 The respondents frequently cited their financial situation as a reason for price-sensitivity, indicating that they are not able to afford responsible alternatives, even if they desire to purchase them. Opting for responsible alternatives was seen as difficult or unfeasible for those facing financial constraints and low incomes. When struggling to make ends meet, prioritizing responsibility over essentials becomes chal enging or even impossible, as the following quote demonstrates. It would be great to buy organic and organical y produced/certified products, but at this income level, you choose the cheapest or don't buy at all. – Female, 46 Despite facing obstacles, a number of respondents asserted that they were making their best effort to be responsible, and believed that they would be more inclined to do so if they had more financial means. Purchasing used items, such as those found at online flea markets, was seen as a practical way to make responsible choices. Nevertheless, some respondents also acknowledged that their stinginess contributed negatively to responsible purchasing; they incline to prioritize low prices over responsibility, opting for less sustainable alternatives to save money. 342 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Information. The lack of information regarding responsible practices posed an important hindrance to making responsible online purchases (81 references). Obtaining information on the responsibility of companies or their products was considered a challenging task, with inadequate transparency and quality of available information. According to the respondents, detecting the product's origin is complicated. Online stores do not provide sufficient details about the raw materials used in the products, their origin, the parties and methods involved in the production process, and the manufacturing country. Consequently, it is difficult to determine responsibility; there is no means to verify if unethical practices such as child labor were involved in the production, for example. Despite having good intentions, one may engage in irresponsible purchasing because of lack of information and knowledge. Reliable information about the entire value chain of the product is not available. […] The ecological, social and ethical responsibility of the product cannot be traced. – Nonbinary person, 48 Information about the manufacturing process of the products is not easily available, so an unethical product may be bought by mistake. – Female, 19 There was an expectation for online retailers to offer greater transparency throughout the entire customer journey, encompassing not only the source of the products but also post-purchase events. Specifical y, there was a desire for more comprehensive information from these retailers regarding the fate of returned products, including whether they are discarded as waste or resold. Almost al online stores have very poor transparency, e.g., about what happens if you return something. Do they, for example, go straight to the trash? – Female, 22 Additional y, the topic of used products was raised, with a recognition of their inherent responsibility in recycling, but also acknowledging that information on these items is often even more chal enging to obtain than for new products. T. Kemppainen, M. Makkonen, L. Frank: What Prevents Consumers from Making Responsible Online Purchases? 343 Responsibility can be achieved by facilitating the online search for used products, ensuring their appeal to customers, and presenting relevant product information in a manner consistent with that of new items. – Female, 23 Due to the insufficient information available, the buyer needs to make even more assumptions regarding the responsibility of a product when dealing with used items. Availability. The availability of responsible products (74 references) was found to be inferior to that of non-sustainable options. In some product categories, it is possible that responsible alternatives may not be available at all, which limits the consumer's ability to make sustainable choices. Furthermore, even when responsible options are available, the selection may be inadequate in terms of product characteristics such as size and design. Foreign online retailers may offer a better selection, but this comes at the cost of longer transportation distances. Therefore, buyers must consider whether their ultimate choice is more responsible, weighing the environmental impact of delivery against their preference for sustainable products. Narrow selection. For example, I'm looking for a certain product that needs to be ordered from abroad, which potential y increases emissions. – Female, 21 On the other hand, it was noted that buying from foreign stores can also be difficult, because not all stores deliver products to foreign countries, such as Finland. 4.2 Internal customer related factors Self-indulgence. The respondents also emphasized self-indulgence (52 references) – the role of their own consumer attributes and the desire for convenience in terms of the shopping process and the product that is purchased. A purchasing process that is quick and uncomplicated is preferred, and products are expected to possess certain features that may not be available in responsible alternatives. There are many online stores that are very irresponsible. So a large selection and an easy shopping experience can be tempting at times. – Female, 22 344 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY One of the reasons for irresponsible choices was the fast-paced nature of daily life. It was observed that addressing sustainability concerns is time-consuming and often impractical, especial y when purchases need to be made quickly. Several respondents also admitted to being lazy and neglecting their responsibility due to this. While responsible shopping is acknowledged as an important concept, it often takes a backseat to other pressing concerns in everyday life. The preventing factors are probably hurry and laziness. – Female, 28 Consequently, many respondents reported having inadequate knowledge about responsibility and emphasized the need to start with the basics. In addition, some respondents revealed that they are impulsive and easily swayed by attractive deals, making shopping a source of pleasure rather than a responsibility issue. Trust. Issues of trust (48 references) were also highlighted as a barrier to responsible purchases. Numerous respondents expressed skepticism towards the notion of sustainability, suspecting that companies engage in greenwashing when communicating their efforts to be responsible. This is due to a lack of evidence to support their claims, or the provision of insufficient evidence. There was a perception among many respondents that online stores engage in deceptive practices, misleading or cheating consumers. It was believed that online stores are withholding information about products, misleading consumers with inaccurate information, using certificates without authorization, and leveraging claims of responsibility as marketing tools. The consumer is being cheated. Many products are advertised as green and responsible, and there are many certificates that ultimately do not guarantee anything. – Female, 27 Many respondents expressed difficulties in distinguishing genuine responsibility claims made by companies. To address this problem, responsible and familiar stores were prioritized. By shopping at stores that exclusively offer responsible products or if the store has been previously self-evaluated as responsible, there is no need to evaluate each product individually. T. Kemppainen, M. Makkonen, L. Frank: What Prevents Consumers from Making Responsible Online Purchases? 345 Trust-related factors were found to be relevant also in the context of online C2C (consumer to consumer) trade and flea markets. Even though flea markets inherently promote responsible behavior by facilitating recycling, there may be apprehension and unease associated with making purchases. When consumers engage in transactions with each other, the absence of easily accessible aid can create challenging situations. Moreover, assessing the quality of second-hand products, particularly when shopping online, can pose a chal enge because the product cannot be physical y experienced before the purchase. 5 Discussion This study's findings enhance comprehension of the factors that contribute to the discrepancy between attitudes and actions (e.g, Park & Lin, 2020; Su et al., 2022) regarding responsible online purchasing. The study demonstrates that external factors related to online store and internal factors related to consumers themselves can prevent responsible online purchases. According to this study, consumers tend to attribute their irresponsible spending mostly to online stores: pricing of responsible products, information about responsibility issues and availability of responsible products. In addition, consumers recognize their own responsibility and the influence of personal factors on practicing responsible behavior while shopping online. As previous research has indicated (e.g., Demarque et al., 2015), opting for responsible choices often entails trade-offs and various costs for consumers. According to this study, consumers tend to perceive such trade-offs and costs particularly in terms of price. The study highlighted that pricing and perceptions related to pricing are the primary factors that hinder responsible online shopping. Consumers tend to view responsible products as costly or more expensive than other products, which limits their willingness to purchase them. Previous research has also emphasized the crucial role of price in responsible consumption (e.g., Kemppainen et al., 2021). Given that pricing is (still) a significant obstacle to responsible behavior, further investigation is necessary to understand why consumers perceive responsible products as expensive and what can be done to the negative price perceptions. Future studies should delve deeper into the (lower) pricing of responsible products and explore ways to enhance consumers' price perception and motivation to purchase responsible options. 346 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY In line with previous studies (e.g., Song et al., 2020; Kemppainen et al., 2021), the findings underscore the significance of the responsibility information provided by online stores. Online retailers do not provide comprehensive information about the social and environmental impacts of their products. Lack of transparency and information about products' sourcing, production processes and their impact makes responsible shopping chal enging. The responsibility information – and the inadequacy of it – is also linked to another obstacle identified in this stydy: consumers' trust. According to the findings, consumers exhibit a degree of skepticism and suspect that companies engage in greenwashing. Such perceptions hinder responsible buying and may negatively impact online stores and other companies that genuinely prioritize responsibility themes. Future research should examine how issues related to responsibility could be communicated more effectively to buyers. Clarifying the communication of responsible practices could enhance consumers' understanding of responsible products and promote their purchasing behavior. Another critical aspect to explore is how to build consumer trust in the information provided. Consumers are becoming increasingly aware of the ethical and environmental implications of their purchases and are seeking more information to make informed choices. By understanding these issues, online services can be developed to promote responsible consumption and better cater to consumers' needs in this regard. Previous studies have noted that practices such as feedback concerning the carbon footprint of a shopping basket (Kanay et al. 2021) can promote responsible online purchasing. Hence, future studies should delve deeper into other solutions and online store characteristics that could help solving the wel -identifies information gap. The findings also indicate that the availability of responsible products is an obvious chal enge to shop responsibly. Consistent with these results, prior research has demonstrated that sustainable consumption choices are impeded by limited opportunities (Tong et al., 2023) and inadequate availability (Vermeir & Verbeke, 2006). As responsible product availability has been recognized as a challenge for quite some time, it prompts a question of why it continues to present a major hurdle. It is worth exploring the reasons behind the perceived inadequacy of the supply and identifying strategies to make responsible alternatives more accessible to consumers. T. Kemppainen, M. Makkonen, L. Frank: What Prevents Consumers from Making Responsible Online Purchases? 347 Finaly, the study underscored the influence of consumer attributes and convenience-seeking behavior – self-indulgence – on responsible online purchases. Amidst the business of daily life, responsible products or the associated purchasing procedures may not seem appealing enough. The act of responsible online purchasing is often viewed as unpleasant, excessively challenging and requiring significant amounts of time – the tradeoff is perceived as excessively large (Olson, 2013). Hence, sustainable consumption decisions are hindered by lack of motivation to consume sustainably (Tong et al., 2023). Future research should delve deeper into ways to impact the convenience of responsible shopping, such as exploring how online store design and characteristics can be utilized to enhance the ease of responsible shopping. To summarize, this study identified factors that prevent responsible online purchasing based on survey responses. The identified obstacles are largely in line with previous research related to responsible consumption in different (online/offline) contexts. It is therefore worth asking, why these obstacles have persisted in a similar fashion, year after year. The obstacles of responsible online shopping should be investigated more comprehensively and through various methodologies to gain a deeper understanding and solutions regarding these issues. References Antonides, G., & Welvaarts, M. (2020). Effects of default option and lateral presentation on consumer choice of the sustainable option in an online choice task. Sustainability, 12(13), 5484. Armstrong, G., Kotler, P., & Opresnik, M. (2019). Marketing: An Introduction (14th ed.). Pearson Education Limited. Berger, M., Müller, C., & Nüske, N. (2020). Digital nudging in online grocery stores – Towards ecological y sustainable nutrition. In Proceedings of the 41st International Conference on Information Systems. Atlanta, GA: AIS. Carter, K., Jayachandran, S., & Murdock, M. R. (2021). Building A Sustainable Shelf: The Role of Firm Sustainability Reputation. Journal of Retailing, 97(4), 507–522. De Canio, F., Martinelli, E., & Endrighi, E. (2021). Enhancing consumers' pro-environmental purchase intentions: the moderating role of environmental concern. International Journal of Retail & Distribution Management, 49(9), 1312-1329. Demarque, C., Charalambides, L., Hilton, D. J., & Waroquier, L. (2015). Nudging sustainable consumption: The use of descriptive norms to promote a minority behavior in a realistic online shopping environment. Journal of Environmental Psychology, 43, 166–174. Fol ows, S. B., & Jobber, D. (2000). Environmental y responsible purchase behaviour: A test of a consumer model. European Journal of Marketing, 34(5/6), 723–746. Gossen, M., Jäger, S., Hoffmann, M. L., Bießmann, F., Korenke, R., & Santarius, T. (2022). Nudging sustainable consumption: A large-scale data analysis of sustainability labels for fashion in German online retail. Frontiers in Sustainability, 3, 922984. 348 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Grimmer, M., & Bingham, T. (2013). Company environmental performance and consumer purchase intentions. Journal of Business Research, 66(10), 1945–1953. Hollaus, M., & Schantl, J. (2022). Incentivizing consumers towards a more sustainable online shopping behavior: A study on nudging strategies in B2C e-commerce. In Proceedings of the 13th International Conference on E-Education, E-Business, E-Management, and E-Learning (pp. 372–376). New York, NY: ACM. Jain, V. K., Dahiya, A., Tyagi, V., & Sharma, P. (2022). Development and validation of scale to measure responsible consumption. Asia-Pacific Journal of Business Administration, Ahead-of-print. Kanay, A., Hilton, D., Charalambides, L., Corrégé, J. B., Inaudi, E., Waroquier, L., & Cézéra, S. (2021). Making the carbon basket count: Goal setting promotes sustainable consumption in a simulated online supermarket. Journal of Economic Psychology, 83, 102348. Kemppainen, T., Frank, L., & Luhtanen, V. (2022). What is meaningful for responsible shoppers in online fashion retail? In Proceedings of the 8th International Workshop on Socio-Technical Perspective in IS Development. Kemppainen, T., Frank, L., Makkonen, M., & Hyvönen, O.-I. (2021). Barriers to responsible consumption in e-commerce: Evidence from fashion shoppers. In A. Pucihar, M. Kljajić Borštnar, R. Bons, H. Cripps, A. Sheombar & D. Vidmar (Eds.), Proceedings of the 34th Bled eConference (pp. 323–335). Maribor, Slovenia: University of Maribor Press. Kilbourne, W. E., & Pickett, G. M. (2008). How materialism affects environmental beliefs, concern, and environmental y responsible behavior. Journal of Business Research, 61(9), 885-893. Lehner, M., Mont, O., & Heiskanen, E. (2016). Nudging – A promising tool for sustainable consumption behaviour? Journal of Cleaner Production, 134(A), 166–177. Michels, L., Ochmann, J., Günther, S. A., Laumer, S., & Tiefenbeck, V. (2022). Empowering consumers to make environmental y sustainable online shopping decisions: A digital nudging approach. In Proceedings of the 55th Hawai International Conference on System Sciences (pp. 4707– 4716). Honolulu, HI: University of Hawai’i at Mānoa. Olson, E. L. (2013). It’s not easy being green: The effects of attribute tradeoffs on green product preference and choice. Journal of the Academy of Marketing Science, 41(2), 171–184. Park, H. J., & Lin, L. M. (2020). Exploring attitude–behavior gap in sustainable consumption: Comparison of recycled and upcycled fashion products. Journal of Business Research, 117, 623–628. Schumacher, K., Peters, L., & Feste, J. (2022). Please mind the stress: The influence of technostress on mindset-driven sustainable consumption in an online shopping context. In Proceedings of the 55th Hawai International Conference on System Sciences (pp. 4378–4387). Honolulu, HI: University of Hawai’i at Mānoa. Song, W., Zhu, Y., & Zhao, Q. (2020). Analyzing barriers for adopting sustainable online consumption: A rough hierarchical DEMATEL method. Computers & Industrial Engineering, 140, 106279. Su, M., Fang, M., Kim, J., & Park, K. S. (2022). Sustainable marketing innovation and consumption: Evidence from cold chain food online retail. Journal of Cleaner Production, 340, 130806. Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving Decisions About Health, Wealth, and Happiness. New Haven, CT: Yale University Press. Tiwari, S., & Singh, P. (2011). E-Commerce: Prospect or Threat for Environment. International Journal of Environmental Science and Development, 211–217. Tong, L., Toppinen, A., Lei, W., & Berghäl , S. (2023). How motivation, opportunity, and ability impact sustainable consumption behaviour of fresh berry products. Journal of Cleaner Production, 401, 136698. Vermeir, I., & Verbeke, W. (2006). Sustainable food consumption: Exploring the consumer “attitude-behavioral intention” gap. Journal of Agricultural and Environmental Ethics, 19(2), 169-194. Yang, S., Li, L., & Zhang, J. (2018). Understanding consumers’ sustainable consumption intention at China’s Double-11 online shopping festival: An extended theory of planned behavior model. Sustainability, 10(6), 1801. Zhang, M. (2023). Sustainability Transitions in E-commerce Research—Academic Achievements and Impediments. Circular Economy and Sustainability, 1-22. ETHICAL CONSIDERATIONS OF AUGMENTED REALITY IN HIGH-TECH MANUFACTURING SANDER VAN DER HOEK,1 MARLIES VAN STEENBERGEN,2 PASCAL RAVESTEIJN3 1 ASML, Eindhoven, Netherland sander.van.der.hoek@asml.com 2 HU University of Applied Sciences Utrecht, Research Group Digital Ethics, Utrecht, Netherland marlies.vansteenbergen@hu.nl 3 HU University of Applied Sciences Utrecht, Research Group Process Innovation & Information Systems, Utrecht, Netherland pascal.ravesteijn@hu.nl The use of Augmented Reality (AR) in industry is growing rapidly, driven by benefits such as efficiency gains and ability to Keywords: augmented overcome physical boundaries. Existing studies stress the need reality, to take stakeholder values into account in the design process. In mixed this study the impact of AR on stakeholders' values is reality, OHMD, investigated by conducting focus groups and interviews, using human value sensitive design as a framework. Significant impacts were values, found on the values of safety, accuracy, privacy, helpfulness and value sensitive autonomy. Twenty practical design choices to mitigate potential design, negative impact emerged from the study. industry 4.0 DOI https://doi.org/10.18690/um.fov.4.2023.22 ISBN 978-961-286-751-5 350 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Driven by the travel restrictions caused by the Covid-19 outbreak, industrial companies that need to perform complex maintenance tasks sought out possibilities for overcoming the issue of not being able to have their engineers on site. These companies began to quickly adopt augmented reality technology that enabled senior engineers to instruct less experienced staff on site1.The vast number of professional and academic publications that mention Covid-19 and Augmented Reality (AR) keywords indicates that this not only happened in the industrial sector but also in other sectors such as healthcare. The use of AR might provide an effective solution for overcoming travel restrictions (Lamberti et al., 2014), but an important question is whether companies are ready to use these technologies in an effective and sustainable way and whether they understand the possible implications. Chal enges and success factors for implementing AR are of technological, organizational and/or environmental type (Egger & Masood, 2020). Masood & Egger (2019) show that though academic research focuses more on the technological aspects, professionals are more concerned with the organizational consequences. All but one of the success factors professionals labeled as most relevant were user-centric: user acceptance, visibility of information, ergonomics, and usability of the user interface. Other research shows that, if human values are not incorporated effectively enough, companies may not be able to leverage al of AR’s potential (Burleigh et al., 2020; Hofmann et al., 2017; Rousi, 2016). This study addresses ethical aspects of using AR technology in an Industry 4.0 context. The main research question for this study is: How can the application of AR technology in the high-tech semiconductor industry take human values into account to positively influence the impact on its stakeholders? Examples of stakeholders are the users of AR glasses, but also their managers, group leads and customers. These stakeholders al have different stakes in regard to the use of AR technology. The key concepts investigated in this study are stakeholders and their values. These concepts are studied within the context of specific AR use cases in the semi-1 See for instance https://www.smartindustry.com/tools-of-transformation/augmented- reality/article/11414929/how-covid-19-boosted-augmented-reality-and-virtual-reality-in-manufacturing). S. van der Hoek, M. van Steenbergen, P. Ravesteijn: Ethical Considerations of Augmented Reality in High-Tech Manufacturing 351 conductor industry, to stimulate discussion of values in more detail. AR technology might impact the personal values of stakeholders in a positive or negative way. Based on insights into these impacts on stakeholder values, designs can be adapted to foster the positive impact and limit the negative impact, resulting in a design that is in balance with the values of the different stakeholders. In section 2 we discuss the theoretical background to our study, fol owed by a description of our research method in section 3. Section 4 presents the results, which are discussed in section 5. The paper ends with conclusions in section 6. 2 Theoretical background Within the high-tech industry the concepts of industry 4.0 and 5.0 are often mentioned. Industry 4.0 represents the 4th industrial revolution with high levels of mechanization, automation, digitalization and miniaturization (Lasi et al., 2014). While in the smart manufacturing paradigm of Industry 4.0 the role of human workers has become increasingly smaller, Industry 5.0 is bringing human workers back into the picture (Longo et al., 2020). The 5th Industrial Revolution combines human and machine, creating synergy between the two (Nahavandi, 2019). It is expected that the use of AR technology wil significantly increase with Industry 5.0 concepts as new technical improvements are introduced (Fernández del Amo et al., 2018). Technical obstacles seem to become less inhibitory in large-scale usage and AR appears to be effective in both transferring knowledge to users and discovering and capturing knowledge from users (Fernández del Amo et al., 2018). Future factories will combine different technologies such as Big Data, AI and AR (Shi et al., 2020), enabling individualized human-machine interaction (Xu, 2021). Though Industry 5.0 is purported to be human-centric instead of technology-driven (Xu, 2021), Longo et al. (2020) raise ethical concerns as the new technologies of the factories of the future are expected to have great impact on human values. They argue that more attention must be paid to the means by which technologies in Industry 5.0 systems can be designed for human values rather than treating these values as a side issue. This is the more relevant when the boundaries between human and technological capabilities are blurred, as is the case with AR technology. Regarding AR outside the context of Industry 4.0, Pase (2012) argues that AR is a persuasive technology, having the potential to intentional y influence or modify 352 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY behaviors, values or attitudes, and as such raises ethical concerns. The main impacted values he discusses are manipulation of people, privacy and physical safety. Slater et al. (2020) explains the concept of physical vs. psychological realism. They discuss ethical issues that may arise when it becomes more difficult to differentiate the real from the virtual world. A third example of research regarding ethical considerations in AR is by Hofmann et al. (2017), who found a variety of ethical issues related to privacy, safety, justice, change in human agency, accountability, responsibility, social interaction, power and ideology. 3 Research Method This study uses the Value Sensitive Design (VSD) approach. VSD is a theoretical y based approach that takes human values into account in technical design. In VSD human value is defined as “What is important to people in their lives, with a focus on ethics and morality” (Friedman & Hendry, 2019, p. 24). VSD distinguishes between three types of investigations into stakeholder values: conceptual (general knowledge from literature or experts), empirical (eliciting actual values and norms from stakeholders) and technical (translating values into technical affordances). These investigations are intended to be iterative, feeding each other and allowing the designer to continuously modify the design (Friedman & Hendry, 2019). In this study, first, values that might be impacted by technology design were distilled from existing academic literature (conceptual investigation). Next, these values were validated by a focus group (Bell et al., 2019) existing of four AR specialists who were involved with AR initiatives in the same high tech industry company: 1) a business leader 2) a business process specialist and expert in AR 3) a business solutions architect and 4) a systems architect. The focus group also identified two relevant scenarios of AR technology use in high tech industry as wel as the stakeholders that might be impacted in these use cases. The scenarios and value list were input for semi-structured interviews with representatives from the stakeholder groups that were identified (empirical investigation). These stakeholders were part of the same company but worked mainly at customer sites (other companies) or were associated with them. In total, 13 employees with different roles (engineers, project leads, site managers and team leads) and work locations in Northern America, Europe and Asia were interviewed. S. van der Hoek, M. van Steenbergen, P. Ravesteijn: Ethical Considerations of Augmented Reality in High-Tech Manufacturing 353 During the interviews the participants were asked how they are involved in and affected by this technology, what values they think are most important in this context, how these values are impacted and how they think the impact can be influenced. The two scenarios from the focus group were discussed. The first consisted of Guided Work Instructions (GWI), where an engineer wears AR glasses to view step-by-step work instructions for a specific task in the form of text, holograms, audio and video. The second scenario was GWI+ where GWI is extended with more interaction between engineer and technology, because AI can “observe” the environment and work of the engineer and interact more with the engineer based on these observations. Interviewees were also asked to rank human values in order of importance to them in the context of the two given scenarios. The data gathering phase of this study took place in the summer and fal of 2021. A thematic analysis was performed across the interviews to identify themes and search for patterns involving repetitions, similarities and differences. This was accomplished by using the code groups: harms, benefits, constraints, do’s, don’ts, enablers, values, value tensions and mitigating measures. During the coding, at first the inductive method was used to code al content, while the grouping of codes used a more deductive approach (Linneberg, 2019). Relations between items and code groups were identified. Benefits, harms, do’s and don’ts were linked to values in cases where there was a clear and strong relationship. Two kinds of relationships were used: relationships that indicate a positive contribution to the value and relationships that indicate a negative impact on the value. Creating this structure al owed further analyses of the raw interview data and helped find patterns and draw conclusions. Subsequently, a formula was used to calculate the overall value ranking based on all interviews. One point was given to a value if it was selected in third place, 2 points if it was rated as second, and 3 points were given to values chosen as most important. These scores were added up, leading to a ranking with overal scores per value. In addition to the scores, two other variables are used to compare the outcome of the value ranking: groundedness and density. Groundedness refers to the number of quotations associated with a code, and density refers to the number of codes linked with a value. 354 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY After the interviews a second focus group session was conducted with the aim of validating and complementing the mitigation measures that emerged from the interviews (technological investigation). 4 Results In the interviews the fol owing values emerged in descending order of priority: safety, accuracy, privacy, helpfulness, autonomy, care, comfort, self-control, trust, independence, power, efficiency, security, accountability, pride, social interaction, cleanliness, professionalism, health, knowledge, quality, accessibility, change in human agency, reliability, effectiveness, ergonomics, responsibility as displayed in figure 1 (note that the scales of the four dimensions scenario 1, scenario 2, groundedness and density are different; dimensions can thus be compared between values but not within one and the same value). Figure 1: Value Ranking S. van der Hoek, M. van Steenbergen, P. Ravesteijn: Ethical Considerations of Augmented Reality in High-Tech Manufacturing 355 The five values with highest priority (in italic above) are discussed in the following paragraphs, as wel as the mitigating measures (numbered below) that were suggested by the participants to prevent potential negative impact on these values. Many times, the phrase “safety first” was mentioned. Stakeholders identified several issues regarding safety. The wearer of an AR device is more vulnerable to falling, hitting something with their head or arms or when controlling a heavy moving object. Persons who work with lasers were cautious about open-beam hazards since a safety glass was not integrated in the AR headset. Safety sensors on the device that warn about potential safety hazards were seen as a significant improvement (1). Incorporation of safety glasses that protect the eyes from laser beams (2) was also suggested as a mitigating measure. The focus group indicated that safety regulations (3) and work instructions accompanied by safety tips (4) were the most feasible and effective measures to improve safety. Improvements in the viewing angle (5) were also mentioned. A narrator function could help prevent text objects from appearing in the field of sight (6). Advanced object recognition and safety hazard detection (7) are further possible future solutions. Accuracy was the second most important value mentioned. The use of AR technology potentially enables the participants to do their jobs more accurately with fewer errors, and requiring less interpretation compared to the existing method of working. Accuracy can be considered an industry value in the semiconductor industry. Performing their jobs more accurately was important to the participants, but the accuracy of the AR technology itself is also important. Several measures for improvement were given such as: more accurate work instructions (8), and higher resolutions and less screen jitter (9). More accurate control of the device (10) was also often mentioned. Privacy became an important value in the second scenario that contained more intel igent interaction between user and technology, which requires an AI function that observes the environment as wel as the person. In this scenario participants became more sensitive about their privacy, and they began to raise concerns about being evaluated using AR and AI technology. Some mentioned that people behave differently when someone is wearing an AR device because others might think they are being recorded. The most significant measures to ensure privacy were to adhere to privacy laws (11), anonymize personal data (12), prevent storage of personal 356 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY information (13) and not share the data with management (14). Privacy-ensuring measures such as blurring faces (15) and transparent privacy controls such as a blinking light when the device is recording (16) were also mentioned by participants. Helpfulness was rated as the fourth important value for the stakeholders; the interviewees stated that if AR was not helpful to them, they were not going to use it. AR technology is considered a potential y helpful technology, but if it is not helpful on the spot, there is no time to make it so. Others mentioned that they need more help in using the tool to make it more helpful to them. The focus group suggested getting context-aware help in the form of projected suggestions (17). Additionally, reducing the administrative burden for engineers was proposed through intel igently assessing the components used during the assembly or maintenance task (18) and making the license portal accessible via the headset (19). It became clear during the interviews that autonomy is a two-sided topic; autonomy is influenced in both positive and negative ways according to the interviewed stakeholders. They indicated a positive influence because an engineer can do more work autonomously as the information to do the task is at hand in a clear and easy to understand way. On the negative side stakeholders indicated that AR can also take away autonomy. Work instructions can become more prescriptive and rigid, leaving less room for the creativity of the engineer. The engineer becomes more dependent on the technology itself to do the job. Stakeholders expressed that if the technology, specifical y in the form of work instructions, is too strict, workers would start feeling like robots or puppets, making them feel less appreciated as professionals. There appears to be a tension between working as an independent professional, and ensuring efficiency and quality by providing uniform detailed instruction; we wil return to this topic later in this section. The focus group suggested giving the engineers more control and providing them with the space to make decisions as part of the work instructions by applying a state-based design principle (20). Besides the individual values, the VSD concepts of value dams (features that some stakeholders strongly object to), value flows (features that many stakeholders support) and value tensions (values contradicting each other) were identified. The most significant value dam mentioned during the interviews involved security, more specifical y the lack of intel ectual property (IP) protection in the current design. Value dams are explained as design options to which even a smal percentage of S. van der Hoek, M. van Steenbergen, P. Ravesteijn: Ethical Considerations of Augmented Reality in High-Tech Manufacturing 357 stakeholders strongly object (Friedman & Hendry, 2019). For companies in the semiconductor industry, protecting their IP is a major concern. IP is often considered one of the most valuable assets of a company working in this domain. Companies in the semiconductor industry often do not al ow devices from suppliers that they do not control in their cleanroom as these can potential y be used to observe and steal IP-sensitive information. Thus, the technology is not easily accessible to the end user because lengthy approval processes and security controls build a metaphorical dam that hinders fostering other values such as autonomy, efficiency and helpfulness. A solution to overcome this value dam mentioned by the focus group was the capability to blur specific objects based on object recognition or virtual curtains placed on specific coordinates. A value tension was identified between efficiency and autonomy or independence. Efficiency can be reached in the area of doing work faster by reducing execution, learning and travelling time, making fewer mistakes and having the right information at hand, leading to higher productivity. This higher productivity can be attained by building AR technology that provides more consistent work instructions, understands the context and how the activities are performed by the user, and can even correct users. Potentially, this could lead to more prescriptive and rigid ways of instructing users. Increasing autonomy means providing individuals with more freedom to determine their working procedures and task scheduling (Niessen & Volmer, 2010). This tension shows the importance of finding a proper balance between AR technology and its capabilities to instruct engineers and the level of freedom given to the engineers. Especial y as autonomy is a critical aspect of both justice and wellbeing (Calvo et al., 2020). Lu et al. (2022) describes this phenomenon as empathic machines or empathic robot control where empathy skil s generate human-robot shared actions based on understanding of the human state and situational circumstances. 5 Discussion Comparing the values mentioned by the participants with existing research about Industry 4.0/5.0 and AR values, it seems that the values mentioned are more related to the AR technology itself than to the context of the semi-conductor industry. However, context does influence some of the values in terms of prioritization. Safety, privacy, comfort and accountability seem to be tightly related to the 358 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY technology itself. Accuracy, care, helpfulness, and autonomy seem to have a stronger relation to the semiconductor industry. As mentioned in the result section safety stood out as the most significant value. This relates to research by Lu et al. (2022) which positions Safety as a ground level value in their Industrial Human Needs Pyramid (Lu, 2022). The first aspect that must be ensured in a manufacturing environment is safety which can also be concluded from our study in the area of AR technology usages. Once safety is ensured other values become more important. During the interviews two scenarios (GWI and GWI+) were described to the interviewees. In the second scenario other values emerged from the interviews. The values autonomy and privacy were considered more important in Scenario 2. The stakeholders worried that if artificially intel igent technology becomes too prescriptive, they may lose their autonomy. AR technology opens new opportunities to measure the performance of employees by using cameras and sensors. Performance monitoring of employees has been a topic of study for many years. Hawk (1994) investigated the effects of computerized performance monitoring, concluding that it can result in health problems, stress and the perception of unfairness. He urges organizations to be selective in which tasks are to be measured by computerized monitors. In 2017 an empirical study on the effects of employee surveillance showed that it leads to reduced sense of privacy and self-esteem and increased uncertainty, employee vulnerability and changes in behavior (Indiparambil, 2017). There seems to be a very thin line between using observation technology and AI for the good of the employees and using it in such a way that individuals change their behaviors for the bad, resulting in undesired behaviors and lack of intentional benefits. Although the value of trust was not selected in the top five values during the interviews, it was mentioned quite a few times that positive outcomes on certain values such as helpfulness, privacy, autonomy, and security will build trust in AR technology, which leads to more usage and broader adoption. Since experience with AR is limited, and the technology is immature in many ways, building trust in the technology is important (Mcknight et al., 2011). Humanity’s trust in technology has always been somewhat problematic, however (Taddeo, 2010). Considering the overal importance of trust in technology, the growing maturity level of AR, the S. van der Hoek, M. van Steenbergen, P. Ravesteijn: Ethical Considerations of Augmented Reality in High-Tech Manufacturing 359 potential negative outcomes and value dams and the high level of human-computer interaction (HCI) that this technology brings (Dünser et al., 2007), trust should play an important role in the further design and development of this technology. 6 Conclusion The stakeholders of AR can be impacted by this technology in a positive but also in a negative way. This study revealed that the values stakeholders are most concerned with are safety, accuracy, privacy, helpfulness and autonomy. Twenty appropriate measures that can be embedded in the design to reduce potential negative impact on these values, were formulated during the value discussions. In addition to these five values, the importance of building trust in AR technology frequently emerged. The most significant value tension that this study disclosed was between autonomy and efficiency. Important design questions need to be answered for AR applications that provide work instructions. How much freedom will an engineer have in using GWI and will they be able to develop a personal way of working? Will GWI take existing experience into account? Wil it adapt the instructions based on the level of details an engineer needs? How wil mistakes or faulty executions of the work instructions by the engineer be reported? A limitation to this study is that is performed in the context of only one company in the semiconductor industry. However, although the persons interviewed were employees of this company, they often executed their work as engineer at other major semiconductor companies. Therefore, they were able to some extent to share perspectives from those companies. Situational factors such as local culture might influence stakeholder values and how they perceive the impact that AR has on those values. Situational factors are taken into account in this study by considering multiple stakeholders and stakeholders from different regions in the world. Interviews were held with stakeholders from three different regions (US, EU, APAC) to represent the whole population of the stakeholder group. Since the interviewees from the US, Taiwan, Japan, Korea and the Netherlands represented the global footprint of the semiconductor industry (Semiconductor Industry Association, 2020) this research provides reliable results on a global level by ruling out local influences. 360 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY The use of AR technology in the semiconductor industry is a novelty. Some interviewees had more extensive experience with AR, but overal , the experience was limited. As a minimum requirement the stakeholders selected for the interviews needed to have experimental experience with AR. Ideal y, when using VSD, al participants can experience the technology extensively, including new design improvements. That was not the case in this research. Envisioned scenarios were presented that required some imagination and interpretation by the participants. We recommend to continue using VSD to validate future development. Valuable future research could be performed in developing working instructions for AR devices. There is stil a great deal to explore in developing personalized, effective work instructions. Many work instructions consist of step-by-step guides that do not take the existing skills, learning curves and personal preferences of the users into account. Since AR technology wil significantly drive human computer interaction, and users need to work more and more in symbiosis with the technology, this area will be an important area for future research. We argue that the further development of AR technology needs to profoundly take human values into account. AR technology wil further mature over time, experience new iterative design cycles, and new applications wil be combined with other technologies like AI. Embedding a value-sensitive design approach may prevent harm and balance value tensions in future designs. Acknowledgements We would like to thank al participants in the focus groups and interviews for their time and valuable insights. References Bel, E., Bryman, A., & Harley, B. (2019). Business research methods (Fifth edition ed.). Oxford University Press. Burleigh, C., Kroposki, M., Magabo, M., & Bailey, L. (2020). Ethical Considerations in Designing Virtual and Augmented Reality Products--Virtual and Augmented Reality Design with Students in Mind: Designers' Perceptions. Journal of Educational Technology Systems, 49(2), 219-238. https://hu.on.worldcat.org/oclc/8686785393 Calvo, R. A., Peters, D., Vold, K., & Ryan, R. M. (2020). Supporting human autonomy in AI systems: A framework for ethical enquiry. Ethics of digital well-being: A multidisciplinary approach, 31-54. S. van der Hoek, M. van Steenbergen, P. Ravesteijn: Ethical Considerations of Augmented Reality in High-Tech Manufacturing 361 Dünser, A., Grasset, R., Seichter, H., & Bilinghurst, M. (2007). Applying HCI principles to AR systems design. Egger, J., & Masood, T. (2020). Augmented reality in support of intelligent manufacturing - A systematic literature review. Computers & Industrial Engineering, 14010.1016/j.cie.2019.106195 Fernández del Amo, I., Erkoyuncu, J. A., Roy, R., Palmarini, R., & Onoufriou, D. (2018). A systematic review of Augmented Reality content-related techniques for knowledge transfer in maintenance applications. Computers in Industry, 103, 47-71. 10.1016/j.compind.2018.08.007 Friedman, B., & Hendry, D. G. (2019). value sensitive design. MIT Press. https://doi.org/10.7551/mitpress/7585.001.0001 Hawk, S. R. (1994). The effects of computerized performance monitoring: An ethical perspective. Journal of Business Ethics, 13(12), 949-957. Hofmann, B., Haustein, D., & Landeweerd, L. (2017). Smart-Glasses: Exposing and Elucidating the Ethical Issues. Science and Engineering Ethics, 23(3), 701-721. 10.1007/s11948-016-9792-z Indiparambil, J. J. (2017). An empirical study on the detrimental effects of employee surveil ance in India. International Journal of Research in Computer Application & Management, 7(12), 48-51. Lamberti, F., Manuri, F., Sanna, A., Paravati, G., Pezzol a, P., & Montuschi, P. (2014). Chal enges, opportunities, and future trends of emerging techniques for augmented reality-based maintenance. IEEE Transactions on Emerging Topics in Computing, 2(4), 411-421. Lasi, H. D., Fettke, P. P. D., Kemper, H. P. D., Feld, T. D., & Hoffmann, M. D. (2014). Industrie 4.0. Wirtschaftsinformatik, 56(4), 261-264. 10.1007/s11576-014-0424-4 Linneberg, M. S., & Korsgaard, S. (2019). Coding qualitative data: A synthesis guiding the novice. Qualitative research journal, 19(3), 259-270. Longo, F., Padovano, A., & Umbrel o, S. (2020). Value-Oriented and Ethical Technology Engineering in Industry 5.0: A Human-Centric Perspective for the Design of the Factory of the Future. Applied Sciences, 10(12), 4182. 10.3390/app10124182 Lu, Y., Zheng, H., Chand, S., Xia, W., Liu, Z., Xu, X., Wang, L., Qin, Z. & Bao, J. (2022). Outlook on human-centric manufacturing towards Industry 5.0. Journal of Manufacturing Systems, 62, 612-627. Masood, T., & Egger, J. (2019). Augmented reality in support of Industry 4.0—Implementation chal enges and success factors. Robotics and Computer Integrated Manufacturing, 58, 181-195. 10.1016/j.rcim.2019.02.003 Mcknight, D. H., Carter, M., Thatcher, J. B., & Clay, P. F. (2011). Trust in a specific technology: An investigation of its components and measures. ACM Transactions on Management Information Systems (TMIS), 2(2), 1-25. Nahavandi S. (2019). Industry 5.0 - A Human-Centric Solution. Sustainability, 11(16) Niessen, C., & Volmer, J. (2010). Adaptation to increased work autonomy: The role of task reflection. European Journal of Work and Organizational Psychology, 19(4), 442-460. Pase, S. (2012). Ethical considerations in augmented reality applications. Paper presented at the Proceedings of the International Conference on E-Learning, E-Business, Enterprise Information Systems, and E-Government (EEE), 1. Rousi, R. (2016). Using human-values as a guide for understanding worthy design directions in augmented reality. AcademicMindtrek 2016 - Proceedings of the 20th International Academic Mindtrek Conference, , 243-252. 10.1145/2994310.2994322 Semiconductor Industry Association. (2020). 2020 STATE OF THE U.S. SEMICONDUCTOR INDUSTRY. https://www.semiconductors.org/wp-content/uploads/2020/06/2020-SIA- State-of-the-Industry-Report.pdf Shi, Z., Xie, Y., Xue, W., Chen, Y., Fu, L., & Xu, X. (2020). Smart factory in Industry 4.0. Systems Research and Behavioral Science, 37(4), 607-617. 10.1002/sres.2704 Slater, M. (2020). The Ethics of Realism in Virtual and Augmented Reality. Frontiers in Virtual Reality, 1 362 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Taddeo, M. (2010). Trust in Technology: A Distinctive and a Problematic Relation. Knowledge, Technology Policy : Knowledge, Techology Policy (Untill 2011), 23(3-4), 283-286. Xu, X. (2021). Industry 4.0 and Industry 5.0 Inception, conception and perception. Journal of Manufacturing Systems, 61, 530-535. THE EFFECT OF E-LEADERSHIP ON DIGITAL TRANSFORMATION IN THE DUTCH PUBLIC SECTOR PASCAL RAVESTEIJN,1 TON CAMU,2 PAUL MORSCH,1 GUIDO ONGENA1 1 HU University of Applied Sciences, Research Group Process Innovation & Information Systems, Utrecht, Netherland pascal.ravesteijn@hu.nl, paul.morsch@hu.nl, guido.ongena@hu.nl 2 HU University of Applied Sciences, Master of Informatics, Utrecht, Netherland ton.camu@icloud.com A public sector that adequately makes use of information technology can provide improved government services that not only stimulates business development it also intensifies citizen participation and economic growth. However, the effectiveness of IT and its governance at both national as wel as on municipality level leaves much to be desired. It is often stated that this is due to a lack of digital skil s needed to manage the IT function and alignment with business. Therefore, the aim of this study is to determine the effect that digital leadership Keywords: competences and IT capabilities have on digital transformation public readiness within Dutch municipalities. Based on an analyses of sector, municipalities, survey data from 178 respondents we recommend municipalities digital to implement a range of activities that all are related to realize the leadership, ability to constantly apply strategic thinking and organizational IT capabilities, leadership to exploit the capability of Information Technology digital to improve the business. transformation DOI https://doi.org/10.18690/um.fov.4.2023.23 ISBN 978-961-286-751-5 364 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The change toward a public sector that adequately makes use of the advantages that information technology has to offer to optimize processes and services has been going on for many years. Dunleavy at al. (2006) already stated that a shift is needed toward ‘‘digital-era governance’ that is realized by ' reintegrating functions into the governmental sphere, adopting holistic and needs-oriented structures, and progressing digitalization of administrative processes' (Dunleavy et al., 2006). In recent years, the Dutch government has been investing more effort in digital transformation to improve its services that are vital towards society (Blacquière et al., 2021). Several studies show that improved government services through digitization not only stimulate business development it also intensifies citizen participation and economic growth (Alvarenga et al., 2020, Alnuaimi et al., 2022). Other benefits that are often mentioned are: better transparency and accountability, improved access to government data, increased support for innovation, a more responsive supply chain, and support for environmental initiatives (Alnuaimi et al., 2022). However, the effectiveness of IT and its governance at both the national as wel as on municipality level leaves much to be desired (Elias et al., 2015). Within the public sector implementation of management of information systems causes issues regarding coordination between organisational units and achieving organisational goals (Broadbent et al., 1989). The IT function of an organisation needs to be managed and aligned between business and technology focused IT departments as it is regularly the missing link between the often artificial boundaries within government organisations (Njanka et al., 2021). However, for successful digital transformation, it is also necessary for managers and employees to have the required digital competences. Ravesteyn & Ongena (2019) stated that e-leadership competences derived from the European e-competence framework (e-CF) have a positive significant effect on digital transformation. Furthermore, Nwankpa & Roumani (2016) found that IT capabilities have a positive effect on digital transformation. Based on this the objective of the study described in this paper is to determine the effect that digital leadership competences and IT capabilities have on digital transformation readiness within Dutch municipalities. The remainder of this paper is structured as follows, after an overview of the theoretical background in the next section we describe the research approach (section 3) and the results of our analyses in section 4. Practical recommendations P. Ravesteijn, T. Camu, P. Morsch, G. Ongena: The Ef ect of e-Leadership on Digital Transformation in the Dutch Public Sector 365 that can be implemented by municipalities are discussed in section 5 and the paper ends with conclusions in section 6. 2 Theoretical background and conceptual model Digital transformation in municipalities is not just about digitising paper-based work processes but about a change of methods and communication that will improve the efficiency and service quality of the organization by improving transparency, lead times and reducing the phenomenon of being ‘pushed from pil ar to post' (Layne & Lee, 2001; Lindgren et al., 2019). Digital transformation involves the use of digital capabilities and technologies to influence different aspects of the organisation to create value. It is important to understand specifical y how different types of digital technologies, in combination with capability, influence certain aspects of the organisation (Morakanyane et al., 2017). In addition, digital transformation is an evolutionary process whereby digital technologies and skil s are deployed to add value to business models, operational processes and to customer experiences (Morakanyane et al., 2017). Due to a lack of shared strategic vision, empowerment, and innovation the strategic contribution of the IT-function to its business counterpart is less effective within the ecosystems of municipalities (Elias et al., 2015). IT is not merely instrumental in cost reduction, but it is a business enabler that adds value to the organisation (Njanka et al., 2021). Working towards digital transformation there is a shortage of administrative skills, data availability, a lack of resources, a lack of technological capabilities (Alnuaimi et al, 2022) and a lack of competences (Elias et al., 2015; Hüsing, 2013). In a survey among Dutch public administrations, Tangi et al. (2021) provide insight into their transformation efforts and find that the public bodies that were studies only act on exogenous input and that there is no sense of urgency other than sufficient external pressure. The European Union recognised the lack in IT related competences and therefore in the Malmö (European Union, 2009) and Tallinn (European Union, 2017) declarations 32 participating countries unanimously adopted an e-Governance initiative to develop digital leadership skills for civil servants at all levels. However, there is no common understanding on what entails digital leadership. Klein (2020) defined characteristics of digital leadership catagorized into three groups: digital business, social attitude, and general mindset. Though, none of the listed characteristics seem to be related to technical skil s. In contrast, McCarthy, Sammon 366 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY & Alhassan (2021) identified eight digital transformation leadership characteristics that contain both business oriented (digital strategist, digital culturalist, customer centrist, organisational agilist, business process optimizer, and digital workplace landscaper) as wel as more technical (digital architect, and data advocatte) characteristics. Looking at the European e-Competence Framework standard (e-CF, 2019; NEN, 2018) we find 30 generic roles and 41 competences that are defined which contribute to uniform development and common principles for competence development for ICT professionals. Ravesteyn and Ongena (2019) used the e-CF competences of the Digital Transformational Leader Role to examine the effect on IT Capabilities and Digital Transformation readiness within organizations. These competences, with business and technical orientation, are Business Plan Development (A3), Architecture Design (A5), Innovating (A9), Business Change Management (E7), Information Systems Governance (E9)1. Given the fact that the e-CF standard is adopted by Dutch government we have selected this role and corresponding competences for this study. That IT Capabilities have a positive effect on digital transformation readiness is shown in earlier research by Lu and Ramamurthy (2011), and Nwankpa and Roumani (2016). Lu and Ramamurthy (2011) conceptualise IT capabilities as a latent construct that is reflected in three dimensions. First, IT infrastructure; the ability of an organisation to deploy IT-based managed data services, its architecture and network services in addition to managing the application portfolio and services delivered. Second, IT business spanning; the ability of an organisation's management to leverage IT infrastructure in support of business objectives (Lu & Ramamurthy, 2011; Mao et al., 2015). And finally, proactive stance, the ability of an organisation to actively explore ways to embrace IT innovations that contribute to its business objectives (Lu and Ramamurthy, 2011). As earlier studies have found positive relations but didn’t explicitly do research in the context of government organizations this study focuses on municipalities to help them better prepare for digital transformation. Derived from the literature our conceptual model (figure 1) has three reflective exogenous latent constructs related to IT Capabilities: 1) IT Proactive Stance, 2) IT Infrastructure, and 3) IT Business Spanning. Furthermore, digital leadership consists 1 The code between brackets refers to the competence label in the e-CF framework. P. Ravesteijn, T. Camu, P. Morsch, G. Ongena: The Ef ect of e-Leadership on Digital Transformation in the Dutch Public Sector 367 of five formative exogenous latent constructs: 1) Innovating, 2) Architecture Design, 3) Business Plan Development, 4) Business Change Management, and 5) Information Systems Governance. Finally, the endogenous construct is Digital Transformation. Figure 1: Conceptual Model As the objective of our research is to determine the effect that digital leadership competences and IT capabilities have on digital transformation readiness within Dutch municipalities the following hypotheses were formulated in line with the conceptual model: H1: Digital leadership competences have a positive effect on digital transformation. H2: IT capabilities have a positive effect on an organization’s digital transformation. 3 Research Method To investigate the influence of the e-leader towards the relationship between IT capabilities and digital transformation a quantitative research method was used that emphasises the quantification of the data collection and provides for analysis at the 368 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY ordinal level. In the operationalization of this research, the existing e-competence framework has been used for col ecting data in Dutch municipalities. The questionnaire, sent to over 1800 civil servants, contained al the constructs of the conceptual model which are measured using multiple items. As there were seven respondents with too many missing values and one that scored 1 on each question, the number of usable respondents after cleaning the dataset was 178. Respondents worked across different sizes of municipalities based on the number of employed persons (table 1). The period of data col ection was between February 14th, 2022, and May 30th, 2022. Table 1: Municipality size Sample (N=178) Proportion (%) Municipality size (number of employees) Less than 250 42 23.6% 250-999 90 50.6% 1000-4999 35 19.63% Above 5000 11 6.17% For this study, both business stakeholders and ICT stakeholders were invited to participate. Unfortunately, the municipal government has no uniformity regarding job titles, distribution of work and mandate (De Tuya et al., 2020). Therefore, to determine the target group, the most common job roles that are responsible and/or bear co-responsibility for digital transformation-related subjects were sought in the immediate vicinity of the researchers. From an inventory within 8 municipal organisations, 18 roles were frequently encountered. Participation requests were therefore sent to the following job roles: IT director, ICT manager, manager I&A, IV manager, CIO, alderman IT/IS, functional manager, technical manager, Information consultant, ICT director, Service Level Manager, CISO, Tiso (technical information security officer), data protection officer, municipal secretary, IT director, transition manager. The initial low response rate was partly mitigated by requesting some of the respondents to forward the request for participation within their own organisation towards the target group. This means that a request for snowbal sampling was also used for approximately 25% of the invitations (Baarda et al., 2021). P. Ravesteijn, T. Camu, P. Morsch, G. Ongena: The Ef ect of e-Leadership on Digital Transformation in the Dutch Public Sector 369 In regard to the survey (containing 47 questions in total), questions on digital transformation and IT capabilities are taken from Nwankpa & Roumani (2016) who derived them from different studies: 4 on Infrastructure (Bharadwaj et al., 2000; Ross et al. 1996; Weill et al., 2002), 4 on Business Spanning (Bharadwaj et al., 2000; Mata et al., 1995), 4 on Proactive Stance (Lu & Ramamurthy, 2011; Weill et al., 2002), and 9 related to Digital Transformation (Aral & Weill, 2007). Questions (21) regarding Digital Leadership competencies have been reused from Ravesteyn and Ongena (2019). Besides this five general questions were posed to determine size of the organization (#employees and population), funtion of the respondent, and worklevel (strategic, tactic, operational). The questions were presented via a 7-point Likert scale (Taherdoost, 2019). Across the entire data set (after cleaning) there were 60 missing values and as there was no more than 5% missing in any individual question we opted for the option ‘mean replacement’ during analysis in SmartPLS (Hair et al., 2021). 4 Results 4.1 Evaluation of Measurement Model Given that it places less of a burden on the measurements and normal distribution (Gefen, et al., 2000; Hair, et al., 2011), partial least squares (PLS) is employed to evaluate the model (Chin, 1998). To evaluate the statistical significance of the loadings and the path coefficients, a bootstrap approach was performed (Hair et al., 2021). A non-parametric method of estimation cal ed ‘bootstrapping’ involves resampling the original data with replacement in order to estimate each parameter in the PLS model (Chin, 2001). Prior to evaluating the structural model and testing the hypotheses, the measurement model is assessed to determine its reliability and validity using the software SmartPLS (Ringle et al., 2015). 4.1.1 Reflective Constructs As shown in Table 2, Cronbach’s alpha, Composite reliability and Average variance extracted (AVE) show that the internal consistency and convergent validity requirements are met. Also, the outer loadings are between 0.785 and 9.41 so meet the rule of thumb >0.7 hence al items can be retained. 370 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 2: Construct reliability and validity Average Composite Composite variance Cronbach’s reliability reliability extracted alpha (rho_a) (rho_c) (AVE) Digital Transformation 0.942 0.943 0.963 0.897 IT Business Spanning 0.917 0.922 0.942 0.802 IT Infrastructure 0.831 0.838 0.887 0.662 IT Proactive Stance 0.919 0.921 0.943 0.805 The Fornel -Larcker criteria to assess discriminant validity is met (Fornell & Larcker, 1981) and this is also true for the assessment of cross-loadings. Finally, we used the heterotraitmonotrait ratio (HTMT) to accurately assess discriminant validity as there is discussion about whether the first two methods are effective in empirical applications (Franke & Sarstedt, 2019; Henseler, Ringle, & Sarstedt, 2015). Based on Bootstrapping with 10,000 samples and a 0.05 significance level (one-sided) table 3 shows that discriminant validity is established as HTMT ratios are less than 0.90 and even below the more conservative threshold of 0.85 (Henseler, Ringle, & Sarstedt, 2015) therefore we conclude that the reflective measures are valid and can be used as a basis for further analysis. Table 3: Heterotrait-monotrait ratios BS INF PS DT BS Business Spanning INF Infrastructure 0.772 PS Proactive stance 0.684 0.654 DT Digital Transformation 0.553 0.584 0.744 P. Ravesteijn, T. Camu, P. Morsch, G. Ongena: The Ef ect of e-Leadership on Digital Transformation in the Dutch Public Sector 371 4.1.2 Formative Constructs To evaluate our formative measurement model (i.e. the constructs for digital leadership), we first test for possible collinearity issues. For this, we use the variance inflation factor (VIF) for which values of 5 or higher indicate significant collinearity issues (Hair, Risher, et al. 2019). We found that al VIF values are below the threshold of 5 therefore we conclude that there are no critical levels of col inearity. Subsequently, we used bootstrapping to test the significance and relevance of the outer weights of our model. All the weights present satisfactory significance levels except two. Items A5_2 of the Architecture Design construct and E7_2 of the Business Change Management construct are marked non-significant. However, when an indicator’s weight is not significant, but the corresponding item loading is relatively high (>=0.50), or statistically significant, the item can be retained (Hair, Hult, Ringle, & Sarstedt, 2021). Since the loadings of A5_2 and E7_2 are respectively 0.847 and 0.800 both with p-value 0.000 we decided not to remove these items from the model. 4.2 Evaluation of Structural Model We examined the significance and relevance of the structural model relationships. Here we not only discuss the significant direct effects but also any indirect effects on the endogenous construct of Digital Transformation (see figure 2) as the goal of our study is to advise municipalities on how to improve their readiness for digital transformation. 372 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Figure 2: Estimated relationships of structural model If we look at the relevance of the constructs, we find that from the digital leadership constructs Business Plan Development ( β = 0.295, t = 2.491, p < 0.013) is significant while for the IT capability constructs the IT Proactive Stance path ( β = 0.461, t = 5.263, p < 0.000) has a significant relation with Digital Transformation. However, since the IT capability constructs also act as a mediator on the key target variable Digital Transformation, we also need to look at the total effect. We then see that Business Plan Development ( β = 0.428, t = 3.412, p < 0.001) has a much larger total effect on digital transformation (see table 4). Furthermore, we find that there is also a significant indirect effect of the Digital Leadership competence Innovating on Digital Transformation via Proactive Stance ( β = 0.245, t = 2.183, p < 0.029). P. Ravesteijn, T. Camu, P. Morsch, G. Ongena: The Ef ect of e-Leadership on Digital Transformation in the Dutch Public Sector 373 Table 4: Total and Indirect Effects Original Standard sample (O) deviation t-value p-values Total Effects Businessplan dev.\DT 0.428 0.125 3.412 0.001 Innovating\IT Proactive Stance 0.245 0.112 2.183 0.029 Specific Indirect Effects Businessplan dev.\IT Proactive Stance\DT 0. 117 0.051 2.284 0.022 Innovating\IT Proactive Stance\DT 0.113 0.055 2.055 0.040 The explanatory power of our structural model (figure 2) is determined by examining the R 2 value of the endogenous latent variable Digital Transformation which with R 2=0.551 explains 55.1% of the variance and is considered moderate (Hair, Risher, Sarstedt, & Ringle., 2019). The effect sizes for Business Plan Development and IT Proactive Stance are respectively f2=0.064 and f2=0.218. R2 however only indicates the model’s in-sample explanatory power (Shmueli 2010). To assess the model’s out-of-sample predictive power. Shmueli et al. (2016) developed a holdout-sample-based procedure that generates case-level predictions on an item or a construct level to reap the benefits of predictive model assessment in PLS-SEM. To assess the predictive power of the model we executed a k-fold cross-validation with PLSpredict. The first step is to check whether Q2 values are above zero which indicates that the model outperforms the most naïve predicted benchmark (Shmueli, Sarstedt, et al. 2019). This is the case for all the indicators in the dataset used in our study. The second step then is to compare the root mean squared error (RMSE) against the naïve linear regression model (LM) benchmark. An increasingly higher number of indicators that yields lower prediction errors in terms of RMSE when comparing the PLS-SEM analysis to the naïve LM benchmark shows a higher predictive power (Shmueli, Sarstedt, et al. 2019). Concerning the 374 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY dependent variable Digital Transformation two PLS-SEM RMSE’s values are higher than LM RMSE’s (indicators DT_1 and DT_2). This indicates that the model has a low predictive power (Shmueli, Sarstedt, et al. 2019). Comparing RMSE’s with regard to IT Proactive Stance, IT Business Spanning and IT Infrastructure shows that all indicators yield a lower prediction error in the PLS-SEM analysis. This indicates that the model has a high predictive power (Shmueli, Sarstedt, et al. 2019). Based on these analyses we conclude that the overal model has high predictive power and can therefore be used as a foundation for our recommendations to municipalities. 5 Recommendations and Discussion To provide clear guidelines to municipalities that want to improve the way they use digital technologies to improve services we need to look at the underlying weights of the items in the constructs that we found to have a significant effect. For Business plan development we find that item A3_3 (0.481) scores highest compared to A3_1 (0.306) and A3_2 (0.355). Therefore, we recommend that municipalities should ‘ constantly apply strategic thinking and organizational leadership to exploit the capability of Information Technology to improve the business’ (A3_3). This confirms the attention for strategy found by McCarthy et al. (2021) who state: “prioritizing digital transformation as a strategic objective by influencing top management to put it top of their agenda” (p.10). Similarly, for Proactive Stance the item weights are respectively ITCPS_1 (0.280), ITCPS_2 (0.277), ITCPS_3 (0.261) and ITCPS_4 (0.297). Although the items are more evenly weighted, ITCPS_4 ‘ We constantly seek new ways to enhance the ef ectiveness of IT use’ is rated highest and should therefore be a process that municipalities internalize. Furthermore, the importance of a focus on innovation (ITCPS_1) ‘ To constantly keep current with new information technology innovations’ supports the notion by Klein (2020) who found that “the most distinguished leadership characteristic in the era of digital transformation [. ] is to be innovative visionary” (p.895). Another capability that municipalities should emphasize is to make sure to ‘ have people that are capable of and continue to experiment with new IT as necessary’ (ITCPS_2), which coincides with the “digital talent scout” characteristic mentioned by Klein (2020). Also important is the need to ‘ Create a climate that is supportive of trying out new ways of using IT’ (ITCPS_3), and this corresponds with the “digital culturalist” characteristic of P. Ravesteijn, T. Camu, P. Morsch, G. Ongena: The Ef ect of e-Leadership on Digital Transformation in the Dutch Public Sector 375 digital transformation leadership found by McCarthy et al. (2021). Lastly, we also found a significant indirect effect of the Innovating competence on Digital Transformation via Proactive Stance. Looking at the underlying items we find that A9_1 (0.599) has the highest score compared to A9_2 (0.463). Even though the Innovating competence seems to be less important in this study (compared to Klein (2020)) it is worth mentioning as the underlying items are clearly in support of A3_3 and ITCPS_4 as the related principles that municipalities should adopt are ‘ to constantly apply independent thinking and technology awareness to lead the integration of disparate concepts for the provision of unique solutions’ (A9_1), and ‘ to constantly chal enge the status quo and provide strategic leadership for the introduction of revolutionary concepts’ (A9_2). Finally, what we didn’t find in this study is an effect of the architecture design competence even though in the literature study by McCarthy et al. (2021) the “digital architect” characteristic was the second most found characteristic after “digital strategist”. 6 Conclusion With this study, we tried to determine the effect that digital leadership competences and IT capabilities have on digital transformation readiness within Dutch municipalities. We found that from the digital leadership competences only Business Plan Development had a direct effect on Digital Transformation while Innovating had an indirect effect via the IT Proactive Stance capability, which by itself also has a direct effect on Digital Transformation. Based on this we can state that the hypotheses we formulated are accepted for only some of the underlying constructs. While we advise municipalities to implement a range of activities that should enable them to ‘constantly apply strategic thinking and organizational leadership to exploit the capability of Information Technology to improve the business’, some final words of caution are necessary. Specifical y, the number of respondents from large municipalities was limited so the difference between organization sizes has not been analyzed and might go unnoticed. Furthermore, the sample of respondents contains employees with a wide range of roles, which we haven’t analyzed to determine if there are different views amongst specific groups that have traits in common. Similarly, cultural aspects might play a role during digital transformations so the outcomes in other countries might be different. Future studies are needed to provide 376 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY more contextual insights into how Digital Leadership competences, IT Capabilities and Digital Transformation interact. References Alnuaimi, B. K., Sing, K. S., Ren, S., Budhwar, P., & Vorobyev, D. (2022). Mastering digital transformation: The nexus between leadership, agility, and digital strategy. Journal of Business Research 145, 636-648. Alvarenga, A., Matos, F., Godina, R., & C. O. Matias, J. (2020). Digital Transformation and Knowledge Management in the Public Sector. Sustainability, 12(14), 5824. https://doi.org/10.3390/su12145824 Aral, S., & Weill, P. (2007). IT assets, organizational capabilities, and firm performance: How resource al ocations and organizational differences explain performance variation. Organization science, 18(5), 763-780. Baarda,B., Bakker, E., T.Fischer,T., Julsing & Vianen. (2021). Basisboek Methoden en Technieken Kwantitatief praktijkgericht onderzoek op wetenschappelijke basis: Vol. Zevende druk (Nummer Sociale methoden en technieken). Noordhoff Uitgevers BV. https://web-p-ebscohost- com.hu.idm.oclc.org/ehost/detail/detail?vid=0&sid=66738126-89fe-453e-9ffc-7d6aa815e62d%40redis&bdata=JnNpdGU9ZWhvc3QtbGl2ZSZzY29wZT1zaXRl# Bharadwaj, A.S. (2000). A Resource-Based Perspective on Information Technology Capability and Firm Performance: An Empirical Investigation. MIS Quarterly, 24(1), 169-196. https://doi.org/10.2307/3250983 Blacquière, R., de Groot, M., & Serruys, M. (2021). I-strategie Rijk 2021 – 2025: Doorpakken op digitale transformatie. https://www.digitaleoverheid.nl/document/i-strategie-rijk-2021-2025/ Broadbent, M., Hansel , A., Dampney, N.G., Gilmour, P., Hardy, G. (March 1989) Business Strategic Orientation, Information Systems Strategic Orientation, and Strategic Alignment. International Journal of Information Management Chin, W.W. (1998). The partial least squares approach to structural equation modeling. In Modern Methods for Business Research, edited by G.A Marcoulides, 1295–1336. Mahwah, N.J: Lawrence Erlbaum Associates. Chin, W.W. (2001). PLS-Graph user’s guide. CT Bauer College of Business, University of Houston, USA, 15, 1-16. De Tuya, M., Cook, M., Sutherland, M., & Luna-Reyes, L.F. (2020). The leading role of the government CIO at the local level: Strategic opportunities and chal enges. Government Information Quarterly, 37(3). https://doi.org/10.1016/j.giq.2017.01.002 Dunleavy, P., Margetts, H., Bastow, S., & Tinkler, J. (2006). New public management is dead: long live digital-era governance. Journal of Public Administration Research and Theory 16, 467-494 E-Competence Framework (e-CF) (2019). A common European Framework for ICT Professionals in al sectors—NEN- EN 16234-1:2019. European Union (2009). Ministerial Declaration on eGovernment (2009). Accessed on July 20, 2022: https://administracionelectronica.gob.es/pae_Home/pae_Estrategias/pae_lineas_ccoperaci on/pae_Cooperacion_Internacional/pae_estrategias_de_administracion_electronica/pae_A mbito_Europeo_- _Las_Declaraciones_Ministeriales.html?idioma=en#.YtfBZy-Qn0o European Union (2017). Tal inn Declaration on eGovernment at the Ministerial Meeting during Estonian Presidency of the Council of the EU (2017). Accessed on April 15, 2022. https://digital-strategy.ec.europa.eu/en/news/ministerial-declaration-egovernment-tallinn- declaration Elias, T., Ulenbelt, P., Fokke, M., Bruins Slot, H., & Van Meenen, P. (2015). Parlementair onderzoek naar ICT-projecten bij de overheid. ’s Gravenhage: Tweede Kamer. Accessed June 10th 2022: P. Ravesteijn, T. Camu, P. Morsch, G. Ongena: The Ef ect of e-Leadership on Digital Transformation in the Dutch Public Sector 377 https://www.tweedekamer.nl/sites/default/files/field_uploads/33326-5-Eindrapport_tcm181-239826.pdf Fornell, C., & Larcker, D.F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of marketing research, 18(1), 39-50. Franke, G., & Sarstedt, M. (2019). Heuristics versus statistics in discriminant validity testing: a comparison of four procedures. Internet Research. Gefen, D., Straub, D.W., & Boudreau, M. (2000). Structural equation modeling techniques and regression: Guidelines for research practice. Communications of AIS 4 (7): 1-77. Hair, J.F., Hult, G.T.M., Ringle, C.M., & Sarstedt, M. (2021). A primer on partial least squares structural equation modeling (PLS-SEM). Sage publications. Hair, J.F, Ringle, C.M, & Sarstedt, M. (2011). PLS-SEM: Indeed a silver bullet. Journal of Marketing Theory and Practice 19 (2): 139-152. Hair, J.F, Risher, J.J., Sarstedt, M., & Ringle, C.M. (2019). When to use and how to report the results of PLS-SEM. European Business Review 31(1), 2-24. Henseler, J., Ringle, C.M., & Sarstedt, M. (2015). A new criterion for assessing discriminant validity in variance-based structural equation modeling. Journal of the Academy of Marketing Science 43 (1): 115-135 Hüsing, T., Korte, W. B., Fonstad, N., Lanvin, B., Cattaneo, G., Kolding, M., Lifonti, R., & van Welsum, D. (2013). e- Leadership: E-Skills for Competitiveness and Innovation Vision, Roadmap and Foresight Scenarios Final Report. 180. Klein, M. (2020). Leadership Characteristics in the Era of Digital Transformation. Business & Management Studies: An International Journal, 8(1), 883–902. Layne, K., & Lee, J. (2001). Developing fully functional E-government: A four stage model. Government Information Quarterly, 18(2), 122–136. https://doi.org/10.1016/ S0740-624X(01)00066-1 Lindgren, I., Madsen, C.Ø., Hofmann, S., & Melin, U. (2019). Close encounters of the digital kind: a research agenda fort he digitalization of public services. Government Information Quarterly, 36, 427-436. Lu, Y., & Ramamurthy, K. (2011). Understanding the Link Between Information Technology Capability and Organizational Agility: An Empirical Examination. MIS Quarterly, 35(4), 931-954. https://doi.org/10.2307/41409967 Mao, H., Liu, S., Zhang, J., & Deng, Z. (2016). Information technology resource, knowledge management capability, and competitive advantage: The moderating role of resource commitment. International Journal of Information Management, 36(6), 1062–1074. https://doi.org/10.1016/j.ijinfomgt.2016.07.001 Mata, F. J., Fuerst, W. L., & Barney, J. B. (1995). Information technology and sustained competitive advantage: A resource-based analysis. MIS quarterly, 19(4), 487-505. McCarthy, P., Sammon, D., & Alhassan, I. (2021): Digital Transformation Leadership Characteristics: A Literature Analysis, Journal of Decision Systems, DOI: 10.1080/12460125.2021.1908934 Morakanyane, R., Grace, A., & O’Reilly, P. (2017). Conceptualizing Digital Transformation in Business Organizations: A Systematic Review of Literature. Digital Transformation – From Connecting Things to Transforming Our Lives, 427–443. https://doi.org/10.18690/978-961-286-043- 1.30 NEN (2018). European ICT professionals role profiles – Part 1 – CWA 16458-1. https://www.nen.nl/cwa-16458-1- 2018-en-250293 Njanka, S. Q., Sandula, G., & Colomo-Palacios, R. (2021). IT-Business Alignment: A Systematic Literature Review. Procedia Computer Science, 181, 333–340. https://doi.org/10.1016/j.procs.2021.01.154 Nwankpa, J., & Roumani, Y. (2016). IT Capability and Digital Transformation: A Firm Performance Perspective. ICIS 2016 Proceedings. https://aisel.aisnet.org/icis2016/ISStrategy/Presentations/4 378 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Ravesteyn, P., & Ongena, G. (2019). The Role of e-Leadership in Relation to IT Capabilities and Digital Transformation. 12th IADIS International Conference Information Systems 2019, 171–179. https://doi.org/10.33965/is2019_201905L022 Ringle, C.M, Wende, S., & Becker, J.-M. (2015). SmartPLS 3. Boenningstedt: SmartPLS GmbH. http://www.smartpls.com. Ross, J.W., Beath, C.M., & Goodhue, D. L. (1996). Develop long-term competitiveness through IT assets. MIT Sloan Management Review. Shmueli, G. (2010). To explain or to predict? Statistical Science 25 (3): 289-310. Shmueli, G., Hair, J.F., Sarstedt, M., Cheah, J.-H., Ting, H., Vaithilingam, S., & Ringle, C.M. (2019). Predictive model assessment in PLS-SEM: guidelines for using PLSpredict. European Journal of Marketing 53 (11): 2322-2347. Shmueli, G., Ray, S., Velasquez Estrada, J.M., & Shatla, S.B. (2016). The elephant in the room: Evaluating the predictive performance of PLS models. Journal of Business Research 69 (10): 4552-4564 Taherdoost, H. (2019). What Is the Best Response Scale for Survey and Questionnaire Design; Review of Different Lengths of Rating Scale / Attitude Scale / Likert Scale; International Journal of Academic Research in Management. Tangi, L., Janssen, M., Benedettie, M., & Noci Guiliano (2021). Digital government transformation: a structural equation modeling analysis of driving and impeding factors. International Journal of Information Management 60. https://doi.org/10.1016/j.ijinfomgt.2021.102356 Weill, P., Subramani, M., & Broadbent, M. (2002). IT infrastructure for strategic agility. Available at http://dx.doi.org/10.2139/ssrn.317307 THE IMPACT OF INNOVATION OBJECTIVES ON INDUSTRY-ACADEMIA COLLABORATION. A LOOK TOWARDS SUSTAINABILITY MARÍA DE LAS MERCEDES GRACIA-LABORDA, CAROLINA LÓPEZ-NICOLÁS, GABRIEL LOZANO-REINA, ÁNGEL MEROÑO-CERDÁN, FRANCISCO JOSÉ MOLINA-CASTILLO University of Murcia, Faculty of Economics and Business, Murcia, Spain mariamercedes.gracial@um.es, carlopez@um.es, gabriel.lozano@um.es, angelmer@um.es, fjmolina@um.es The literature on innovation has been very prolific in highlighting the importance of companies developing new products, processes or business models in order to be more competitive in the marketplace. Empirical studies have shown that this innovative activity has translated into superior results for companies that have engaged in innovation. However, most of these initiatives have been studied mainly from the company's point of view without considering the contribution that academia can make to these innovation processes. This paper explores precisely how it is possible to achieve better results in innovation objectives through industry-academia col aboration (IAC). To Keywords: Industry- this end, a sample of 7638 Spanish companies is analysed, Academy distinguishing between those that have linked their innovation collaboration; innovation objectives to collaboration between the company and higher objectives; education centres. The results reveal that this IAC helps business reinforce innovation objectives, demonstrating that the union of model innovation; the academic and business worlds improves the results of economic, business innovation processes. This has important theoretical social, and implications as it offers new insights into the analysis of environmental levels, innovation processes and business implications as it proves that sustainable there is a need to develop platforms that encourage IAC. BMI DOI https://doi.org/10.18690/um.fov.4.2023.24 ISBN 978-961-286-751-5 380 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction In the last few years, Business Model Innovation (BMI) has gained significant interest among academics and professionals as an avenue for business development and for achieving above-average performance (Ghinoi & Di Toma, 2022). Its implementation is essential in order to foster long-term sustainable competitive advantage and to recognize new techniques to business organization (Kraus et al., 2020). Despite the growing literature in this research field, there are still many issues regarding BMI that remain limited (Foss & Saebi, 2017). The relationship between BMI and industry-academia collaboration (IAC) is one of them. The engagement between university and industry or multi-stakeholder collaboration has the potential to generate synergies both for industry and academia (Haug, 2018) and results in higher levels of innovation and advances in knowledge, technological enhancements, and industry objectives (Arshed et al., 2022). Universities provide workforce that can be useful to firms, and are the source of innovative ideas to set up new business ventures (Ahmed et al., 2022). Based on this, when industry require research in unknown areas where they do not have access yet, they look for academia. Besides, this educational institution has the knowledge expertise and the research methods to designed solutions which are valid and relevant, so that industries should benefit from its collaboration (Burova et al., 2021). Several topics regarding IAC and innovation have been addressed in the literature, special y related to technology transfer (Blundi et al., 2019; Ravi & Janodia, 2022). To the best of our knowledge, there is no research that assesses the relationship between BMI objectives and IAC. We consider that BMI objectives may lead a company to decide to col aborate with research institutions as a means to achieve them. To fill this gap, this research has a twofold objective: on the one hand, to determine whether the importance that firms give to BMI objectives is higher when they collaborate with universities; and, on the other hand, to identify which of these innovation objectives contribute to IAC. On this purpose, data was drawn from the Spanish (Eurostat) Community Survey (CIS) for 2014-2016 in order to evaluate if there is an influence of BMI objectives on the collaboration between industry and university. In addition, through hierarchical logistic regression a set of seven BMI objectives, classified with the Triple Layered M. de las Mercedes Gracia-Laborda, et al.: The Impact of Innovation Objectives on Industry-Academia Col aboration. A Look Towards Sustainability 381 Business Model Canvas (TLBMC), have been identified which contribute to collaboration between companies and universities. The remaining paper is structured as follow. First, we provide a review of the salient literature on BMI and Industry-Academia Collaboration. Next, the methodology to collect data from 7,638 Spanish organizations is detailed. Then, the results of the empirical analysis are discussed. Finally, conclusions are summarized in section five. 2 Business Model Innovation Objectives and Industry-Academia Collaboration: Literature review The term business model was introduced for the first time by Bel man et al. (1957) (Groesser & Jovy, 2016), yet it was the arrival of internet and the expansion of information and communication technologies (ICT) that prompted it to be explored to a greater extend in the 1990s (DaSilva & Trkman, 2014). Despite the fact that there is a large number of definitions provided regarding the concept; many contributions to the literature agree with the notion of business model as “the logic the firm fol ows to operate its resources and to create and capture value for external and internal stakeholders” (Ammar & Chereau, 2018, p. 2). A business model is built on the three main value dimensions of a business: the creation, delivery, and value capture (Clauss et al., 2020). Value creation occurs when a firm matches the customer’s demands with a re-organisation of its resources which lead to enhanced efficient (Kraus et al., 2020); value delivery explains the mechanisms how to bring the created value to the customers (Dahan et al., 2010; Spieth et al., 2021), and the value capture indicates how a firm will obtain money from developing its activity (Osterwalder & Pigneur, 2010). In this context, innovation constitutes one of the cornerstones of business model (Budler et al., 2021). BMI is essential for business sustainability (Breier et al., 2021) and represents a source of a firm’s competitive advantage (Latifi et al., 2021; Pieroni et al., 2019). Although there is no consensus on its definition, we follow Foss & Saebi, (2017, p. 201) who state that it can be understood as “designed, novel, nontrivial changes to the key elements of a firm’s business model and/or the architecture linking these elements”. While business model objectives refer to the overal business objective that a company seeks to achieve in developing or shaping its business model (Molina-Castillo et al., 2019); innovation objectives comprises a 382 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY firm’s distinguishable aims that depict intentions and strategies that involve innovation efforts (OECD & European Comission, 2018). Based on this, we refer to BMI objectives as those that involve innovation efforts and imply novel, trivial and designed modifications in a firm’s business model. Thus, these innovation objectives may lead to innovation activities and performance (Meroño-Cerdan & López-Nicolas, 2013). Literature on BMI has pointed out several tools and path to design and assist it (Heikkila et al., 2016). Nowadays, businesses are required to innovate their business models by creating sustainable value on its economic, social, and environmental levels (López-Nicolás et al., 2021). The TLBMC thus constitutes a tool which provides a ful understanding of the business model of a company, includes the three levels previously mentioned and that support sustainability-oriented BMI (Joyce & Paquin, 2016). More specifical y, the economic layer consists of the Business Model Canvas proposed by Osterwalder & Pigneur (2010), which distinguish nine modules which are interrelated with each other, being these: customer value proposition, segments, customer relationships, distribution channels, key resources, key activities, partners, costs and revenues. Moreover, Joyce & Paquin (2016) describe the aim and composition of the aforementioned two other layers of the business model, the environmental and the social. These authors state that the former has as its main purpose to assess how a firm can produce further environmental benefits than environmental effects and encompasses functional value, materials, production, supplies and outsourcing, distribution, use phase end-of-life and environmental impacts and benefits; the latter attempts to capture what are the key social impacts of the organization that derive from its relationships; and its components are social value, employee, governance, communities, societal culture, scale of outreach, end-users and social impacts and benefits. We consider that there are several BMI objectives that not only do focus on specific areas of the business model, but they also drive firms to collaborate with other stakeholders. The IAC may be a way for achieving BMI and for generating synergies both for industry and for academia (Arshed et al., 2022). In fact, among the main benefits resulting from this collaboration, we can remark the emergence of general solutions to issues related to products or service offering, which encourages innovation in the business model (Sjöö & Hellström, 2021). For achieving this, (Vico et al., 2015) find as the main reasons to start a col aboration the following: assistance M. de las Mercedes Gracia-Laborda, et al.: The Impact of Innovation Objectives on Industry-Academia Col aboration. A Look Towards Sustainability 383 in problem solving, provision of specialized services, patent generation and introducing innovations (e.g., new products, processes, findings from research, etc.). In this research, several BMI objectives have been classified within the different layers of the TLBMC to determine which of them encourage the company to collaborate with universities or IAC. Specifically, Table 1 shows to which business model dimension and layer the analyzed BMI objectives would correspond – remarking that most of the objectives are linked to the dimension of value creation. Table 1: Integration of dimensions layers of business model and business model innovation objectives Dimension Layer BMI Objective Expand the grade of good and services Replace obsolete products or processes Improve quality of goods or services Economic Improve flexibility for producing goods or delivering services Increase the capacity of producing goods or delivering services Value creation Reduce material per unit of output Reduce energy per unit of output Environmental Reduce negative environmental impacts/ deliver environmental benefits Improve public health, safety or security Comply with mandatory regulations Total employment growth Social Increase in qualified employment Maintenance of employment Value delivery Economic Enter new markets Increase market share Economic Reduce labour costs per unit of output Reduce energy per unit of output Value capture Reduce negative environmental impacts/ Environmental deliver environmental benefits Improve public health, safety or security Comply with mandatory regulations Source: Own el aboration based on Joyce & Paquin, (2016) and OECD & European Comission, (2018) 384 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3 Methodology Our dataset comes from Spanish Community Innovation Survey (CIS). The CIS questionnaire is extensively used in most European countries, especially in the UK, France, Spain, and Italy (e.g., Aronica et al., 2022; Evangelista & Vezzani, 2010; Ganter & Hecker, 2013; Hervas-Oliver et al., 2014, 2015; Lubacha & Wendler, 2021; Wei et al., 2022) but also considered the most influential innovation questionnaire even in non-EU countries (Wei et a., 2022). It has become an interesting source of research data to study complementarities between different forms of innovation (Ballot et al., 2015). This survey, conducted by the Spanish National Statistics Institute, provides information about the innovation process, its structure, the relationships between that process and firms’ technological strategy, the factors affecting their capability to innovate and companies’ performance. The respondent units (companies) are sent a letter of presentation of the survey, which includes the user and password for online completion. Since 2013, access to web completion is carried out via the secure protocol page https://iria.ine.es. Once this letter has been received, companies have a period of 15 days to complete and send the questionnaire. The Statistics Institute establishes an initial telephone contact with the company to check that the questionnaire has been received. If the completed questionnaire has not been received by the deadline, the necessary telephone and written complaints will be made. The monitoring of the data col ection schedule and the quality control of the information has been carried out by Statistics Institute’s Central Services. The response rate was 93,17 percent. The 2016 survey addressed innovation activity for the period 2014-2016. The sample for that period consisted of 7,638 companies with a minimum size of 10 employees operating in different sectors (agriculture, construction, industry, commerce, and services). When using CIS data, a potential bias may arise related to the sample selection problem. Because of the CIS questionnaire structure, some variables regarding innovation are available only for firms which have introduced at least one process or one product innovation. This may create a selection bias if the econometric analysis is limited to that sub-sample of companies which is likely to be not randomly drawn from the larger population. This type of bias may distort coefficients. As in other studies (Evangelista & Vezzani, 2010), the choice of variables used here and our estimation strategy allows us to include all the firms present in the CIS sample, avoiding the selection bias problem mentioned. M. de las Mercedes Gracia-Laborda, et al.: The Impact of Innovation Objectives on Industry-Academia Col aboration. A Look Towards Sustainability 385 A description of variables included in the analysis is given in Table 2. All of them comes from Oslo Manual (OECD & European Comission, 2018; OECD & Eurostat, 2005). Table 2: Variable in the analysis Variable Scale Dependent variable: Dummy variable Industry-Academia Collaboration (in the 2014-2016 (1=collaboration, 0=non- period) collaboration) Independent variables: Continuous variables from 1 Importance of business model innovation objectives (noimportance) to 4 (high importance) Economic objectives: 1-4 interval - Wider range of goods or services (OBJ_1) - Substitution of outdated products or processes (OBJ_2) - Penetration into new markets (OBJ_3) - Increased market share (OBJ_4) - Higher quality of goods or services (OBJ_5) - Greater flexibility in production or service provision (OBJ_6) - Greater production capacity or provision of services (OBJ_7) - Lower labour costs per unit produced (OBJ_8) - Less materials per unit produced (OBJ_9) - Less energy per unit produced (OBJ_10) Specifical y, objectives 1 to 5 refer to “product innovation” and objectives 6 to 10 refer to “innovation in process”. Environmental objectives: 1-4 interval - Less environmental impact (OBJ_11) - Improved health and safety (OBJ_12) - Compliance with Environmental, Health or Safety Regulatory Requirements (OBJ_13) Social objectives: 1-4 interval 386 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY - Increase in total employment (OBJ_14) - Increase in qualified employment (OBJ_15) - Maintenance of employment (OBJ_16) Control variables Firm year Dummy variable: 0= mature company (more than 10 years); and 1= young company (up to 10 years). Firm sector Continuous variable that takes the values: 0=Agriculture, 1=Industry, 2=Construction, 3=Commerce and 4=Services) 4 Results Before stimating the regression model, ANOVA tests were carried out. Although not presented here due to extension limits, ANOVA results show the importance that companies give to BMI objectives depending on whether or not they collaborate with universities in the 2014-2016 period. In general, the means obtained in each objective are higher when the company collaborates with a higher education institution. In addition, these means are especial y high regarding economic objectives –both those referring to product innovation and to process innovation. Going further, the differences in means are statistically significant in all objectives. This al ows us to affirm that the importance that companies give to innovation (economic, environmental, and social) objectives are greater when the organization collaborates with the university compared to those firms which do not collaborate with higher education institution. Table 3 shows the results of the hierarchical logistic regression. The dependent variable in al the models is the company-university collaboration, while the independent and control variables vary depending on the model; in particular, Model 1 includes the constant and firm age; Model 2 adds the sector; and Model 3 adds economic, environmental, and social objectives. Focusing on Model 3 (which is the one that includes BMI objectives as independent variables), we obtain that the seven M. de las Mercedes Gracia-Laborda, et al.: The Impact of Innovation Objectives on Industry-Academia Col aboration. A Look Towards Sustainability 387 objectives that statistically contribute to collaboration between companies and universities are: (i) wider range of goods or services; (ii) penetration into new markets, (iii) increased market share; (iv) less environmental impact; (v) improved health and safety; (vi) increase in total employment; and (vii) increase in qualified employment. Among them, we appreciate how three are economic objectives (all of them related to product innovation), two are environmental objectives, and the remaining two are social objectives. Table 3: Hierarchical Logistic Regression Col aboration with universities Model 1 Model 2 Model 3 Wider range of goods or services (OBJ_1) 0.170*** Substitution of outdated products or processes (OBJ_2) -0.058 Penetration into new markets (OBJ_3) 0.186*** Increased market share (OBJ_4) -0.154** Higher quality of goods or services (OBJ_5) 0.085 Greater flexibility in production or service provision (OBJ_6) 0.000 Greater production capacity or provision of services (OBJ_7) -0.017 Lower labour costs per unit produced (OBJ_8) -0.023 Less materials per unit produced (OBJ_9) -0.040 Less energy per unit produced (OBJ_10) 0.111 Less environmental impact (OBJ_11) 0.379* Improved health and safety (OBJ_12) -0.063*** Compliance with environmental, health or safety regulatory requirements (OBJ_13) -0.034 Increase in total employment (OBJ_14) -0.250*** Increase in qualified employment (OBJ_15) 0.467*** Maintenance of employment (OBJ_16) 0.103 Firm year 0.520 0.500 0.598 Constant -1.436*** -1.429*** -3.761*** Sector Control No Yes Yes Chi-squared (model) 2.114 47.778*** 433.795*** R-squared 0.001 0.017 0.146 388 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 5 Conclusions As highlighted at the beginning of this paper, the study of IAC requires further analysis to discover how to improve the innovation processes that can be developed. The results show that there are essential benefits at a business level when the company actively collaborates with the academic world, demonstrating how unity is a strength and al ows better results to be achieved than when this collaboration is not carried out (Yi et al., 2022). There is no doubt that the company has a deep understanding of the needs of its customers as it has developed its products precisely to meet those needs (Keiningham et al., 2020). But academia also has much to offer to complement this activity as it is responsible for analyzing in detail the cognitive processes by which customers decide to purchase one product rather than another from among those available in the set under consideration (Wijekoon et al., 2021). In the academic world, this is usually carried out with experimental studies with subjects and nowadays it is very common to analyze aspects of neuromarketing applied to the marketing of new products (Kansra et al., 2022). In this vein, we observe how IAC is fundamental when it comes to achieving innovation objectives related to the development and launch of new products on the market (Liu et al., 2022). Analogously, we see how the study results demonstrate that IAC is very useful for penetrating new markets and increasing market share (Canabal & White, 2008). Companies collaborating with universities have probably benefited from al the predictive models of new product adoption that have long been developed in academia (Rogers, 2003). Researchers devote significant effort to understanding how a new product can reach the market earlier and better than the product with which it competes (Suarez et al., 2015). In fact, the modelling of consumer adoption processes developed by academics is becoming increasingly complex. We also see how the results clearly demonstrate the impact on employment aspects through IAC. In this way, objectives related to improving employment and job quality are actively promoted when a business col aborates with academia (Mohammadi et al., 2017). This is undoubtedly a fundamental fact that should be actively considered from a governmental policy point of view. Companies require a skil ed workforce but at the same time a workforce that matches the specific skil s demanded by these organisations (Schweisfurth & Raasch, 2018). The academic M. de las Mercedes Gracia-Laborda, et al.: The Impact of Innovation Objectives on Industry-Academia Col aboration. A Look Towards Sustainability 389 world must try to adapt its teaching processes, learning methodologies and contents to the business reality. Our results show that this IAC could have significant social implications in this sense. However, we also find that other innovation objectives have not shown significant results from IAC. It is, therefore essential to further explore how to improve IAC to better design collaborative business models that al ow for fewer materials per unit produced or greater flexibility in production or service provision (Heirati & Siahtiri, 2019). In the same way, longitudinal studies on these collaborative processes could shed light on the barriers that may exist in these types of collaborations that are difficult to analyze in cross-sectional studies (Bitetti & Gibbert, 2022). To this end, it is essential to support initiatives at the European level to develop platforms that favour IAC. A clear example of this is the venture alliances platform (https://www.ventureal iances.eu) that, for the last year has been helping companies and academics to find the right partner so that they can achieve innovation objectives that to date, have been studied in isolation between industry and academia. This work also reveals the need to replicate this work in other countries to verify the benefits of collaboration and how it contributes to improving the development of new products, processes, and innovative and sustainable business models. References Ahmed, F., Fattani, M. T., Ali, S. R., & Enam, R. N. (2022). Strengthening the Bridge Between Academic and the Industry Through the Academia-Industry Col aboration Plan Design Model. Frontiers in Psychology, 13, 875940. https://doi.org/10.3389/fpsyg.2022.875940 Ammar, O., & Chereau, P. (2018). Business model innovation from the strategic posture perspective An exploration in manufacturing SMEs. European Business Review, 30(1), 38-65. https://doi.org/10.1108/EBR-09-2016-0119 Aronica, M., Fazio, G., & Piacentino, D. (2022). A micro-founded approach to regional innovation in Italy. Technological Forecasting and Social Change, 176, 121494. https://doi.org/10.1016/j.techfore.2022.121494 Arshed, N., Ahmad, W., & Hanif, U. (2022). A Spatial Temporal Exploration of Factors Motivating Academia-Industry Col aboration. Journal of the Knowledge Economy, 13(1), 521-540. https://doi.org/10.1007/s13132-021-00729-6 Ballot, G., Fakhfakh, F., Galia, F., & Salter, A. (2015). The fateful triangle: Complementarities in performance between product, process and organizational innovation in France and the UK. Research Policy, 44(1), 217-232. https://doi.org/10.1016/j.respol.2014.07.003 Bel man, R., Clark, C., Malcom, D., Craft, C., & Ricciardi, F. (1957). On the construction of a multi-stage, multi-person business game. Operations Research, 5(4), 469-503. 390 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Bitetti, L., & Gibbert, M. (2022). The ROAD to continuous business model innovation: A longitudinal study unveiling patterns of cognitive sensing dynamic capabilities. Creativity and Innovation Management, 31(1), 123-140. https://doi.org/10.1111/caim.12477 Blundi, D., Nonato, A. C., Medeiros, S., Filgueiras, J., Veiga, F., Travassos, G., & Oring, V. (2019). Technology Appropiation and Technology Transfer in the Brazilian Mining Sector. Breier, M., Kal muenzer, A., Clauss, T., Gast, J., Kraus, S., & Tiberius, V. (2021). The role of business model innovation in the hospitality industry during the COVID-19 crisis. International Journal of Hospitality Management, 92, 102723. https://doi.org/10.1016/j.ijhm.2020.102723 Budler, M., Župič, I., & Trkman, P. (2021). The development of business model research: A bibliometric review. Journal of Business Research, 135, 480-495. https://doi.org/10.1016/j.jbusres.2021.06.045 Burova, A., Heinonen, H., Becerril Palma, P., Keskinen, T., Hakulinen, J., Opas, V., Mäkelä, J., Ronkainen, K., Siltanen, S., Raisamo, R., & Turunen, M. (2021). Toward Efficient Academia-Industry Col aboration: A Case Study of Joint VR System Development. Academic Mindtrek 2021, 176-185. https://doi.org/10.1145/3464327.3464367 Canabal, A., & White, G. O. (2008). Entry mode research: Past and future. International Business Review, 17(3), 267-284. https://doi.org/10.1016/j.ibusrev.2008.01.003 Clauss, T., Bouncken, R. B., Laudien, S., & Kraus, S. (2020). Business model reconfiguration and innovation in SMEs: A mixed-method analysis from the electronics industry. International Journal of Innovation Management, 24(2). https://doi.org/10.1142/S1363919620500152 Dahan, N., Doh, J., Oetzel, J., & Yaziji, M. (2010). Corporate-NGO col aboration: Co-creating new business models for developing markets. Long Range Planning, 43(2-3), 326-342. DaSilva, C. M., & Trkman, P. (2014). Business Model: What It Is and What It Is Not. Long Range Planning, 47(6), 379-389. https://doi.org/10.1016/j.lrp.2013.08.004 Evangelista, R., & Vezzani, A. (2010). The economic impact of technological and organizational innovations. A firm-level analysis. Research Policy, 39(10), 1253-1263. https://doi.org/10.1016/j.respol.2010.08.004 Foss, N. J., & Saebi, T. (2017). Fifteen Years of Research on Business Model Innovation: How Far Have We Come, and Where Should We Go? Journal of Management, 43(1), 200-227. https://doi.org/10.1177/0149206316675927 Ganter, A., & Hecker, A. (2013). Deciphering antecedents of organizational innovation. Journal of Business Research, 66(5), 575-584. https://doi.org/10.1016/j.jbusres.2012.02.040 Ghinoi, S., & Di Toma, P. (2022). Conceptualising business model innovation: Evidence from the managers’ advice network. Innovation, 24(2), 251-271. https://doi.org/10.1080/14479338.2021.1885298 Groesser, S. N., & Jovy, N. (2016). Business model analysis using computational modeling: A strategy tool for exploration and decision-making. Journal of Management Control, 27(1), 61-88. https://doi.org/10.1007/s00187-015-0222-1 Haug, G. (2018). Universidades y empresas: Apuntes para crear sinergias con sentido. Heikkila, M., Bouwman, H., Heikkila, J., Haaker, T., Lopez-Nicolas, C., & Riedl, A. (2016, junio 19). Business Model Innovation Paths and Tools. 29th Bled eConference Digital Economy, Bled, Slovenia. Heirati, N., & Siahtiri, V. (2019). Driving service innovativeness via col aboration with customers and suppliers: Evidence from business-to-business services. Industrial Marketing Management, 78, 6-16. https://doi.org/10.1016/j.indmarman.2017.09.008 Hervas-Oliver, J.-L., Sempere-Ripol , F., & Arribas, I. (2015). Asymmetric modeling of organizational innovation. Journal of Business Research, 68(12), 2654-2662. https://doi.org/10.1016/j.jbusres.2015.04.005 Hervas-Oliver, J.-L., Sempere-Ripol , F., & Boronat-Moll, C. (2014). Process innovation strategy in SMEs, organizational innovation and performance: A misleading debate? Smal Business Economics, 43(4), 873-886. https://doi.org/10.1007/s11187-014-9567-3 M. de las Mercedes Gracia-Laborda, et al.: The Impact of Innovation Objectives on Industry-Academia Col aboration. A Look Towards Sustainability 391 Joyce, A., & Paquin, R. L. (2016). The triple layered business model canvas: A tool to design more sustainable business models. Journal of Cleaner Production, 135, 1474-1486. https://doi.org/10.1016/j.jclepro.2016.06.067 Kansra, P., Oberoi, S., Gupta, S. L., & Singh, N. (2022). Factors limiting the application of consumer neuroscience: A systematic review. Journal of Consumer Behaviour, cb.2131. https://doi.org/10.1002/cb.2131 Keiningham, T., Aksoy, L., Bruce, H. L., Cadet, F., Clennel , N., Hodgkinson, I. R., & Kearney, T. (2020). Customer experience driven business model innovation. JOURNAL OF BUSINESS RESEARCH, 116, 431-440. https://doi.org/10.1016/j.jbusres.2019.08.003 Kraus, S., Filser, M., Puumalainen, K., Kailer, N., & Thurner, S. (2020). Business Model Innovation: A Systematic Literature Review. International Journal of Innovation and Technology Management, 17(06), 2050043. https://doi.org/10.1142/S0219877020500431 Latifi, M.-A., Nikou, S., & Bouwman, H. (2021). Business model innovation and firm performance: Exploring causal mechanisms in SMEs. Technovation, 107. https://doi.org/10.1016/j.technovation.2021.102274 Liu, A., Gu, J., & Liu, H. (2022). The fit between firm capability and business model for SME growth: A resource orchestration perspective. R&D Management, 52(4), 670-684. https://doi.org/10.1111/radm.12513 López-Nicolás, C., Ruiz-Nicolás, J., & Mateo-Ortuño, E. (2021). Towards Sustainable Innovative Business Models. Sustainability, 13(11), 5804. https://doi.org/10.3390/su13115804 Lubacha, J., & Wendler, T. (2021). Do European firms obey the rules? Environmental innovativeness in light of institutional frameworks. Industry and Innovation, 28(9), 1196-1223. https://doi.org/10.1080/13662716.2021.1929869 Meroño-Cerdan, A. L., & López-Nicolas, C. (2013). Understanding the drivers of organizational innovations. The Service Industries Journal, 33(13-14), 1312-1325. https://doi.org/10.1080/02642069.2013.815736 Mohammadi, A., Broström, A., & Franzoni, C. (2017). Workforce Composition and Innovation: How Diversity in Employees’ Ethnic and Educational Backgrounds Facilitates Firm-Level Innovativeness: WORKFORCE COMPOSITION AND INNOVATION. Journal of Product Innovation Management, 34(4), 406-426. https://doi.org/10.1111/jpim.12388 Molina-Castillo, F.-J., Meroño-Cerdan, A.-L., & López-Nicolás, C. (2019). Impact of business model objectives on marketing innovation activities: A comparison between manufacturing and service firms. European Journal of Innovation Management, 23(1), 177-195. https://doi.org/10.1108/EJIM-12-2018-0259 OECD, & European Comission (Eds.). (2018). Oslo manual 2018: Guidelines for collecting, reporting and using data on innovation (4th edition). OECD Publishing. https://doi.org/10.1787/9789264304604-en OECD, & Eurostat. (2005). Oslo Manual: Guidelines for Col ecting and Interpreting Innovation Data. (Third Edition). Osterwalder, A., & Pigneur, Y. (2010). Business model generation. (Self published). Deuto. Pieroni, M. P. P., McAloone, T. C., & Pigosso, D. C. A. (2019). Configuring New Business Models for Circular Economy through Product-Service Systems. Sustainability, 11(13). https://doi.org/10.3390/su11133727 Ravi, R., & Janodia, M. D. (2022). Factors Affecting Technology Transfer and Commercialization of University Research in India: A Cross-sectional Study. Journal of the Knowledge Economy, 13(1), 787-803. https://doi.org/10.1007/s13132-021-00747-4 Rogers, E. M. (2003). Diffusion of innovations (5th ed). Free Press. Schweisfurth, T. G., & Raasch, C. (2018). Absorptive capacity for need knowledge: Antecedents and effects for employee innovativeness. Research Policy, 47(4), 687-699. https://doi.org/10.1016/j.respol.2018.01.017 Sjöö, K., & Hel ström, T. (2021). The two sides of the coin: Joint project leader interaction in university-industry col aboration projects. R&D Management, 51(5), 484-493. https://doi.org/10.1111/radm.12452 392 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Spieth, P., Laudien, S. M., & Meissner, S. (2021). Business model innovation in strategic alliances: A multi‐layer perspective. R&D Management, 51(1), 24-39. https://doi.org/10.1111/radm.12410 Suarez, F. F., Grodal, S., & Gotsopoulos, A. (2015). Perfect timing? Dominant category, dominant design, and the window of opportunity for firm entry: Perfect Timing? Dominant Category and Window of Entry. Strategic Management Journal, 36(3), 437-448. https://doi.org/10.1002/smj.2225 Vico, M. L., Nina, M. R., & Torres, J. V. (2015). ¿Qué quieren las empresas de las universidades? Oportunidades de colaboración para desarrol ar innovaciones. Forum Empresarial. Centro de Investigaciones Comerciales e Iniciativas Académicas., 3(1.1), 1-11. Wei, J., Li, Y., Liu, X., & Du, Y. (2022). Enterprise characteristics and external influencing factors of sustainable innovation: Based on China’s innovation survey. Journal of Cleaner Production, 372, 133461. https://doi.org/10.1016/j.jclepro.2022.133461 Wijekoon, A., Salunke, S., & Athaide, G. A. (2021). Customer heterogeneity and innovation‐based competitive strategy: A review, synthesis, and research agenda. Journal of Product Innovation Management, 38(3), 315-333. https://doi.org/10.1111/jpim.12576 Yi, Y., Chen, Y., & Li, D. (2022). Stakeholder ties, organizational learning, and business model innovation: A business ecosystem perspective. Technovation, 114, 102445. https://doi.org/10.1016/j.technovation.2021.102445 YOUNG ADULTS’ ATTITUDE TOWARDS DIGITAL PAYMENT METHODS AND FINANCIAL RESPONSIBILITY EMMA ROSENLIND,1 ULRIK SÖDERSTRÖM,1 OLE NORBERG,1 HELEN CRIPPS,2 THOMAS MEJTOFT1 1 Umeå University, Digital Media Lab, Umeå, Sweden emma.rosenlind@umu.se, ulrik.soderstrom@umu.se, ole.norberg@umu.se, thomas.mejtoft@umu.se 2 Edith Cowan Univerity, School of Business and Law, Perth, Australia h.cripps@ecu.edu.au There are numerous online payment methods when shopping online. This paper presents the evolution of payment methods through the years and investigates which payment methods young adults prefer and investigates if and how modern payment methods affect financial responsibility among young adults. The study was conducted through a survey and fol ow-up interviews with Swedish young adults. 75% claimed they make online Keywords: purchases 1-3 times a week or more, of which 35% always check payment their bank account before paying. The study suggests that direct methods, financial payment using their smart device (Swish) is the most preferred responsibility, payment method. The research suggests that it is the finance, combination of the different payment methods available that is purchase behaviour, impacting financial responsibility among young adults rather Bled than the payment method itself. eConference DOI https://doi.org/10.18690/um.fov.4.2023.25 ISBN 978-961-286-751-5 394 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Traditionally money (cash) was handled banks. When you wanted to use the money, you had to walk to the bank so the tel er could hand it over to you. Then you had to take the money to the store if you wanted to buy something (Eriksson, & Sandhill, 2019). In the early 60s, debit cards were launched as the first step toward digitalizing payments (Leblebici, 2012; Lauer, 2020). The debit card was similar to the check, but its infrastructure and payment guarantee to the sel er eventual y outcompeted checks and cash (Wisehn, 2019). About 10 years later, Michael Aldrich invented a system that came to be the foundation of e-commerce (Aldrich, 2011). The breakthrough for online payments and e-commerce happened in the early 90s and big tech companies such as PayPal and Amazon worked hard to lead the market forward (Ferrera & Kessedjian, 2019). In 2005, The National Federation coined the term “Cyber Monday”, due to the high number of online transactions carried out during Thanksgiving that year (McFerrin, 2021). For 30 years, online payments and e-commerce have constantly developed. Since almost al payments and transactions are something that can be done by yourself without interaction with another human party, the tel er in the bank, is now no longer needed. The list of different ways of paying on the internet is long, but al of them have one thing in common, they strive toward making the experience for the customers as easy as possible (van der Cruijsen, & Plooij, 2018; Page, 2021). 2 Research Objectives With the rise of the cashless society and the ephemeral nature of digitalization payments means that young people have grown up in an environment where payments are more straightforward, accessible, and frictionless (Senali et al., 2022; Tatum, 2022). The digitalization of payment methods and the rise of “buy now pay later” schemes have seen increasing numbers of young people experiencing debt and financial distress (Coffey et al., 2023). This paper aims to understand young adult’s approach to finance in the context of how different payment methods can affect financial responsibility among young adults. To answer this, the following research questions have been formulated: E. Rosenlind, U. Söderström, O. Norberg, H. Cripps, T. Mejtoft: Young Adults’ Attitude Towards Digital Payment Methods and Financial Responsibility 395 1. Which online payment methods do young adults use and prefer today? 2. How do online payment methods affect young adult’s level of financial responsibility? 3 Literature Review The fol owing section wil introduce some key concepts of this study. 3.1 Financial responsibility Financial responsibility is a term that is used when talking about the process of handling money and assets in a way that is considered productive and in the best interest of the individual. Financial responsibility is about having a mindset where you can look beyond the wants of today to be able to provide the needs of tomorrow (Senali et al., 2022; Tatum, 2022). 3.2 BankID BankID is a Swedish digital ID launched in 2002. It is an electronic ID document that can be compared with a national ID card and is based on the Swedish personal number that is unique for each citizen. BankID makes it possible for companies, banks, authorities, and other organizations to identify and enter into agreements with private persons via the internet. BankID has become an effective addition to the individual security device that many banks provide (BankId, n.d.-a). Today, 12 of Sweden’s biggest banks use BankID as an identification service which means that when paying online with third-party payment methods, such as Klarna and Swish, the identification wil be handled through BankID. The identifying process, when paying online, only consists of reading through the information about the purchase (amount of money and receiver) and confirming with either a code consisting of digits or any other smartphone unique verification procedure, such as fingerprint or facial recognition (BankID, n.d.-b). 396 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3.3 Digital Payment Methods The fol owing section describes five different payment methods that wil be covered in this study. The payment methods are commonly used in Sweden, and some only operate in Sweden (Arvidsson, 2019; Insight Intelligence, 2020). 1) Klarna: Klarna is a global leading payment and shopping service that currently provides 150 million active customers with a flexible and smart purchase experience. In Sweden, Klarna offers three different ways of paying – (1) Direct payment, which means that the money is drawn from the account straight away, (2) Pay after delivery with “Klarna – Pay in 30 days” (“Klarna – Få först. Betala sen” in Swedish), which means that the customer do not have to pay until receiving the goods, and (3) Instalment plans where the customer can spread the cost of their purchase over up to 36 months. Al the payments and purchases can be handled in Klarna’s mobile app or website (Klarna, 2022). 2) Swish: Swish is a Swedish mobile-based payment method where the user can send money through an app that is directly connected to their bank account (Arvidsson, 2019). It was launched in 2012 as a service for payments between private users. In 2014, Swish expanded to companies and organizations (Swish, n.d.). As of 2022, Swish has over 8.3 million private users and over 300,000 connected companies, and during June 2022 around 40 billion SEK was sent via Swish (Arvidsson, 2019; Swish, 2022). 3) Apple Pay: Apple Pay is a way of handling payments, both in physical stores and online. The bank card is connected to Apple Pay through a unique device number and a unique transaction code. This means that the personal card number is never stored directly on the device and runs less chance of getting hijacked. When paying online with Apple Pay no confirmation with an external verification service is needed. Depending on the device the confirmation can be done with either Face ID, Touch ID, or passcode (Apple, n.d.). E. Rosenlind, U. Söderström, O. Norberg, H. Cripps, T. Mejtoft: Young Adults’ Attitude Towards Digital Payment Methods and Financial Responsibility 397 4) Invoice: An invoice is a physical or electronic document that contains a payment claim. It can be sent from one company to another company or to a private person. All online banks work differently, but paying an invoice is often carried out similarly (Hansson, 2022). 5) Card payment: Card payment is when the payment is done directly by using a bank card. To complete the purchase, the card number, the card’s validity period, and the CVC code need to be entered. Often the purchase wil have to be confirmed via a third-party security service, such as BankId (Swedbank, n.d.). 4 Method The study was carried out by using a survey with fol ow-up interviews. The survey was created with the tool Typeform and was sent out via Facebook and Slack. The target group of the study is young adults, which are defined as men and women between 18-25 years old. The respondents were selected from Umeå University and We Know It, a consultant firm for students. The reason behind the choice was for the respondents to preferably be young adults and somehow be experienced with digital payment methods. 4.1 Survey questions The first part of the study aimed to collect demographic data about the user. Questions such as age, gender, and daily occupation were asked. The second part of the survey consisted of questions regarding which payment method the respondent prefers and most frequently uses. Questions such as “Which payment method do you prefer?” and “Rank the following payment methods according to which you are most likely to use” were asked. The last part of the survey consisted of questions regarding how the respondents view their own financial responsibility. Questions such as “Do you know how much money you have in your bank account right now?” and “Do you check your bank account before you make a purchase?” were asked. At the end of the survey, the respondents were asked to enter their e-mail addresses to participate in follow-up interviews. After the answers were analyzed, four people were contacted for interviews. Based on their answers in the survey, questions regarding their payment habits and financial responsibility were asked. The study only investigates payment methods that are used in Sweden. All questions were 398 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY written in Swedish but were translated into English in the result section of this paper. Another limitation is that most of the participants were Swedish young adults studying at the university and may therefore not give a correct representation of the whole target group. 5 Results 5.1 Demographic data A total of 57 respondents within the target group (18-25 years old) answered the survey (63% male and 37% female). Due to the target group and the choice for distributing the survey, 95% of the population were students and only 5% were employed. 5.2 Payment methods The data collected through the survey showed that most of the participants prefer Swish as a payment method when shopping online (Figure 1). When asked about the factors that influenced their choice of payment method, most part of the participants cited simplicity and flexibility as the main reasons for preferring Swish. The second most popular payment method was card payment. Respondents supported their choice of card payment due to the ability to save your card details in your browser, which makes it simple and easy or feeling of familiar and not knowing any other payment option. The third most popular payment method, with 21%, was “Klarna – Pay in 30 days. One factor that made them choose this payment was that they want to receive the item before they pay of it. Another factor was that they do not want to pay for something they may want to return. Invoice and Klarna instalment plan were also included as alternatives in the question, but these options were by any of the respondents as their preferred payment method. E. Rosenlind, U. Söderström, O. Norberg, H. Cripps, T. Mejtoft: Young Adults’ Attitude Towards Digital Payment Methods and Financial Responsibility 399 Figure 1: Which payment method do you prefer when shopping online? The respondents were also asked to rank the payment methods according to which they are most likely to use. The answers agreed with the previous results, with Swish ranked in first place. Of the respondents 54% have tried the “Klarna – Pay in 30 days” function as part of the ranking. When asked why they tried it, 14% of respondents were motivated by not having enough money for the purchase. However, the most common reason for using it was the ability to see the item before paying. For 30% of the respondents the most common reason for using invoices was that there was no other payment method available and for 17% they used invoices when they wanted to buy something but did not have the money for it. 11% stated they use invoices when the purchase was made on behalf of someone else or for their employer, and three said they only use them when buying expensive items. Finally, 51% of the respondents claimed that there has been at least one occasion that they have not fulfilled a purchase due to limited payment methods (Figure 2). Figure 2: Has it happened that you have not fulfil ed a purchase due to limited payment methods? 400 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 5.3 Financial responsibility The part of the survey regarding financial responsibility consisted of seven questions. The first three questions asked about the respondent’s online purchase behaviour. 61% of the respondents make online purchases 1-3 times a month, and 23% make online purchases a couple of times per year. Only one person never makes online purchases and only one person makes online purchases a couple of times per week. 37% of the respondents always checked their bank account balance before they make the purchase (Figure 3). Figure 3: Do you check your balance before making an online purchase? Respondents were asked if they borrow money from friends and family. Only one person answered that they do it often, approximately 16% answered that it can happen, but rarely, and the remaining 82% claimed that it never happens. The 18% of the people who say they borrow money from friends and family stil claim to have a buffer with easily accessible money, of more than 500 USD (5,000 SEK). In total, 79% of the respondents have a buffer of over 500 USD with 18% have a buffer of under 500 USD and only 3% answered that they do not have a buffer at all. In the final question the respondents ranked their financial responsibility from 1-10. The mean answer was 7.2 (Figure 4). Comparing to the previous answers of the respondents who considered themselves less financial y responsible (3 or lower), none of them uses “Klarna – Pay in 30 days” or Invoice due to not having enough money at the time of purchase. Only one person that answered that they never or rarely check their bank account before purchase (Figure 3) also answered that they are less financial y responsible (3 or lower). E. Rosenlind, U. Söderström, O. Norberg, H. Cripps, T. Mejtoft: Young Adults’ Attitude Towards Digital Payment Methods and Financial Responsibility 401 Figure 4: Do you consider yourself as financial responsible? 5.5 Follow up interviews Of the respondents 14 participants chose to enter their email addresses and therefore volunteered to be a part of the follow up interviews. After analysing their results, four people who were representative of the range of survey respondents were contacted and asked to attend an interview. A semi-structured interview was conducted where the questions were adapted to the participants’ previous answers. Some of the questions that were asked were: • How often do you make impulse purchases? • Develop further what makes you prefer Klarna/Swish/Card Payment. • “Klarna – Pay in 30 days” and invoice work basical y the same way. What makes you prefer “Klarna – Pay in 30 days”? • How do you keep track of when it is time to pay for your purchase? • How do you keep track of your purchases? • Tell more about the situation where you did not proceed with your purchase due to limited payment methods. The common findings from the interviews, which corroborates the results of the survey, was that people prefer to use Swish since it is easy, feels safe and the money is drawn from the account straight away. The major reason for preferring “Klarna – Pay in 30 days” was the shortage of money at the time for the purchase and the possibility of receiving the item before payment was required. One respondent stated that it is easier to decide whether you want the item since you can see and feel it before paying. But the same person also highlighted that they often forget to return 402 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY items or, decide to keep them, even though they might not need or want them, due to the effort required to return the item. One respondent expressed that he easily loses track of how much he spends as it is so easy to purchase with Klarna. He said that he often buys takeaway food with Klarna since it is the fastest way of paying when you are on the go, and sometimes they only offer “Klarna – Pay in 30 days”, not direct payment. Another person preferred “Klarna – Pay in 30 days” over invoice since Klarna does not charge any extra cost and you pay for your purchase through the app or website directly, without filling in any invoice details. The same person also mentioned the difficulties of keeping track of when the pay date expires on an invoice as a reason for using “Klarna – Pay in 30 days” and appreciated the notices from Klarna as he considered them crucial to remember to complete the payment. All four participants expressed strain of keeping track of purchases and getting an accurate overview of their finances since purchases can be spread out over different platforms and payment services. 6 Discussion When analysing the results, Swish is currently the most popular payment method among young adults in Sweden. Swish was chosen due to its simplicity, flexibility, and the fact that the money is drawn from the account immediately. Card Payment is the second most popular payment method because of reliance and simplicity, with the addition that it is mainly preferred if the user already has the card details saved in the browser. The third most popular payment method is “Klarna – Pay in 30 days” and it was chosen due to the flexibility when making returns and not needing the money at the time of the purchase (Senali, et al., 2022; Kumar et al., 2023). However, it needs to be taken into consideration that almost al the respondents were students. Students may be less likely to make as many online purchases as someone who works, but at the same time, students may be more versed in using digital payment methods with is consident with previous research (Yuk et al., 2021; An et al., 2023). Students also may be more aware of their private economy since they are on a strict budget. To answer these questions and make a more accurate study, the number of respondents would need to increase, but also expanded with a greater variation in occupation. This could be done by not only spreading the survey to university students but also reaching out to young adults working. Common to E. Rosenlind, U. Söderström, O. Norberg, H. Cripps, T. Mejtoft: Young Adults’ Attitude Towards Digital Payment Methods and Financial Responsibility 403 the three preferred payment methods mentioned is that they al use the third-party BankID to complete the payment this is consistent with previous research that has identified security as a key determent of the use of digital payments (Patil et al., 2017; van der Cruijsen and Plooij, 2018; Senali, et al., 2022). BankID is an important part of the payment process since it handles security in a much more efficient and fast way than using e.g., a specific security device. One could therefore argue that it is not the paying method itself that needs to be considered in the paper, but the whole purchase and payment process. Regarding the second research question, there is not enough evidence to conclude whether these payment methods can affect financial responsibility among young adults. To investigate this matter further, additional questions regarding purchase behaviour and financial responsibility need to be included in further research. Another obstacle when gathering information for the second research question is that the respondent’s idea of how financially responsible, they are, does not always align with the reality of how they act as found by Garbinsky et al., (2021). Regardless, one interesting finding was that the motivation to complete a purchase decreased if the choice of payment methods was limited (van der Cruijsen and Plooij, 2018; Senali, et al., 2022). If the respondent never completed the purchase, this could suggest that the payment methods that are not as fast and flexible may make the customer more thoughtful when making a purchase (Coffey et al., 2023). However there is a point where the purchase is so important that this over rides any difficulties with the payment method. If this is some sense of financial responsibility by deferring discretionary purchases, or just laziness, it is hard to tell. Since no follow-up questions were asked about the situation, this is a subject for further research. Another interesting finding was the struggle of getting an overview of purchases when they are spread out over different types of platforms and payment methods. Further research is required to determine if it is not the payment method itself that affects the financial responsibility among young adults, it is the situation and combination of different payment methods available at the time of purchased (Senali, et al., 2022). As this insight is drawn from the four people during the follow-up interviews it would be necessary to interview more people and add questions about the matter to any future survey. One conclusion that can be made is that “Klarna – Pay in 30 days”, which was the third most popular payment method, 404 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY creates a way to temporarily avoid financial responsibility. By being able to buy things for money that you do not have, you avoid the fundamental idea of financial responsibility, which is about giving up today’s desires to be able to provide for the needs of tomorrow. 7 Conclusion This paper aims to investigate which payment method young adults prefer and use today and if the payment method has any impact on their financial responsibility. This study shows that Swish is currently the most preferred payment method among young adults in Sweden. Whether the payment method affects financial responsibility among young adults cannot be entirely determined by the result of the study. Still, the results show that an individual’s choice of payment method can play a role in their financial responsibility. Some respondents highlighted that “Klarna – Pay in 30 days” has a (negative) impact on their financial responsibility since it is easy to use money that they do not have. The findings of the study also implicate that it can be the combination of different payment methods that affect young adults’ financial responsibility. References Aldrich, M. (2011). Online Shopping in the 1980s. Annals of the History of Computing, 33(4), 57–61. An, Q., Hong, W. C. H., Xu, X., et al. (2023). How education level influences internet security knowledge, behaviour, and attitude: a comparison among undergraduates, postgraduates and working graduates. International Journal of Information Security, 22, 305–317. https://doi.org/10.1007/s10207-022-00637-z Apple. (n.d.). Apple Pay. Available: https://www.apple.com/apple-pay/ Arvidsson, N. (2019). Building a Cashless Society: The Swedish Route to the Future of Cash Payments. Springer. BankID. (n.d.-a). Mobilt BankID, Available: https://support.bankid.com/sv/bankid/mobilt-bankid BankID. (n.d.-b). Vad är BankID?. Available: https://support.bankid.com/sv/bankid/vad-aer-bankid Coffey, J., Senior, K., Haro, A., et al. (2023). Embodying debt: youth, consumer credit and its impacts for wellbeing. Journal of Youth Studies. https://doi.org/10.1080/13676261.2022.2162376 Eriksson, B., & Sandhil , U. (2019). Cashless: A dead end for Sweden?. In A. Larsson & R. Teigland (Eds.), Digital Transformation and Public Services (pp. 235–242). Routledge. Ferrera, C., & Kessedjian, E. (2019). Evolution of E-commerce and Global Marketing. International Journal of Technology for Business, 1(1), 33-38. Garbinsky, E. N., Mead, N. L., & Gregg, D. (2021). Popping the positive illusion of financial responsibility can increase personal savings: Applications in emerging and western markets. Journal of Marketing, 85(3), 97-112. Hansson, A., (2022). Betala räkning – Hur betalar man en faktura? Available: https://buffert.se/betala- rakningar-fakturor/ Insight Intelligence. (2020). Sverige betalar 2020. Available: E. Rosenlind, U. Söderström, O. Norberg, H. Cripps, T. Mejtoft: Young Adults’ Attitude Towards Digital Payment Methods and Financial Responsibility 405 https://www.svenskhandel.se/globalassets/dokument/aktuel t-och-opinion/rapporter-och-foldrar/betalfragor/sverige-betalar-2020.pdf Klarna. (2022). About Klarna. Available: https://www.klarna.com/international/about-us/ Kumar, S., Prashar, S., & Shah, A. (2023). Creating brand love for payment apps through emotions. Marketing Intelligence & Planning, 41(3), 279–292. https://doi.org/10.1108/MIP-07-2022-0314 Lauer, J. (2020). Plastic surveillance: Payment cards and the history of transactional data, 1888 to present. Big Data & Society, 7(1), 1–14. https://doi.org/10.1177/2053951720907632 Leblebici, H. (2012). The evolution of alternative business models and the legitimization of universal credit card industry: Exploring the contested terrain where history and strategy meet. In S. J. Kahl, B. S. Silverman, & M. A. Cusumano (Eds.) History and Strategy (Advances in Strategic Management, Vol. 29) (pp. 117–151). Emerald Group Publishing Limited. https://doi.org/10.1108/S0742-3322(2012)0000029009 McFerrin, J. (2021). The history of ecommerce: from its origins to modern day. Available: https://www.iwdagency.com/blogs/news/the-history-of-ecommerce-how-did-it-all-begin Page, G., (2021). Focus: Evolution of payment methods. Available: https://cigp.com/insights/evolution-of-payment-methods Patil, P.P., Dwivedi, Y.K., Rana, N.P. (2017). Digital Payments Adoption: An Analysis of Literature. In: Digital Nations – Smart Cities, Innovation, and Sustainability, I3E 2017. Lecture Notes in Computer Science, vol 10595. Springer. https://doi.org/10.1007/978-3-319-68557-1_7. Senali, M. G., Cripps, H., Meek, S., & Ryan, M. M. (2022). A comparison of Australians, Chinese and Sri Lankans’ payment preference at point-of-sale. Marketing Intelligence & Planning, 40(1), 18–32. Swedbank, (n.d.). Handla på nätet med kort. Available: https://www.swedbank.se/privat/kort/sa-anvander-du-ditt-kort/handla-med-kort-pa-natet.html Swish (n.d.). About Swish. Available: https://www.swish.nu/about-swish Swish. (2022, June). Swich statistics – June 2022. Available: https://assets.ctfassets.net/zrqoyh8r449h/4oohpQ2xPiUFJFWZLwifZ/fd1be4bfd8d10dd65 c20c64157f38b6f/Swish_statistics___downtime_june_2022.pdf Tatum, M., (2022). What is financial responsibility? Available: https://www.smartcapitalmind.com/what-is-financial- responsibility.htm van der Cruijsen, C., & Plooij, M. (2018). Drivers of payment patterns at the point of sale: stable or not?”, Contemporary Economic Policy, 36(2), 363–380. https://doi.org/10.1111/coep.12245 Wisehn, I., (2019). Sveriges första kreditkort, Available: https://www.foretagskallan.se/foretagskallan- nyheter/lektionsmaterial/sveriges-forsta-kreditkort/ Yuk, M. T., Ka, Y. C., Luchen, H., Yun, K. I., & Wan, Y. (2021). Financial innovation in digital payment with wechat towards electronic business success. Journal of Theoretical and Applied Electronic Commerce Research, 16(5), 1844–1861. https://doi.org/10.3390/jtaer16050103 406 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY DIGITAL COACHING MOTIVATING TOWARDS PHYSICAL ACTIVITY IN JAPAN DURING THE COVID-19 PANDEMIC EEVA KETTUNEN,1 KRISTIINA JOKINEN,2 GENTIANE VENTURE,3 WILL CRITCHLEY,4 LAURI FRANK1 1 University of Jyvaskyla, Faculty of Information Technology, Jyvaskyla, Finland eeva.k.kettunen@jyu.fi, lauri.frank@jyu.fi 2 National Institute of Advanced Industrial Science and Technology, Tokyo, Japan kristiina.jokinen@aist.go.jp 3 University of Tokyo, Graduate School of Engineering, Tokyo, Japan venture@g.ecc.u-tokyo.ac.jp 4 Jyvaskyla, Finland wcritchley@gmail.com COVID-19 influenced peoples lives enormously and its impact was also seen in physical activity and exercising. This mixed methods intervention study highlights some of the impact the pandemic had on the physical activity and exercising behaviors of 10 Japanese information systems employees and students. This study also focuses on the influence that using sport and wel ness technology digital coaching had on peoples’ physical activity, exercise motivation and exercise self-efficacy during the pandemic. Based on the findings digital coaching can have motivational elements to support people towards a more physical y active lifestyle for example by making users more aware of their own physical activity. However, digital coaching is Keywords: not enough if users do not have initial motivation which may be digital coaching, understandable in the chal enging times of the COVID-19 exercise pandemic. The findings highlight that digital coaching could be self-efficacy, further developed to better support its users. This study and the COVID-19, Japan, results bring insight for digital coaching developers and users as Bled wel as people working in health care field. eConference DOI https://doi.org/10.18690/um.fov.4.2023.26 ISBN 978-961-286-751-5 408 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Japan was the second country outside of China to have an identified case of COVID-19, in early 2020. Japan's response was considered to be fairly rapid and organized. There was no formal lockdown, as the Japanese government has no power to force businesses to close and people to stay inside. However, local States of Emergency were declared to encourage businesses to close and people to stay home (BBC, 2020). The government encouraged citizens to avoid "the 3Cs": closed spaces, crowded places, and close-contact settings. Despite there being no actual punishment for not following these guidelines, it was recognized that the Japanese populace were general y fol owing them, and thus, crowds of many forms were significantly reduced for much of 2020 (Sayeed & Hossain, 2020). Understandably, these kinds of recommendations could have had a significant impact on sport and exercise participation. In Japan many people use smartwatches and activity trackers to track and measure their daily lives and various activities (Yoshida et al., 2018). Based on the article of IT Media Mobile (2022) 38% of the Japanese population aged between 20-69 have a smart watch. Interestingly, 41.3% of those smart watch owners bought them in 2020. Health management was the top reason for buying a smart watch (IT Media Mobile, 2022). This can be interpreted as showing that the effect of COVID-19 changed people's mind about health management since 29.6% of those who have a smart watch bought it for health management reasons (IT Media Mobile, 2022). This article focuses on exploring the usage and effects of sport and wel ness technology digital coaching in Tokyo, Japan during spring 2020. The aim is to see how the start of the COVID-19 pandemic influenced the exercise behavior of 10 Japanese information systems (IS) field employees and students, how these participants were able to adapt and use sport and wel ness technology digital coach into their lives and how the usage of a digital coach influenced their exercise self-efficacy and motivated them toward physical activity and exercise during the beginning of the pandemic. This study brings insight for digital coaching developers, digital coaching users and for people working in health care field. However, because of the limited number of participants due to the restrictions of the pandemic, the purpose of this article is more to highlight interesting aspects of digital coaching usage during the time of the pandemic and suggest potential areas for future studies. This study is unique because it was done in a cultural y different environment E. Ket unen, K. Jokinen, G. Venture, W. Critchley, L. Frank: Digital Coaching Motivating Towards Physical Activity in Japan During the Covid-19 Pandemic 409 compared to other studies of the same topic, as wel as being conducted during a pandemic. 2 Theoretical background 2.1 Sport and wellness technology digital coaching The maturing of the consumer market for sport and wel ness technology has meant that there are more lower-cost devices of high-quality available, which has also created an opportunity to do human-centered research in this area (James, 2017). Sport and wel ness technology has been shown to increase levels of physical activity in part by the information it can provide a user on their own physical activity and exercising habits (e.g. Larset et al., 2019; Romo et al., 2019). This enhanced awareness of one’s own physical activity and exercising habits has the added benefit of being a motivating factor for some users (e.g. Chan et al., 2004; Faghri et al., 2008). Sport and wel ness technology can also contribute to the goal setting process (e.g. Gordon et al., 2008), both by setting goals and providing feedback on the progress towards those goals, both of which make users more goal-oriented (e.g. Kari et al., 2016). There is, however, not necessarily always a connect between increased awareness of a user’s own physical activity and sustained use of a sport and wellness technology device (Miyamoto et al., 2016). Typical sport and wellness technology that has already been available tend to focus only on providing information and data about previous performances, rather than giving meaningful feedback or guidance on what to do next. Without meaningful or personalized data, a person would need specialized knowledge or professional help to make appropriate decisions in relation to their future exercise plans (Duking et al., 2016). One solution for this demand is digital coaching. Sport and wel ness technology digital coaching refers to a “service on a technological device that not only gives feedback but also offers advice, suggestions and future steps for a user to follow in the pursuit of their wellness and fitness goals” (Kettunen & Kari, 2018 p.3). Typically sport and wel ness technology devices and applications give only performance data and feedback. In addition to this, a digital coach creates a personalised training plan which is continually updated based on the user’s actions (Schmidt et al., 2015). It is important to acknowledge that digital coaching does not refer to the use of digital tools by human coaches, rather, it refers to a device or a solution which in itself is 410 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY the coach and functions independently without human interaction. The potential of sport and wel ness technology digital coach has been recognized in previous IS research (e.g. Kranz et al., 2013; Kettunen et al. 2021) and interest toward more instructional and personalized interaction has become more popular (Boulos & Yang, 2020; Mezei et al., 2020). However, commercial sport and wellness technology digital coaching devices and applications are stil relatively new and therefore only a few studies have focused on studying the usage experience and the motivational influence of the use of digital coach. Previous studies have suggested that the most suitable processes for digital coaching are related to behavior change techniques, goal setting persuasion, evaluation, interaction and co-creation Chatterjee et al. (2021) whereas the key element influencing adherence and usability was personalization (Lentferink et al., 2017; Chatterjee et al. 2021). Users seem to be more engaged with digital coaches that take into account the psychological needs related to motivation (Sundar, 2012; Chang et al. 2016). 2.2 Exercise Self-efficacy The quantitative part of this study is based on the theory of self-efficacy by Albert Bandura (1977). The concept of self-efficacy refers to a person’s own beliefs about their own capacity related to performing a specific task. It is important to recognize that a person’s self-efficacy does not necessarily correlate to his/her actual capacity to perform a specific task. A person with high levels of self-efficacy is more likely to view a difficult task as an opportunity and a challenge whereas a person with low self-efficacy easily tends to avoid difficult tasks (Bandura 1997). Therefore, it is understandable that self-efficacy can also affect a person’s motivation by impacting the amount of effort they are wil ing to give to overcome a specific task. According to Bandura (1998) there are four main sources of information that affect person’s self-efficacy: vicarious experiences, performance accomplishment, verbal persuasion, and physiological states. Since in this study the topic is physical activity and exercise, the theory of self-efficacy is applied to exercise and physical activity and therefore is cal ed exercise self-efficacy. Self-efficacy plays an important role in exercise since it has been shown to have a high influence in the adoption of physical activity habits (McAuley & Blissmer, 2000). The importance of self-efficacy tends to be even more important in the phase when physical activity has not yet become a habit (Bandura, 1986). Self- E. Ket unen, K. Jokinen, G. Venture, W. Critchley, L. Frank: Digital Coaching Motivating Towards Physical Activity in Japan During the Covid-19 Pandemic 411 efficacy has also been associated with maintaining long-term physical activity (McAuley et al., 2011). People with high levels of self-efficacy may work harder and participate in physical activity more frequently (Bandura, 1986). Self-efficacy is one of the most researched concepts in the field of physical activity and exercise (Krol et al., 2007), and is a well-known theory when studying motivation and self-confidence in the field of sport performance (Feltz, 1988). In this paper, exercise self-efficacy is studied from the perspective of sport and wel ness technology digital coaching and the aim is to see whether the use of a digital coach can influence the users’ exercise self-efficacy during the COVID-19 pandemic and despite the issues and restrictions the pandemic caused related to physical activity and exercising. 3 Methodology 3.1 The Digital Coach Used in the Study The digital coach used for this study was the Suunto 3 Fitness sports watch, created by Finnish company Suunto OY (Suunto, 2019). The device is designed as a training watch for exercisers and is particularly tuned for aerobic exercising such as walking, running or cycling. The watch has features related to exercise timing, wrist-based heart rate monitoring, sleep monitoring, GPS, and 24/7 activity, stress and recovery tracking. The features that make the watch a digital coach is its “adaptive training guidance” and “real time feedback”. The watch creates a personalised weekly training plan based on the goal user has set. An estimation of the user’s fitness level is made using the existing training data. On a training day, the watch provides the workout goals such as the time and the intensity. The digital coach provides real-time guidance during the recommended workout. The guidance is primarily related to staying in a particular heart rate zone. The watch also provides instructions through audio and visual indicators to help the user to keep up with the set goal. After the workout the device provides information about the workout. The training guidance will adapt if a person misses or modifies a workout. 3.2 Research approach, data col ection and analysis The study was a mixed methods intervention study that took place in Tokyo, Japan, at the start of the COVID-19 pandemic from February to April 2020. The intervention lasted three months and included eight male and two female 412 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY participants, ages 22-67, who were students or employees working in IS field. The volunteers were recruited from the National Institute of Advanced Industrial Science and Technology and from Tokyo University of Agriculture and Technology. The participants had different types of physical activity backgrounds. The physical activity background was not a selection criterion. Participants did not have previous experience related to digital coaching although some of them had previous experience on of sport and wel ness technology. In the beginning of the intervention all participants were given quantitative paper questionnaires related to exercise self-efficacy and attitude towards sport and wel ness technology digital coaching. After answering the questionnaire al participants were given a sport and wel ness technology digital coach to use for the duration of the study. The participants were asked to use the digital coach as they felt most suitable for them. In the end of the intervention period the participants were asked to answer again the same questionnaires they answered in the beginning of the intervention. In this second round the questionnaire was an online questionnaire due to the COVID-19 restrictions. The purpose of this quantitative data was to see if there were any changes in their answers due to the intervention. After the intervention the participants were also interviewed about their experiences related to physical activity during COVID-19 and their usage of the digital coach to support their physical activity and exercising. Due to COVID-19 restrictions the qualitative data was collected via online and email interviews in the end of the intervention in early May 2020. The interviews were semi-structured interviews which are the most used interview type in IS qualitative research (Myers & Newman, 2007), consisting of themes related to exercise background, COVID-19 pandemic, adaptation of digital coach, digital coach’s influence on physical activity and exercise and ideal digital coach. Interviews done over email followed a more structured format. The quantitative data was analyzed using Excel. Due to the low number of participants the purpose of the quantitative data analysis was not to provide statistically significant information but instead, by analyzing averages on an item by item basis, to highlight possible trends and areas worth researching more deeply. The qualitative data was analyzed using thematic analysis method which is meant for “analyzing, identifying and reporting patterns within data” (Braun and Clarke, 2006, p. 79) and is widely used in qualitative research (Guest et al., 2012). The analysis E. Ket unen, K. Jokinen, G. Venture, W. Critchley, L. Frank: Digital Coaching Motivating Towards Physical Activity in Japan During the Covid-19 Pandemic 413 process of the studies began with transcribing the relevant parts of the interviews and becoming familiar with the data. The interview themes were already divided into smal er sections in the interview phase During the analysis the answers of the individual participants were compiled into an Excel spreadsheet. This made it easier to see the occurring themes and compare the data. The most highlighted issues and topics were presented in the result section. 4 Findings 4.1 Quantitative findings related to exercise self-efficacy and attitude Exercise self-efficacy was measured using a scale by Kroll et al. (2007). The scale, presented in table 2 below, included statements regarding personal abilities related to physical activity. The self-efficacy was measured on the scale from 1-4 where 1=not true at all, 2= seldom true, 3=somewhat true and 4= completely true. As seen from the table 2 there was no significant changes when comparing the means of the questionnaires before and after the intervention. One statement that seemed to have the biggest change between the two data points was related to exercising when feeling depressed. It seemed that after the intervention the participants were less certain that they are capable of being physical y active and exercise when feeling depressed. The second part of the questionnaire, presented in table 3 below, focused on self-efficacy related to the participants’ overall opinions about their exercising and improving their own fitness. This part included 13 statements, of which six were related to the role of sport and wel ness technology. The statements were measured on a scale of 1-7 where 1 represented “strongly disagree” and 7 represented “strongly agree”. Based on the results participants seemed less confident at the end of the intervention that they are able to train independently without any guidance or coaching. However, they also found it less hard after the intervention to find out how to improve or analyze their own aerobic fitness. Also the belief of the truthfulness of the information provided by sport and wel ness technology was increased during the intervention. 414 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 1: Changes in exercise self-efficacy Start End Statement (Mean) (Mean) I can overcome barriers and chal enges with regard to PA and 3,1 3,3 exercise if I try hard enough I can find means and ways to be physical y active and exercise 3,4 3,5 I can accomplish my PA and exercise goals that I set 3,1 3,4 When I am confronted with a barrier to PA or exercise, I can find 2,8 3,0 several solutions to overcome this barrier I can be physical y active or exercise even when I am tired 2,4 2,2 I can be physical y active or exercise even when I am feeling depressed 2,7 1,9 I can be physical y active or exercise even without the support of 3,1 3,0 my family or friends I can be physical y active or exercise without the help of a therapist 3,3 3,4 or trainer I can motivate myself to start being physical y active or exercising 3,1 3,0 again after I've stopped for a while I can be physical y active or exercise even if I had no access to a 2,7 3,2 gym, exercise, or rehabilitation facility Attitudes towards using a digital coach was measured by five statements shown in table 4 below. The statements focused on the overal attitude (bad vs. good), the experimental aspect (unpleasant vs. pleasant and uncomfortable vs. comfortable), and the instrumental aspect (useless vs. useful and foolish vs. sensible). As can be seen from the means, participants’ attitudes towards digital coaching remained relatively similar. However, after the intervention the participants considered using digital coaching more uncomfortable than before. E. Ket unen, K. Jokinen, G. Venture, W. Critchley, L. Frank: Digital Coaching Motivating Towards Physical Activity in Japan During the Covid-19 Pandemic 415 Table 2: Changes related to improving fitness, and sport and wel ness technology Start End Statement (Mean) (Mean) I know how to create myself an exercising program 4,0 4,3 I need help in creating myself a suitable exercising program 4,1 4,3 I am able to train independently without any guidance or coaching 5,1 4,0 Sport and wel ness technology has an important role in my 4,7 4,6 exercising Sport and wel ness technology provides me with important 5,3 5,6 information that I can use in my exercising I am able to improve my fitness with the help of sport and wel ness 5,3 5,5 technology I believe that sport and wel ness technology provides me with 5,5 5,9 reliable information regarding my own exercising I believe that sport and wel ness technology provides me with 5,3 5,6 accurate information regarding my own exercising I believe that sport and wel ness technology provides me with truthful information regarding my own exercising 5,1 5,9 It is hard for me to find out how to improve my aerobic fitness 4,9 4,1 I do not know how to increase the level of my aerobic fitness 4,3 4,2 It is hard for me to analyze my aerobic fitness 5,0 4,0 I know how to improve my aerobic fitness 4,1 4,6 416 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 3: Changes in attitude towards using a digital coach Start End Statement (Mean) (Mean) I know how to create myself an exercising program The thought of using a digital coach as a support for my training sounds: bad 5,5 5,7 (1) vs. good (7) The thought of using a digital coach as a support for my training 5,7 5,7 sounds: useless (1) vs. useful (7) The thought of using a digital coach as a support for my training 6,0 5,7 sounds: foolish (1) vs. sensible (7) The thought of using a digital coach as a support for my training 5,7 5,4 sounds: unpleasant (1) vs. pleasant (7) The thought of using a digital coach as a support for my training sounds: uncomfortable (1) vs. comfortable (7) 5,6 4,9 4.2 Qualitative findings 4.2.1 COVID-19 influencing exercise behavior The exercise background of the participants varied. Four participants had walking and commuting to work/school as their only exercise and other participants reported also having exercise related hobbies such as strength training, dancing, running or team sports. During the COVID-19 pandemic their exercise behavior changed significantly due to the restrictions. Exercise related hobbies that were held in public places such as gyms and exercise arenas were cancel ed causing a decrease in physical activity level for many participants. Since people worked from home, so the amount of walking decreased from the elimination of commuting. Some participants continued or started running during the pandemic but even running was considered more difficult due to big crowds outside. As one participant stated “I avoided people jams. For instance I started running around the park near home at midnight”. In general participants reported that COVID-19 had a negative influence on their physical activity as well as on their exercise motivation: “I have no motivation to go out anymore”(Male, 25 years), “The frequency of training has decreased”(Male, 24 years). Despite the restrictions and limitations, some participants found a way to remain physical y active: “I decided to do longer walk exercises as my commuting activity disappeared” (Female, 46 years), “I became more E. Ket unen, K. Jokinen, G. Venture, W. Critchley, L. Frank: Digital Coaching Motivating Towards Physical Activity in Japan During the Covid-19 Pandemic 417 motivated to exercise since I felt stressed at home. Exercise became my new routine.” (Male 22). 4.2.2 Adaptation and usage of the digital coach None of the participants had used a sport and wel ness technology digital coach before but a couple of them had used a pedometer or a running application. Most participants had either no expectations or high expectations towards digital coaching hoping that it would guide and encourage them to exercise. Participants felt that starting to use the digital coach was relatively easy. Most wore the digital coach every day during the intervention although some participants felt uncomfortable wearing it overnight. Few participants wore the digital coach only during exercising. The lock down and working remotely affected some participants’ usage of the digital coach: “I used the digital coach al day before I started working from home. After that I wore it occasional y when I went to sleep or for a walk” (Female, 46 years). The feature participants used the most were heart rate, step counting, sleep tracking and exercise tracking. Only a few participants followed the training program offered by the adaptive training guidance since participants thought it was hard to understand or that they would have liked to schedule their training themselves. As some participants noted: “I did not use the adaptive training guidance. It looked rational but I would have liked to schedule the training myself” (Male, 24 years), “I did use it at first but it was a little too hard to understand” (Male, 24 years) The real time feedback feature was perceived somewhat more useful since half of the participants used it. The usage of the digital coach changed as the pandemic continued. Some participants decreased the usage since their exercise levels decreased. 4.2.3 Motivational influence of the digital coach After the intervention most participants reported becoming more aware about their physical activity or lack of physical activity. Most participants had also learned something new about their own physical activity, such as their heart rate levels. As one participant stated: “I realized my heart rate is quite high. I also realized how much I walk every day, I walk more than I thought” (Male, 28 years). Because people were working more from home, some participants found more time to exercise. A few participants started running and felt that having a digital coach along their exercising increased motivation and made running more regular and consistent. 418 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Participants also perceived seeing their exercise progress as wel as seeing their daily activities motivational. According to one participant: “Using the digital coach has improved my health since it gives me numbers and I can set goals. This increases my motivation which will improve my health” (Male, 22 years). COVID-19 decreased the exercise motivation for some participants, and the use of a digital coach was not a good enough motivator to start increasing their exercise levels. According to one participant: “Being unable to exercise outside reduced my motivation” (Male, 22 years). Using the digital coach encouraged some of the participants to make some changes to their exercise and wel being. Some participants reported having made changes to their training and health management due to the usage of the digital coach: “ I learned that I exercised too hard, so I reviewed my training plan” (Male, 24 years), “ Because of using the digital coach, I am now trying to keep an ultra slow pace in early stage of running” (Male 67 years),“Before I thought I slept more but after using the watch I tried to go to bed earlier and have better sleep quality by calming down before going to bed” (Male, 22 years). In general participants found it motivational to receive exercise and health related data as wel as being able to compare the progress. As one participant noted: “Digital coach assists on exercising but the responsibility is on the person. That makes it easy to start with the digital coach” (Male, 28 years). For some participants tracking exercise became fun and easy: “By making it appear numerical y, the exercise became fun and easy” (Female, 60). For some participants the purpose of using a digital coach changed during the intervention: “Digital coach usage has changed from exercise management to life management such as sleep management” (Male, 22 years). 4.2.4 Ideal digital coach Participants were also asked how to develop digital coaching to make it more motivational and suitable for them. Some participants felt the coach should be more encouraging, more clear, and more straight forward. The data should be more accurate and the device itself could be more comfortable. In general participants wanted their ideal digital coach to include features related to weight training, and calorie consumption. It should also connect more effectively with other technologically. In general, participants felt that digital coaching is suitable for different types of people but especial y for people who want to increase their exercise E. Ket unen, K. Jokinen, G. Venture, W. Critchley, L. Frank: Digital Coaching Motivating Towards Physical Activity in Japan During the Covid-19 Pandemic 419 level and learn more about their own physical activity and health. One participant stated: “Digital coach is for people who want to increase their exercise level and do not know a lot about exercising.” (Male, 22 years). When comparing a human coach to a digital coach, participants felt that a human coach can be more flexible, provide more detailed instructions and give more personal advice. A human coach would also possibly earn more respect and as a result motivate the client to work harder. One participant felt that: “With a human coach you have to show respect and work harder and do exercise” (Male, 25 years). The benefits of a digital coach were flexibility, cheaper price, ease of use and rich personal information. As one participant said: “You have more freedom to do things your way. It creates less stress. No schedules are needed and the response is fast” (Male, 28 years). 5 Discussion This study focused on exploring the usage, effect on exercise self-efficacy and the motivational influence of sport and wel ness technology digital coaching during the COVID-19 pandemic in Japan. Based on the findings it seems that the overall attitude towards physical activity and exercising also affected the usage of digital coach during COVID-19. The motivational influence of digital coaching seemed inconclusive. Whereas for some participants digital coach brought extra motivation, for others the COVID-19 restrictions decreased the exercise motivation such that the usage of digital coaching could not bring back the motivation. This result is consistent with previous research (Kettunen et al. 2021) suggesting that motivational elements of the digital coach are not enough if the user themselves are not initial y motivated to exercise. The quantitative findings highlight that digital coach usage can make people more confident about training without outside support and being more in charge of their own fitness development. This finding is also consistent with previous studies (Kettunen et al., 2019; Kettunen et al., 2021). However, it seems that the atmosphere created by COVID-19 has made people less confident that they can train when feeling depressed. The trust in the truthfulness of the data provided by the digital coach increased during the intervention. This finding is opposite to previous similar studies (Kettunen et al. 2019; Kettunen et al., 2020) done in Finland. The attitude towards digital coaching was relatively high already in the beginning of the study and remained about the same throughout the intervention. An exception to this was that people perceived using a digital coach as less comfortable after the intervention. 420 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY According to the findings the digital coach was perceived as interesting and useful in helping people learn about their physical activity as wel as increased the exercise motivation. Digital coaching also makes the training more visible and exciting to some participants by giving numbers and setting goals. Thus, a digital coach has potential in influencing exercise motivation even during difficult times such as during the COVID-19 pandemic. However, digital coaching could be developed much further to increase its motivational influence. For example, enhancing communication between the user and the device was highlighted. It is notable that most technology related to health improvement reaches only those who are already health-conscious, sporty, and active, so the problem seems to be rather how to reach those who do not care about their health. There does not appear to be a good solution for this. Based on this study digital coaching has potential and therefore might be the key to solve that problem, so it would be useful to conduct more studies with different target groups and using different types of digital coaches. It is also important to recognize that technology such as a digital coach may not be able to overcome the individual traits that guide a person’s behavior during extraordinary events like the COVID-19 pandemic was. Thus, the impact of a digital coach on a person’s exercise behavior might have been somewhat different in an otherwise more normal time period. 6 Limitations and suggestions for future research The results of the study are based on a relatively smal and homogenous target group and the usage of one particular digital coaching device. Due to the pandemic and the limited number of digital coaches it was not possible to extend the number of participants. It is also important to note that the participants were al working or studying in the information systems field which might have influenced the adoption of new technology. Since the quantitative data is based on a smal numer of participants the quantitative results serve as highlighting possible trends and areas worth researching more deeply. Doing research on a unique time such as during a pandemic made the topic more interesting but at the same time made analysis harder as it is hard to know for sure whether some results were more related to the usage of the digital coach or to the change in lives due to COVID-19. Future studies could focus on the influence that digital coaching usage has on exercise motivation and self-efficacy especial y. The usage of sport and wel ness technology digital coaching E. Ket unen, K. Jokinen, G. Venture, W. Critchley, L. Frank: Digital Coaching Motivating Towards Physical Activity in Japan During the Covid-19 Pandemic 421 in Japan could also be studied more broadly, since to our knowledge this study is among the few studies, if not the only one, conducted in Japan about the topic. Acknowledgements Thank you to AIST and TUAT staff and students for successful and pleasant col aboration throughout the research and for volunteering as participants for the study. Special thank you to Koki Ijuin for sharing his knowledge, assistance and enthusiasm. References Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review 84 (2), 191-215. Bandura A. (1986). Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, NJ: Prentice Hall. Bandura, A. (1998). Health promotion from the perspective of social cognitive theory. Psychology and Health 13 (4), 623–649. Boulos, M. N. K., & Yang, S. P. (2021). Mobile physical activity planning and tracking: a brief overview of current options and desiderata for future solutions. Mhealth, 7. Braun, V., and V. Clarke (2006). “Using thematic analysis in psychology,” Qualitative research in Psychology 3 (2), 77-101. BBC. (2020). Coronavirus: State of emergency lifted in most of Japan. BBC News. Retrieved September 29, 2022, URL: https://www.bbc.com/news/world-asia-52658551 Chan, C. B., Ryan, D. A., & Tudor-Locke, C. (2004). Health benefits of a pedometer-based physical activity intervention in sedentary workers. Preventive Medicine, 39(6), 1215-1222. Chang, R. C. S., Lu, H. P., Yang, P., & Luarn, P. (2016). Reciprocal reinforcement between wearable activity trackers and social network services in influencing physical activity behaviors. JMIR mHealth and uHealth, 4(3), e5637. Chatterjee A, Gerdes M, Prinz A, Martinez S. 2021. Human Coaching Methodologies for Automatic Electronic Coaching (eCoaching) as Behavioral Interventions With Information and Communication Technology: Systematic Review J Med Internet Res, 23(3):e23533 Faghri, P. D., Omokaro, C., Parker, C., Nichols, E., Gustavesen, S., & Blozie, E. (2008). E-technology and pedometer walking program to increase physical activity at work. The Journal of Primary Prevention, 29(1), 73-91. Feltz, D. (1988) “Self-confidence and sports performance.” Exercise and Sport Sciences Reviews 16 (1), 423-458. Finnish Sports Federation. (2011). Kansallinen liikuntatutkimus 2009–2010: Aikuis- ja senioriliikunta. [National Sports Study 2009-2010: Adult and Elderly Physical Activity (Report), Helsinki: Finnish Sports Federation. Gordon, M., Althoff, T., & Leskovec, J. (2019). Goal-setting and achievement in activity tracking apps: a case study of MyFitnessPal. The World Wide Web Conference, USA, 571-582. Guest, G., MacQueen, K. M. and E. E. Namey (2012). Applied Thematic Analysis. Los Angeles: SAGE. IT Media Mobile. (2022) Smart watch ownership rate is 38.0%, top reason for purchase is "health management". URL: https://www.itmedia.co.jp/mobile/articles/2102/09/news086.html James, D. (2017). Wearable Technology in sport, a convergence of trends. Journal of Advanced Sport Technology, 1(1), 1-4. 422 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Kari, T., Koivunen, S., Frank, L., Makkonen, M., & Moilanen, P. (2016). Critical Experiences During the Implementation of a Self-tracking Technology. PACIS 2016: Proceedings of the 20th Pacific Asia Conference on Information Systems, Taiwan, 129. Kettunen, Eeva (2021) Using Digital Coaching to Promote Physical Activity to University Students with Low Levels of Physical Activity: A Qualitative Intervention Study. In Proceedings of the 54th Hawai International Conference on System Sciences, University of Hawai'i. Kettunen, E., and Kari, T. (2018). ”Can Sport and Wel ness Technology be My Personal Trainer?: Teenagers and Digital Coaching.” in Proceedings of the 31st Bled eConference. Digital Transformation: Meeting the Chal enges”, Bled, Slovenia, pp. 463-476 Kettunen, E., Kari, T., Makkonen, M., Critchley, W., & Sel , A. (2019) Digital Coaching among Physically Inactive University Students: A Quantitative Intervention Study on Exercise Self-efficacy. In Proceedings of the 32th Bled eConference. Humanizing Technology for a Sustainable Society. Kettunen, E., Kari, T., Makkonen, M., Frank, L. & Critchley, W. (2020). Young Elderly and Digital Coaching: A Quantitative Intervention Study on Exercise Self-Efficacy. In the proceedings of the 33rd Bled eConference : Enabling technology for a sustainable society. Maribor: University of Maribor, 469-484 Kranz, M., A. Möl er, N. Hammerla, S. Diewald, L. Roalter, T. Ploetz and P. Olivier (2013). “The Mobile Fitness Coach: Towards Individualized Skill Assessment Using Personalized Mobile Devices” Pervasive and Mobile Computing 9, 2013, pp. 203-215. Krol , T., Kehn, M., Ho, P. S., & Groah, S. (2007). The SCI Exercise Self-Efficacy Scale (ESES): development and psychometric properties. The international journal of behavioral nutrition and physical activity, 4, (34). Larsen, R. T., Christensen, J., Juhl, C. B., Andersen, H. B., & Langberg, H. (2019). Physical activity monitors to enhance amount of physical activity in older adults–a systematic review and meta-analysis. European Review of Aging and Physical Activity, 16(1), 1-13. Lentferink, A. J., Oldenhuis, H. K., de Groot, M., Polstra, L., Velthuijsen, H., & van Gemert-Pijnen, J. E. (2017). Key components in eHealth interventions combining self-tracking and persuasive eCoaching to promote a healthier lifestyle: a scoping review. Journal of medical Internet research, 19(8), e7288. McAuley E. & Blissmer B. (2000). Self-efficacy determinants and consequences of physical activity. Exercise and Sport Sciences Reviews, 28, 85–88. McAuley, E., Szabo, A., Gothe, N., & Olson, E. A. (2011). Self-efficacy: Implications for Physical Activity, Function, and Functional Limitations in Older Adults. American journal of lifestyle medicine, 5(4), Mezei, J., Sell, A., & Walden, P. (2020). Digital coaching-an exploratory study on potential motivators. Proceedings of the 53rd Hawai International Conference on System Sciences, USA. Miyamoto, S. W., Henderson, S., Young, H. M., Pande, A., & Han, J. J. (2016). Tracking health data is not enough: a qualitative exploration of the role of healthcare partnerships and mHealth technology to promote physical activity and to sustain behavior change. JMIR mHealth and uHealth, 4(1), Article e5. Romeo, A., Edney, S., Plotnikoff, R., Curtis, R., Ryan, J., Sanders, I., Crozier, A., & Maher, C. (2019). Can smartphone apps increase physical activity? Systematic review and meta-analysis. Journal of Medical Internet Research, 21(3), Article e12053. https://doi.org/10.2196/12053 Sayeed, U. B., & Hossain, A. (2020). How Japan managed to curb the pandemic early on: Lessons learned from the first eight months of COVID-19. Journal of global health, 10(2), 020390. URL: https://doi.org/10.7189/jogh.10.020390 Sundar, S. S., Bellur, S., & Jia, H. (2012, June). Motivational technologies: a theoretical framework for designing preventive health applications. In International conference on persuasive technology (pp. 112-122). Springer, Berlin, Heidelberg. Suunto, 2022 URL: https://www.suunto.com/en-gb/Support/Product-support/suunto_3/ suunto_3/ E. Ket unen, K. Jokinen, G. Venture, W. Critchley, L. Frank: Digital Coaching Motivating Towards Physical Activity in Japan During the Covid-19 Pandemic 423 Yoshida, Y., Nishimura, T., Jokinen, K. (2018). Biomechanics for understanding movements in daily activities. LREC Workshop "Language and Body in Real Life & Multimodal Corpora" (REAL-MM), 7-12 May 2018, Miyazaki, Japan. http://lrec-conf.org/workshops/lrec2018/W20/ index.html 424 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Appendix 1: Participant information Age Gender Employment Status Activity Level 22 Man Student Health enhancing participant 22 Man Student Health enhancing participant 24 Man Student Health enhancing participant 24 Man Student Fitness participant 25 Man Student Casual participant 28 Man Employed Inactive or sedentary 46 Woman Employed Inactive or sedentary 52 Man Employed Competition athlete 60 Woman Employed Health enhancing participant 67 Man Employed Fitness participant The Appendix 1. above shows more detailed information of the participants. The activity level categorization was collected using a categorization based on the Finnish National Sport Survey (Finnish Sports Federation, 2011), which classifies people into seven categories based on their PA level. These categories, presented in order from the most to least active, were: competition athletes, fitness athletes, fitness participants, physical y active for health, active in commuting and non-exercise, occasional y active, and inactive or sedentary. MOBILE CLINICAL DECISION SUPPORT SYSTEMS: A PATHWAY FROM DESIGN TO COMMERCIALIZATION NALIKA ULAPANE,1 ABDUR RAHIM MOHAMMAD FORKAN,1 PREM PRAKASH JAYARAMAN,1 PENELOPE SCHOFIELD,1 KATE BURBURY,2 NILMINI WICKRAMASINGHE1 1 Swinburne University of Technology, Melbourne, Australia nalika.ulapane@gmail.com, fforkan@swin.edu.au, pjayaraman@swin.edu.au, pschofield@swin.edu.au, nwickramasinghe@swin.edu.au Keywords: 2 Peter MacCallum Cancer Center, Melbourne, Australia clinical kate.burbury@petermac.org decision support A smartphone-based clinical decision support system (CDSS) systems, commercialization, has been designed for perioperative patient management in design cancer care. A systematic design approach to ensure fit-for- science purpose of such mobile CDSSs is lacking. This study attempts to research methodology, fil that void by reporting on the pathway we took from design perioperative, to commercialization. Our pathway is governed by the design surgery, science research methodology and the theory of task technology task technology fit. Our experiences are generalizable and can provide guidance fit, to many mobile clinical decision support solutions in healthcare. smartphone DOI https://doi.org/10.18690/um.fov.4.2023.27 ISBN 978-961-286-751-5 426 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Clinical decision making is a complex and cognitively demanding process. It involves the interplay between tacit and explicit knowledge and includes observation, information, knowledge, experience, caring, and incidental learning (Banning, 2008). Medical errors could result from shortcomings in clinical decision making (Makary & Daniel, 2016) and manifest as adverse health and cost implications. As such, effective and accurate clinical decision making becomes essential for quality healthcare delivery. Technology has been introduced to healthcare over the years to support clinical decision making. Clinical decision support systems (CDSSs) has been one such use of technology. CDSSs have advanced over the years. Starting from paper-based systems they have evolved to legacy-based computer systems (Skyttberg, et. al., 2016), and then more recently to handheld mobile device-based systems (Chahal, et. al., 2020). Latest technology advancements such as Industry 4.0 and Healthcare 4.0 have enabled this evolution. Following this backdrop, designing a Smartphone- based CDSS is the focus of this study. The introduction of mobile CDSSs to healthcare is not simple. A recent scoping review (Ulapane & Wickramasinghe, 2021) has listed some of the major issues encountered in the past five years in attempts to introduce mobile technologies to healthcare as follows: complexity and performance-related issues in the used technologies; difficulty to validate the efficacy of the introduced technologies; costs involved in introducing new technologies; lack of quality of data (when it comes to the use of Artificial Intelligence (AI)); lack of generalizability of certain techniques and technologies used, lack of expandability and scalability of the introduced technologies; lack of streamlining of the technologies with the clinical workflows; privacy and cyber security-related issues; surveillance capitalism; risks and accountability; policy and legislative challenges; slow adoption of certain technologies in healthcare; perceptions and biases of technology users and potential users; and competence (or lack of it) in technology usage among clinicians. The cal for better standardization of mobile technologies in healthcare has also emerged (Lee, et. al., 2018). N. Ulapane, et al.: Mobile Clinical Decision Support Systems: A Pathway from Design to Commercialization 427 The prevalence of diverse issues as noted above, indicates a void in this field. This void comes due to the lack of a systematic approach for the design and development of mobile CDSSs. Motivated by that void, this study attempts to answer the following research question: “How can mobile CDSSs be designed and developed to be of superior fit-for-purpose?” By answering the aforesaid research question this study makes a twofold contribution: A contribution to theory and a contribution to practice. The contribution to theory comes as a systematic design approach for mobile CDSSs. Our approach extends towards validation and commercialization. Our approach combines the theory of Task Technology Fit (TTF) (Goodhue & Thompson, 1995), the Design Science Research Methodology (DSRM) (Hevner & Chatterjee, 2010), and the input from a commercial software development partner. The contribution to practice is a designed and validated mobile CDSS. It includes a smartphone application and a web-based data analytics platform. This CDSS assists clinicians, specifical y doctors, and nurses, to manage surgery patients during perioperative care (i.e., pre-operative and post-operative stages). The CDSS supports with decisions like management of anticoagulant drugs. The specific health focus is prevention of thromboembolism. Thromboembolism is a condition of undesired blood clotting. It is a leading cause of death and complication in surgery patients (Chahal, et. al., 2020). This CDSS has been designed for a leading cancer hospital in Australia. This paper is arranged as follows: Review of related works; relevant theories; methodology; results; discussion, and conclusions. 2 Review of Related Work A literature search was carried out surrounding Smartphone-based CDSSs in perioperative care. The fol owing keyword search was done: ("perioperative" OR "surgery") AND "smartphone". The keyword search was done in the Google Scholar database. Google Scholar is accessible to the public free of charge. Almost all academic databases are enlisted in Google Scholar. The search was done between May 1st and May 5th of 2022. The search was limited to items written in English. The items that included the keywords within the item’s title were considered. Works published since 2021 were considered to capture the latest results. Twelve articles got retrieved matching the search criteria. Our search is deliberately restrictive and 428 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY thus may be incomplete. Our purpose here was to scan the latest literature within the previous year or so for a snapshot of the latest works. The aims and objectives of the retrieved works were reviewed. We noticed that none of the retrieved works had focused on proposing a systematic design approach. As such, our study contrasts those works. The works (Ahmad, et. al., 2022), (Awaludin, et. al., 2022), (Boaro, et. al., 2021), (Panda, et. al., 2022), (Soangra & Lockhart, 2021), (van den Berg, et. al., 2022), (Voglis, et. al., 2022), and (Wu, et. al., 2022) have all reported some degree of design and development. Their primary focus is the technology solution rather than the design approach. As such, their design approaches are specific for their solutions. Such design approaches can be improved and generalized by grounding on theory. Therefore, our study offers an increment to current thinking as we propose a systematic design approach grounded on theory. Our approach can be replicated irrespective of the health or technology context. The works (Jones, et. al., 2021), (Kabbani & Kabbani, 2021), and (Lesher, et. al., 2021) were review articles. The need for a systematic design approach is further emphasized in them. Jones, et. al., (2021) concludes a lack of certification, validation and peer review of applications designed for plastic surgery in the UK. Kabbani & Kabbani, (2021) highlights the importance of codesign and cocreation of applications through collaboration with healthcare professionals. Moreover, Lesher, et. al., (2021) has discussed institutional and regulatory barriers to the adoption of mobile health (i.e., mHealth) applications. Such points complement our argument and reemphasize the need for a systematic design approach so that persistent barriers can be overcome and thereby enable smooth and seamless introduction of digital health solutions to healthcare contexts. 3 Relevant Theories Our attempt is to maximize the fit-for-purpose of mobile CDSSs. A well-known theory to assess fit-for-purpose is the theory of Task Technology Fit (TTF) (Goodhue & Thompson, 1995). We have adopted TTF in this work. Furthermore, our study involves designing an artifact through codesign. Therefore, the Design Science Research Methodology (DSRM) (Hevner & Chatterjee, 2010) is also followed. These theories are summarized in the following subsections. N. Ulapane, et al.: Mobile Clinical Decision Support Systems: A Pathway from Design to Commercialization 429 3.1 The Theory of Task Technology Fit (TTF) Stated in TTF is that Information Technology (IT) systems are likely to be more usable, desirable, and impactful, if the system’s capabilities match the tasks the user must perform (Goodhue & Thompson, 1995). Goodhue & Thompson, (1995) presented a list of factors to measure the influence of TTF on user performance. We have constructed the questionnaire in (TTF Questionnaire, 2022) based on that list. It is tailored for the users of mobile CDSSs. This questionnaire is used for validation of our artifact through user feedback. 3.2 Design Science Research Methodology (DSRM) DSRM (Hevner & Chatterjee, 2010) is a process for systematically creating an artifact so that the artifact’s desirability can be maximized by meeting stakeholder needs. The process includes six steps: (1) Problem identification and motivation; (2) Defining the objectives for a solution; (3) Design and development; (4) Demonstration, (5) Evaluation, and (6) Communication. Research can be integrated at every or any one of the first five steps. Research can target understanding and solving any issues to maximize the artifact’s desirability. The landmark publications (Hevner & Chatterjee, 2010), (Hevner & Wickramasinghe, 2018), and (Peffers, et. al., 2007) are useful for more details. 4 Methodology The DSRM inspired design process we fol owed is depicted in Figure 1. The participants of the codesign process are listed in Table 1. The various stages of the design process are described in the subsections that fol ow. 430 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Figure 1: The design process fol owed Table 1: Details of the research team Researcher’s Description about the Role in the project code participant C1 Senior Hematologist Project Lead C2 Senior Anesthetist Clinician Facilitator C3 Senior Anesthetist Clinician Facilitator A1 Professor in Digital Health Principal Investigator A2 Professor in Behavioral Chief Investigator Science A3 Professor in Computer Chief Investigator Science R1 Senior Research Fel ow in Associate Computer Science Investigator R2 Junior researcher in Digital Junior Investigator Health 4.1 Identification of problem and motivation This step was conducted between January and June 2020 with the participation of C1, A1, A2 and A3. This step was carried out through drafting and reviewing a proposal for this project. The primary aim came out as to design and develop a mobile CDSS for the target health context (i.e., optimization of perioperative patients to reduce the incidence of thromboembolism). A secondary aim was to commercialize this CDSS, extending it to become a gold-standard for all surgical procedures. A project proposal document was the outcome of this phase. N. Ulapane, et al.: Mobile Clinical Decision Support Systems: A Pathway from Design to Commercialization 431 4.2 Definition of the objectives for a solution The specific objectives were identified between May and October 2020. These were identified through 1-hour semi-structured virtual meetings that were arranged once a month. Participants C1, A1, A2, A3, R1, and R2 were regular participants and attended al the meetings. Participants C2 and C3 were invited occasional y by the regular participants to obtain specific clinical expertise. Detailed minutes of each meeting were documented. After each meeting the minutes were shared among all participants for consensus. At this stage, the clinician participants C1, C2, and C3, shared with the team the relevant clinical rules to be implemented as the CDSS. The fol owing were defined with consensus as the objectives and deliverables expected from the design phase: (a) A smartphone based CDSS; (b) A database to capture usage data of the CDSS (usage data include data entered to the CDSS and recommendations displayed by the CDSS); (c) A web-based dashboard to enable data display and analytics (this was expected to be a prototype to inspire further developments, specific analytics requirements were not defined at this stage), and (d) Updating certain clinical rules about anticoagulant drug management. 4.3 Design and Demonstration The Design and Demonstration phases occurred in tandem between October 2020 and October 2021. Participants R1 and R2 led the design and implementation. Participant R2 as one of the first activities translated the clinical rules to editable flowcharts. Examples are available in (CLOTS Dashboard Demo, 2022) and (Ulapane, et. al., 2023). This was done to map clinical rules onto a data structure that is accessible to both clinicians and computer scientists. In our experience, this translation was helpful for liaison between the clinicians and the rest. It also helped in programming the clinical rules into an application. We suggest that translating clinical rules into such more widely accessible data structures is an important intermediate step that helps software implementations of clinical rules. These flowcharts were then shared with participants C1, C2, and C3 to update any clinical rules. Those clinician participants reached consensus among themselves and responded with updates to the clinical rules. 432 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY After obtaining the updated clinical rules, the development activities followed. Participant R1 led the software implementation. First, the mobile application was built enabling the app usage data being recorded in a backend database. Secondly, a web-based front-end was developed to enable data display and analytics. The mobile application with data capture facility and the web-based front end combined is the artifact produced by this study. Snapshots of the artifact are available in the Results section. The layouts for the mobile application and the web-interface were deliberately kept simple. Non-cluttered interfaces, preference for push buttons, and using colors of the partnering client hospital’s logo were taken as the key design considerations for the mobile application and the web interface. The rationale behind these design considerations was to enhanced user-friendliness through decluttering and adequate functionality. Fancified aesthetics was not prioritized. Demonstration was done through 1-hour virtual meetings arranged once a month. These meetings were conducted as semi-structured codesign workshops. Incremental progress was demonstrated, and the clinician participants were given the opportunity to provide feedback and express any ‘would-like-to-have’ sort of wishes. As indicated in Figure 1, opportunity was given for participants to reconsider the original objectives and propose any alterations to them. However, no participant proposed major alterations. Again, detailed minutes were recorded and were shared among the participants after each meeting for consensus. The artifact and the source code were the outcomes of this phase. 4.4 The Artifact The Smartphone App with data capture facility and the web-based front end combined is the artifact produced by this study. Details are in the Results section. 4.5 Assessment This phase was carried out between November 2021 and February 2022. The designed application was made available online (e.g., iPhone TestFlight) to download and use. Participants C1, C2 and C3 were asked to download and test. Meanwhile, R2 constructed the questionnaire in (TTF Questionnaire, 2022) based on TTF to be shared with the users of the mobile CDSS to provide feedback. The intention was communicated to C1 to perform a wider assessment, by inviting more clinicians N. Ulapane, et al.: Mobile Clinical Decision Support Systems: A Pathway from Design to Commercialization 433 onboard and asking them to assess the app and provide feedback through the questionnaire in (TTF Questionnaire, 2022). C1 agreed and recruited 7 clinicians inclusive of C1 and C3. These 7 recruits were invited to test the app and then attend a focus group conducted virtual y. In the focused group the participants were presented the questionnaire in (TTF Questionnaire, 2022) and were asked to provide qualitative and quantitative feedback. The qualitative feedback was recorded as meeting minutes. The qualitative feedback was also compiled as a report. In the report the feedback was summarized under the main themes of TTF, i.e., (1) Characteristics of the clinician’s task involving the technology usage; (2) Characteristics of the technology (i.e., the CDSS), and (3) The impact the CDSS has on the clinician’s performance. Several subthemes emerging from the qualitative feedback (inductive analysis) were also highlighted in the report. These findings are summarized in the Results section. This report was then shared with the participants of the focus group for consensus. This report containing user feedback was the outcome of this phase. 4.6 Partnering with a Commercialization Partner This phase was carried out between March 2022 and August 2022. Different software product development companies were considered as candidates to be recruited as a commercialization partner. The track record of previous work and the experience of the candidates, and any preferences of the partnering client cancer hospital were considered as factors that would weigh our choice of a commercialization partner. A candidate was chosen, and several meetings were held to establish relationship and express our interest. Different plans for commercialization that could be offered by the partner were invited alongside quoted costs. The plans were reviewed by the research participants and the plan that matched the current budget constraints was chosen. Affordable amendments to the plans were also proposed. Then, the relevant contracts and nondisclosure agreements pertaining to intel ectual property were signed. Final y, the developed artifact along with the source code (i.e., the outcome of phase 4.3), and the assessment report (i.e., the outcome of phase 4.5) were submitted to the commercialization partner for review. 434 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 4.7 Commercialization Partner’s Assessment This phase was carried out between September 2022 and November 2022. The commercialization partner conducted two activities. One activity was assessing the source code. The source code was assessed against several factors, such as industry best practices for quality of the source code (according to the commercialization partner’s internal y defined criteria that is partly covered by their intel ectual property rights), the cyber security aspects, and the possibility to integrate with existing infrastructure of hospitals. The second activity was a replication of the assessment phase (i.e., phase 4.5) with end users, but again, according to an assessment criterion that is defined by the commercial partner. A report was submitted by the partner to the research team at the end of the assessment. The report contained the following: (1) A summary of the findings from the partner’s assessment, i.e., code quality and user perceptions; (2) The partner’s recommendations along with suggested pathways for commercialization; and (3) Tentative budgets estimates for each commercialization pathway. A couple of fol ow up meetings were held to clarify any unclear points and to reach consensus. The partner’s report following consensus was the outcome of this phase which is consistent with DSRM approaches of getting consensus among al stakeholders/users. 4.8 Recommendations for Commercialization Currently we are considering the recommendations of the commercial partner and are sourcing funding for pursual. The pathways suggested by the partner for commercialization are depicted in the Results section. 5 Results Reference (CLOTS Demo, 2021) provides a video demonstration of the CDSS smartphone application. The smartphone application is available in (CLOTS App, 2022) for download and use. Reference (CLOTS Dashboard Demo, 2022) provides documented description about the app functionality and the web interface. Figures 2 and 3 provide snapshots of the artifact. End user feedback obtained from the focus group using the questionnaire in (TTF Questionnaire, 2022) are summarized in Figure 4. The feedback is summarized under the three main themes of TTF listed in subsection 4.5. The themes emerging from the data are highlighted in bolded font. N. Ulapane, et al.: Mobile Clinical Decision Support Systems: A Pathway from Design to Commercialization 435 More discussion about these results is available in our previous work (Ulapane, et. al., 2023). The pathways suggested by the commercialization partner are depicted as a flowchart in Figure 5. Ultimately validation of superior fit-for-purpose is successful commercialization; so, this is a necessary first step in this regard. Figure 2: Some screenshots of the smartphone based CDSS app (the font in the figure is legible, please zoom to read) Figure 3: Some screenshots of web interface for data display and analytics (the font in the figure is legible, please zoom to read) 436 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 6 Discussion Our study made a twofold contribution: A contribution to theory and a contribution to practice. The contribution to theory was a systematic design approach for mobile CDSSs. The approach was applied to design a mobile CDSS to a leading Australian cancer hospital. Key phases of the design process were detailed. The outcomes of each phase were mentioned, and it was emphasized that a successful plan for commercialization is a necessary step for commercialization which in turn is the ultimate evidence of superior fit-for-purpose. Our approach extended the typical analysis of fit-for-purpose to include validation and suggested pathways for commercialization. Our approach combines the theory of Task Technology Fit and the Design Science Research Methodology (DSRM). The pathways for commercialization came as input by a software product development partner. Figure 4: Findings from the TTF questionnaire-based focus group carried out to validate the CDSS (the font in the figure is legible, please zoom to read) The contribution to practice was a designed and validated mobile CDSS. It included a smartphone application and a web-based data analytics platform. This CDSS was designed to assist clinicians to manage surgery patients during perioperative care. The specific health focus was prevention of thromboembolism. The CDSS can be downloaded from (CLOTS App, 2022) and be used in a smartphone. Key design considerations were discussed in this paper. More details are available in (CLOTS Dashboard Demo, 2022), (CLOTS Demo, 2021), and (Ulapane, et. al., 2023). N. Ulapane, et al.: Mobile Clinical Decision Support Systems: A Pathway from Design to Commercialization 437 Figure 5: Pathways suggested by the commercial partner to commercialize the CDSS (the font in the figure is legible, please zoom to read) Two major limitations or chal enges were encountered during this study. The first was this being a project conducted during the COVID-19 pandemic. Thus, pandemic measures had to be strictly abided by in al the research activities. This also meant recruiting participants and finding available times being a chal enge—the second limitation. This resulted in our participant groups being quite small. Lessons learned and reflections can be summarized as follows. Making new developments interoperability with existing hospital systems can be chal enging. Developing web-based applications instead of mobile applications has some advantages to alleviate the interoperability issue. Policy and regulatory barriers may exist to access health data especial y on cyber security grounds. 7 Conclusions This study reported a systematic design approach spanning from concept to commercialization. The design approach is governed by the Design Science Research Methodology and the theory of Task Technology Fit. The approach was followed to design a smartphone-based clinical decision support system (CDSS) for perioperative patient management. The CDSS was designed for a leading cancer hospital in Australia. The health focus is prevention of thromboembolism in surgery 438 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY patients by optimaly managing surgery patients during perioperative care. Our design approach and experiences gained are generalizable. They can be resourceful for many health tech developments. Acknowledgements This paper is an outcome from the DHCRC-055 project funded by the Digital Health CRC Limited, the Peter MacCal um Cancer Centre, and the Swinburne University of Technology, Australia. Digital Health CRC Limited is funded under the Commonwealth Government’s Cooperative Research Centres Program. References Ahmad, H. S., Yang, A. I., Joshi, D. S., Cleri, N., Basil, G., Wang, M. Y., .. & Yoon, J. W. (2022). 318 Classifying Peri-operative Course and Outcomes in Spine Surgery Using Smartphone-Based Accelerometry. Neurosurgery, 68(Supplement_1), 71-72. Awaludin, S., Nurachmah, E., Soetisna, T. W., & Umar, J. (2022). The effect of a smartphone-based perioperative nursing intervention: Prayer, education, exercise therapy, hypnosis, and music toward pain, anxiety, and early mobilization on cardiac surgery. Journal of Public Health Research, 11(2). Banning, M. (2008). A review of clinical decision making: models and current research. Journal of clinical nursing, 17(2), 187-195. Boaro, A., Leung, J., Reeder, H. T., Siddi, F., Mezzalira, E., Liu, G., . . & Smith, T. R. (2021). Smartphone GPS signatures of patients undergoing spine surgery correlate with mobility and current gold standard outcome measures. Journal of Neurosurgery: Spine, 35(6), 796-806. Chahal, R., Alexander, M., Yee, K., Jun, C. M. K., Dagher, J. G., Ismail, H., . . & Burbury, K. (2020). Impact of a risk‐stratified thromboprophylaxis protocol on the incidence of postoperative venous thromboembolism and bleeding. Anaesthesia, 75(8), 1028-1038. CLOTS App, App store link to download (2022): https://apps.apple.com/us/app/clots/id1436672491 CLOTS Dashboard Demo. (2022). A document demonstrating the CLOTS App and backend: https://dit.swin.edu.au/clotsapp/Overview.docx CLOTS Demo. (2021). A video demonstration of the CLOTS App. Link (Date of last visit: Sep 21, 2022): https://dit.swin.edu.au/clotsapp/CLOTS_demo.mp4 Goodhue, D. L., & Thompson, R. L. (1995). Task-technology fit and individual performance. MIS quarterly, 213-236. Hevner, A., & Chatterjee, S. (2010). Design science research in information systems. In Design research in information systems (pp. 9-22). Springer, Boston, MA. Hevner, A. R., & Wickramasinghe, N. (2018). Design science research opportunities in health care. Theories to Inform Superior Health Informatics Research and Practice, 3-18. Jones, O., Murphy, S. H., & Durrani, A. J. (2021). Regulation and validation of smartphone applications in plastic surgery: It's the Wild West out there. The Surgeon, 19(6), e412-e422. Kabbani, J., & Kabbani, J. (2021). 452 Exploring the Merits and Availability of Smartphone Applications Related to Bariatric Surgery. British Journal of Surgery, 108(Supplement_2), znab134-208. Lee, Y. L., Cui, Y. Y., Tu, M. H., Chen, Y. C., & Chang, P. (2018). Mobile health to maintain continuity of patient-centered care for chronic kidney disease: content analysis of apps. JMIR mHealth and uHealth, 6(4), e10173. N. Ulapane, et al.: Mobile Clinical Decision Support Systems: A Pathway from Design to Commercialization 439 Lesher, A. P., Gavrilova, Y., Ruggiero, K. J., & Evans, H. L. (2021). Surgery and the Smartphone: Can Technology Improve Equitable Access to Surgical Care? Journal of Surgical Research, 263, 1-4. Makary, M. A., & Daniel, M. (2016). Medical error—the third leading cause of death in the US. Bmj, 353. Panda, N., Solsky, I., Cauley, C. E., Lipsitz, S., Desai, E. V., Huang, E. J., . . & Haynes, A. B. (2022). Smartphone-based assessment of preoperative decision conflict and postoperative physical activity among patients undergoing cancer surgery: A prospective cohort study. Annals of Surgery. Peffers, K., Tuunanen, T., Rothenberger, M. A., & Chatterjee, S. (2007). A design science research methodology for information systems research. Journal of management information systems, 24(3), 45-77. Skyttberg, N., Vicente, J., Chen, R., Blomqvist, H., & Koch, S. (2016). How to improve vital sign data quality for use in clinical decision support systems? A qualitative study in nine Swedish emergency departments. BMC medical informatics and decision making, 16(1), 1-12. Soangra, R., & Lockhart, T. (2021). Smartphone-based prediction model for postoperative cardiac surgery outcomes using preoperative gait and posture measures. Sensors, 21(5), 1704. TTF Questionnaire. (2022). Task Technology Fit Theory-Based Questionnaire, For Health Care CDSS Context. Link (Date of last visit: Sep 21, 2022): https://dit.swin.edu.au/clotsapp/20220119_TTF.xlsx Ulapane, N., & Wickramasinghe, N. (2021). Critical issues in mobile solution-based clinical decision support systems: A scoping review. Optimizing health monitoring systems with wireless technology, 32-45. van den Berg, L., Brouwer, P., Panda, N., Hoogbergen, M. M., Solsky, I., Onnela, J. P., . . & Sidey-Gibbons, C. J. (2022). Feasibility and performance of smartphone-based daily micro-surveys among patients recovering from cancer surgery. Quality of Life Research, 31(2), 579-587. Voglis, S., Ziga, M., Zeitlberger, A. M., Sosnova, M., Bozinov, O., Regli, L., . . & Maldaner, N. (2022). Smartphone-based real-life activity data for physical performance outcome in comparison to conventional subjective and objective outcome measures after degenerative lumbar spine surgery. Brain and Spine, 2, 100881. Wu, M. H., Wu, P. C., Hseih, L. B., Lin, L. M., & Tsai, C. L. (2022). A Smartphone Application for Monitoring Postoperative Symptoms and Signs with an Instant Bi-Direction Feedback System in Patients after Cataract Surgery. Ulapane, N., Forkan, A. R. M., Jayaraman, P. P., Schofield, P., Burbury, K., & Wickramasinghe, N. (2023). Using Task Technology Fit Theory to Guide the Codesign of Mobile Clinical Decision Support Systems. Proceedings of the 56th Hawai International Conference on System Sciences. pp. 2870-2879. 440 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY FICTION OR REALITY – WHICH GAME STORY PROMOTES LEARNING OUTCOME MORE? LINDA GROGORICK, MORITZ ROHDE , BIJAN KHOSRAWI-RAD, SUSANNE ROBRA-BISSANTZ Technische Universität Braunschweig, Institute for Business Information Systems, Braunschweig, Germany l.grogorick@tu-bs.de, moritz.rohde@tu-bs.de, b.khosrawi-rad@tu-bs.de, s.robra- bissantz@tu-bs.de Digital Game-based Learning (DGBL) has achieved several positive results in recent years, e.g., increased fun, motivation, or learning outcome. However, many DGBL applications fail, which makes an isolated consideration of individual game elements and their influence on learning necessary to better Keywords: design future DGBL applications. One widely used game game element is the game story. As there is little research on how a story, serious game story should be designed to promote learning, this paper game, conducts an experiment comparing a fictional game story with a presentation realistic one. The results show that both game stories lead to a skil s, motivation, significant objective knowledge gain. In terms of learning learning outcome, both game stories achieved similar results. outcome DOI https://doi.org/10.18690/um.fov.4.2023.28 ISBN 978-961-286-751-5 442 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction With the introduction of the term digital game-based learning (DGBL) by Prensky (2001), the topic of combining games and learning has received a lot of attention in both research and practice (Breuer, 2010). Studies show that the use of game elements for educational purposes leads to positive effects such as increased motivation, engagement, creativity, or fun (Oliveira et al, 2023; Grogorick, 2021; Prensky, 2003). There are two forms of DGBL design: Gamification and serious games. Gamification is the integration of individual game elements into a non-game context (e.g., education) (Deterding et al., 2011). An example is the teaching concept 'GamEducation,' i.e., individual game elements are integrated into an existing lecture to create a 'gameful experience' in learning strategic decision-making and planning of digital businesses (Siemon and Eckardt, 2017). In contrast to gamification, a serious game does not simply integrate individual game elements into an existing environment, such as an existing lecture. Instead, a serious game is a complete game with fixed rules and objectives (Deterding et al., 2011). One example is 'Adapt or Die' to learn business model development (Schlimbach and Robra-Bissantz, 2022). The two forms of integrating game elements into education, gamification and serious game, are not always clearly distinguishable, so the terms are often used synonymously. However, both forms create a situation where learners expect fun and enjoyment and are thus motivated, promoting learning outcomes as the overriding goal (Grogorick, 2021; Manzano-León et al. 2021). Nevertheless, many DGBL applications fail because learning is not fun. The reason for this is that DGBL designers focus on the achievement of learning rather than creating an environment that, in addition to the achievement of learning aims, also provides an intense gaming experience characterized by motivation and fun (Zichermann and Cunningham, 2011; Eckardt et al., 2018; Van Roy and Zaman, 2017). In order to support learning as effectively as possible, the development of DGBL focuses not only on the design of the application in general but also on the concrete design of individual game elements (Grogorick, 2021). The game story is a commonly used game element (Hamari et al. 2014). While not al DGBL applications use an accompanying game story, storytelling is still relevant to the learning experience and provides a context for processing learning content (Clark et al., 2016; Kapp, 2012). Studies conducted so far show contradictory results as to whether or not the integration of a game story positively impacts learning and engagement. L. Grogorick, M. Rohde , B. Khosrawi-Rad, S. Robra-Bissantz: Fiction or Reality – Which Game Story Promotes Learning Outcome More? 443 Some studies argue that learners use their cognitive capacities to process the game story instead of the learning content (Wouters et al., 2013; Adams et al., 2012). Other work has found that manipulating the game story influences learning. For example, a complex game story has a negative effect compared to a minimal y designed game story (McQuiggan et al., 2008), or triggering surprises in a game story leads to a greater depth of knowledge processing (Spek, 2011; Hoeken and van Vliet, 2000). As the effects of manipulated game stories are stil unexplored (Wouters et al., 2013), we want to address this topic in this paper. Game stories are often fictional or based on reality (Nicholson, 2015). In commercial computer games, research has shown a realistic storyline has an inspiring effect, making the game more attractive to players (Olson, 2010). But other studies have also demonstrated the positive effect of fictional game stories, as one can become completely immersed in them and forget about 'real world' problems (Fullerton, 2014). In an educational context, for example, Kruse et al. (2014) chose a story and characters for learning mathematics due to the popularity of vampire series and films. However, many students did not identify with this and felt that they were not taken seriously, which led to a high drop-off in the use of the application. In addition, Eckardt and Robra-Bissantz (2018) evaluated a serious game for learning information literacy skil s, in which students on a technical course take part in a fictional research expedition that more closely resembles a real-life story. Although this application achieved positive results regarding the learning outcome (Eckardt and Robra-Bissantz, 2018), these are not clearly attributable to the choice of the game story. Another study used an analog board game to investigate whether a fictional or a realistic storyline was more motivating to learn and found no significant differences regarding the motivational effect (Eckardt et al., 2019). As previous results are mixed and a direct comparison between a fictional and a realistic story in learning has, to the best of our knowledge, only been done with a non-digital board game, this paper will report the results of a comparative study using a DGBL application, testing a realistic versus a fictional story. Therefore, the research question (RQ) of this paper is as follows: RQ: How does the learning outcome differ between a fictional and a realistic story in digital game-based learning? 444 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY By answering our RQ, we aim to provide new knowledge about the learning outcome of fictional and realistic game stories in DGBL and inspire game story design in future DGBL applications. Overall, we would like to contribute to the improvement of game story design and thus supporting to solve the problem of DGBL's failure. Our paper is structured as fol ows. First, we start with the research background, addressing the story as a game element and the learning outcome concept. Second, we explain the design of our serious game with a fictional and realistic story. Third, we present the study we conducted, starting with an explanation of our experimental design, before going on to explain and then discuss our results. Finally, in our conclusion, we summarize our findings and show possibilities for further research. 2 Research Background 2.1 Game Story Storytelling is highly relevant to the learning experience as it provides a narrative context for processing and applying learning content (Kapp, 2012). In general, stories are part of the real world. However, they create an artificial y perceived fictional world in the minds of those who immerse themselves in them. For example, a game creates an imaginary world in the player's mind. Players gain new knowledge while immersed in the story by using their previous experience to transfer it to the story's meaning, thereby gaining new knowledge and skil s that are relevant to real-life situations (Lugmayr et al., 2017). Integrating storytelling also enriches classroom activities. For example, storytel ing can be one of the most relevant learning experiences, as shown by its implementation in HIV and AIDS education (Duveskog and Sutinen, 2013). The learning experience is shown through the consumption of the story. The story moves forward whenever learners have made learning progress (Lugmayr et al., 2017). There are two ways to create a story in a learning application: fictional or realistic. Many films we watch in the cinema or on Netflix, and even games such as Super Mario, where Princess Peach has to be rescued again and again, have a fictional plot with people or characters that do not exist in real life. These are cal ed fictional stories. The opposite of this is a realistic story. The realistic story actually happened, and the protagonists real y existed (or still exist) (Andrews, 2010). L. Grogorick, M. Rohde , B. Khosrawi-Rad, S. Robra-Bissantz: Fiction or Reality – Which Game Story Promotes Learning Outcome More? 445 Overall, storytelling serves to structure individual actions and characters so that they have meaning (Kapp, 2012). Accordingly, the player's experience should become meaningful, leading to a narrative presence with which the player can identify (Laschke and Hassenzahl, 2011). This narrative presence is seen as an element of immersion that can lead to a positive wil ingness to use an application (Ryan et al., 2006). For DGBL, this means that the integration of stories can contribute to learners wanting to learn with the developed application because the DGBL application creates a sense of immersion. When learners want to learn with an application, they are also motivated, which can positively impact the learning outcome (Grogorick, 2021). 2.2 Learning Outcome A successful game story supports the creation of a meaningful learning outcome. Therefore, we now consider the learning outcome and how it can be operationalized. Defining and measuring the variable learning outcome is difficult because many factors influence it (Kerres, 2001; Häussler, 2007). In general, learning outcome can be described as the result of al didactic activities and therefore means more than memorizing facts, events, or processes, to which it is often reduced, as shown by conducting numerous written knowledge tests (Kerres 2001; Adam, 2004). Kerres (2001) defines several factors that influence the learning outcome. Those relevant to the measurement carried out in this paper are listed below and then explained in more detail: • Objective learning success or objective knowledge, measured at different time intervals. • Subjective satisfaction with learning behavior and outcome, i.e., subjectively perceived knowledge gain. • Emotional responses in terms of motivation to learn and identification with the learning application or game story. Objective and Subjective Knowledge. In general, Probst et al. (2006) define the concept of knowledge as the totality of knowledge and skills that individuals use to solve problems. According to them, it is only through the combination of information and its application that knowledge is created, which in turn can become a skill (Mescheder and Sallach, 2012). In defining knowledge researchers distinguish 446 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY between objective and subjective knowledge. A person's actual stored knowledge, also called factual knowledge, describes objective knowledge (Brucks, 1985). Factual knowledge is often col ected by answering questions to verify what has been learned (Johann, 2008). Subjective knowledge is the assessment of a person's knowledge in a particular area. This knowledge assessment can be made by oneself or another person (Brucks, 1985). Motivation. Emotional responses, such as learner motivation, can be measured using different models. One model for measuring learners’ motivation is the ARCS model by Keller (1987). ARCS is an acronym for four factors that promote motivation, which we explain in the following: attention, relevance, confidence in success, and satisfaction. Arousing and maintaining the learner’s interest is the attention dimension. As part of the learning process, learners need to be shown how the skil s they are learning wil be helpful to their everyday or professional lives. The dimension of relevance summarizes this aspect. Raising learners’ expectations of success is important and can be achieved, for example, by making learning objectives clear. As a result, learners develop a sense of control over their own learning outcome, to which the motivational dimension of confidence in success refers. Satisfaction and motivation are closely related. If learners are satisfied with the learning application, they will be motivated to learn. Identification. Identification with the learning application or game story is another emotional response. It is characterized as a mental process in which the user takes on a character's perspective and thus imagines that character's experiences, emotions, and perceptions (Konijn and Hoorn, 2005). In the design of a DGBL application, the narrative context determined by the story of the game is one element that significantly influences the end user’s identification with the game (Reeves and Read, 2009). 3 Serious Game for Learning Presentation Skills The DGBL application of this paper focuses on learning presentation skil s. Effective presentation skil s are essential for al students as they are associated with individual career opportunities and benefits (Chaney and Green, 2002). In both versions (fictional and realistic), the serious game has the same content for better comparability (e.g., aims and structure of convincing presentations). The DGBL L. Grogorick, M. Rohde , B. Khosrawi-Rad, S. Robra-Bissantz: Fiction or Reality – Which Game Story Promotes Learning Outcome More? 447 application is a browser-based board game. Browser games have the advantage that they are easily accessible to students; for example, users do no not need to instal additional software (Costu et al., 2009). The digital board game is more like a serious game, as it has fixed rules and objectives (Deterding et al., 2011). First of al , players can choose an avatar. In the fictional story of the game, the characters are fantasy creatures cal ed 'Badanklas.' As a Badankla, the player wants to improve the vil age and therefore intends to present an idea to the Great Council of Badanklas. Knowing that he/she does not have the best presentation skil s, the Badankla wants to improve them before the presentation to convince the Great Council of his/her idea. The realistic game story is about a student that aims to finish his bachelor thesis and wants to work on his presentation skil s before the oral defense. A game story with a theme closely linked to learning aims (as we do with presentation skil s) can enhance the impact of a narrative (Wouters, 2013). Figure 1 shows the board game. Figure 1: Screenshot of the Board Game Up to four learners can play the digital board game at the same time. They are initial y on the 'Start' field. Each player rolls a die and, depending on the number they roll, moves one step closer to the goal. The goal represents the final of the storytel ing by applying the learned presentation skills in giving a presentation. On the way to the goal, players wil come across event and knowledge cards. Players can draw event 448 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY cards when they land on squares with explosions. The event cards are designed to fit and relate to the story of the game. For example, an event card might say in the fictional story, 'Your Badankla friends tel you that your presentation is getting better and better. Take two steps forward.' In the realistic story, it might say: 'You take a break and go for a run. Afterward, you feel much fitter mental y. Take one step forward.' In contrast, participants can draw knowledge cards when they land on squares with lamps. Each knowledge card contains new learning content that will be presented to al players. Figure 2 summarizes the game elements used in the DGBL application. With the exception of storytel ing, al game elements are designed identical y in both versions of the game so that the influence on learning outcome depends as much as possible on the choice of the story (fictional or realistic). Figure 2: Game Elements of the Board Game L. Grogorick, M. Rohde , B. Khosrawi-Rad, S. Robra-Bissantz: Fiction or Reality – Which Game Story Promotes Learning Outcome More? 449 4 Fictional or Realistic Game Story: Study on Learning Outcome 4.1 Study Design We conducted a controlled experiment under laboratory conditions lasting about approx. 30 minutes to evaluate how learners assess the two variants. The participants were randomly assigned into two groups (fictional vs. realistic game story) at the beginning of the study. The experiment begins with the completion of an online questionnaire to assess learners’ prior knowledge. After a short explanation, the serious game follows, either with the fictional story (group 1) or the realistic story (group 2). After playing the serious game, the participants have to complete a post-test with questions on knowledge and emotional reactions (identification and motivation) to assess the learning outcome. Figure 3 shows the research process in summary. Figure 3: Study Design We assessed the learning outcome using the factors defined by Kerres (2001) (see Chapter 2.2). We evaluated subjective knowledge according to Flynn and Goldsmith (1999). Our study includes knowledge questions and free-text questions to measure objective knowledge, as these al ow the evaluation of a high processing depth of knowledge (Kibler and Eckardt, 2018; Biggs, 1999). We asked eight knowledge questions, e.g., about the aims of presenting or about recognizing different steps in the process of presenting. We measured motivation based on the ARCS model (Keller, 1987) using a scale by Chen and Chan (2008) adapted to the DGBL context. The scale divides motivation into attention, relevance, challenge, and satisfaction. The item "I am looking forward to the next session to use this game” (satisfaction) is not part of the questionnaire because participants only use the application during the experiment. Moreover, we used the identification scale according to Homburg 450 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY et al. (2009). We measured subjective knowledge, motivation, and identification using a 5-point Likert scale (1 = strongly disagree, .. , 5 = strongly agree). Regarding the evaluation of objective knowledge, a value between zero (al questions answered incorrectly) and one (al questions answered correctly) corresponds to the result. 4.2 Study Results 52 learners participated in the experiment (24 male and 28 female). Of these, 27 participants chose the realistic game story, and 25 chose the fictional one. The average age was 24 years. In the following, we report the experiment results by giving each construct's mean values (MV) and standard deviations (SD). In addition, we conducted a Student's t-test for independent samples to test whether significant differences exist between the assessment of the fictional and realistic game story for the respective constructs (α = 0.05). Table 1 shows our experiment results. Table 1: Experiment Results Fictional Story vs. Realistic Story Construct Fictional Realistic t-Test Story Story MV SD MV SD T P Objective Knowledge (Pre- 0.34 0.43 0.37 0.44 0.69 0.49 Test) Subjective Knowledge (Pre- 2.54 1.05 2.46 1.00 0.54 0.59 Test) Objective Knowledge (Post- 0.56 0.47 0.54 0.46 0.38 0.71 Test) Subjective Knowledge (Post- 2.49 1.00 2.32 0.95 1.23 0.22 Test) Attention (Motivation) 3.12 0.96 3.01 0.85 0.9 0.19 Relevance (Motivation) 3.03 0.95 2.95 0.91 0.59 0.28 Chal enge (Motivation) 3.18 0.99 3.02 0.88 1.25 0.11 Satisfaction (Motivation) 3.01 1.2 2.98 0.89 0.23 0.41 Identification 3.25 1.15 3.13 0.91 0.82 0.21 First, we tested in a pre-test whether there were differences in prior knowledge (subjective and objective) between the two experimental groups (fictional vs. realistic story). We found no significant knowledge differences in the pre-test. Regarding the L. Grogorick, M. Rohde , B. Khosrawi-Rad, S. Robra-Bissantz: Fiction or Reality – Which Game Story Promotes Learning Outcome More? 451 post-test, the two groups had only minimal differences in objective knowledge. However, the mean value for subjective knowledge gain was higher for the fictional story group (2.49 > 2.32), although not significantly (p > 0.05). Regarding the motivation effects according to the ARCS model as well as the identification, the fictional game story was also consistently rated slightly better than the realistic game story, although we found no significant effects. In addition, the mean value for the fictional story for motivation and identification was consistently above the scale mean (mean value > 3.00), so learning with the game story tended to be perceived positively. Overall, the participants perceived the fictional story slightly better after completing the experiment. In addition, we conducted further t-tests to check whether the participants perceived a subjective or objective knowledge gain after playing the serious game (α= 0.05). Table 2 presents the results of the knowledge gain analysis. Table 2: Results Knowledge Gain Construct Pre-Test Post-Test t-Test MV SD MV SD T P Objective Knowledge 0.34 0.43 0.56 0.47 4.89 < 0.01 (Fictional Story) Objective Knowledge 0.37 0.44 0.54 0.46 3.95 < 0.01 (Realistic Story) Subjective Knowledge 2.54 1.05 2.49 1.00 0.35 0.37 (Fictional Story) Subjective Knowledge 2.49 1.00 2.32 0.95 1.05 0.15 (Realistic Story) The results show that the participants subjectively perceive no gain in knowledge. For instance, the mean values of the post-test after learning with the serious game are even slightly worse for both forms of the game story than for the pre-test (not significant). However, we found a significant objective knowledge gain for both game stories (p < 0.01). Therefore, participants assumed they had not gained any knowledge, but their responses provided evidence that they had objectively gained new knowledge. 452 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 4.3 Discussion of Results Objective knowledge improved significantly after the learning process in both the realistic and fictional game story. This was not the case for subjective knowledge. The reason for this may be an inaccurate self-assessment, which can be explained by a person's level of competence. According to this, incompetence often leads to overestimating one's own abilities, whereas increasing competence (in this case, increasing objective knowledge) leads to an underestimation of one's own level of knowledge (Kim et al., 2016). This phenomenon has also been shown in other studies of knowledge assessment in an educational context (Tashiro et al. 2021; Eckardt et al., 2019). The comparison of the learning outcome, measured by knowledge (objective and subjective), motivation, and identification, shows no significant differences in both versions (fictional and realistic story). Both versions were perceived identical y, except for slightly different mean scores. A study that also compared a fictional with a realistic game story shows similar results, but these refer to an analog serious game (Eckardt et al., 2019). In our study, we confirmed these results in the digital context. Furthermore, other studies that have analyzed whether the inclusion or exclusion of game stories impacts the learning outcome have also found no significant differences (Sailer and Homner 2020; Wouters et al. 2013). A possible reason could be that the storytelling element has less influence on the learning outcome in the context of DGBL than other game elements (e.g., feedback or collaboration). Therefore, the design of the element (e.g., fictional or realistic) or even its presence is not crucial. This finding is also supported by other studies concentrating on an isolated analysis of the impact of individual game elements. For example, the presence of collaboration has the same positive significant impact on the learning outcome as the combined integration of collaboration and competition (Eckardt and Finster, 2019). However, further studies are needed to confirm this, exploring both different designs of individual game elements and the integration or exclusion of individual game elements and their influence on the learning outcome. The need to consider the impact of individual game elements in isolation has also been highlighted in other studies (e.g., Khosrawi-Rad et al., 2021; Landers et al., 2017). L. Grogorick, M. Rohde , B. Khosrawi-Rad, S. Robra-Bissantz: Fiction or Reality – Which Game Story Promotes Learning Outcome More? 453 5 Conclusion and Future Research In this paper, we presented a DGBL application in two versions (with a fictional and a realistic game story) for learning presentation skills. We evaluated it in terms of their learning outcome. In this analysis, we focused on the impact of static game stories on the learning outcome. Future studies could therefore explore the impact of dynamical y designed game stories on the learning outcome or compare dynamic and static game stories. Dynamic stories al ow for more autonomy and choice (Fullerton, 2014), which positively impacts learning according to self-determination theory (Van Roy and Zaman, 2017). We contribute to answering our RQ by testing our serious game in an experiment with 52 students, whereby both versions were primarily rated similarly. Our findings provide new insights for both practice and research. On the one hand, we show that the choice of the game story (fictional or realistic) has no impact on the learning outcome, thus contributing to the isolated consideration of the impact of individual game elements and their manipulation. On the other hand, our results aim to inspire practitioners to design their own DGBL applications, as they can freely choose between realistic and fictional storytel ing to achieve a learning outcome. References Adams, D. M., Mayer, R. E., MacNamara, A., Koenig, A., & Wainess, R. (2012). Narrative games for learning: Testing the discovery and narrative hypotheses. Journal of Educational Psychology, 104, 235–249. Adams, S. (2004). Using Learning Outcomes. Report for United Kingdom Bologna Seminar. Heriot-Watt University. Andrews, D. H. (2010). Story types and the hero story. In D.H. Andrews, T.D. Hul & J.A. Donahue (Eds.), Storytelling as an instructional method (pp. 1-10). Sense Publishers. Biggs, J. (1999). What the student does: teaching for enhanced learning. Higher Education Research and Development, 18(1), 57–75. Breuer, J. & Bente, G. (2010). Why so Serious? On The Relation of Serious Games and Learning. Journal for Computer Game Culture, 4, 7-24. Brucks, M. (1985). The Effects of Product Class Knowledge on Information Search Behaviour. Journal of Consumer Research, 12(1), 1-16. Chaney, L. and Green, C. (2002). Presenter behaviors: Actions often speak louder than words. Supervision, 63(5), 17-19. Chen, Z. H., & Chan, T. W. (2008). Learning by substitutive competition: Nurturing my-pet for game competition based on open learner model. Proceedings of the International Conference on Digital Games and Intel igent Toys Based Education, 124-131. Clark, D. B., Tanner-Smith, E. E., & Killingsworth, S. S. (2016). Digital games, design, and learning: A systematic review and meta-analysis. Review of educational research, 86(1), 79-122. 454 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Costu, S., Aydn, S. & Filiz, M. (2009). Students' Conceptions About Browser-Game-based Learning in Mathematics Education: TTNetvitamin Case. Procedia-Social and Behavioral Sciences, 1(1), 1848-1852. Deterding, S., Khaled, R., Nacke, L. E., & Dixon, D. (2011). Gamification: Toward a Definition. CHI 2011 Gamification Workshop Proceedings. Duveskog, M. & Sutinen, E. (2013). Enrichtung student HIV awareness by digital storytelling. Journal of Educational Multimedia and Hypermedia, 22(4), 383-406. Eckardt, L., & Finster, R. (2019). Kollaboration oder Wettbewerb: ein Vergleich der Motivation beim Game-based Learning. HMD Praxis der Wirtschaftsinformatik, 56(1), 83-93. Eckardt, L., Grogorick, S. & Robra-Bissantz, S. (2018). Play to Learn: Conducting a Playtest Session for Improving an Educational Game. Proceedings der American Conference on Information Systems. Eckardt, L., Röske, D. & Robra-Bissantz, S. (2019). Influence of Belief in a Just World on Knowledge in Game-based Learning. Proceedings of the Bled eConference Humanizing Technology for a Sustainable Society, 507-522. Eckardt, L., Schlaf, S. P., Barutcu, M., Ebsen, D., Meyer, J., & Robra-Bissantz, S. (2019). Empirische Untersuchung des Einflusses der Identifikation mit einer Spielgeschichte auf den Lernerfolg bei einem Serious Game, Proceedings of Teaching Trends, 139-145. Eckardt, L. & Robra-Bissantz, S. (2018). Learning Success: A Comparative Analysis of a Digital Game-Based Approach and a Face-to-Face Approach. Proceedings of the 31th Bled eConference Digital Transformation - Meeting the Chal enges, 331-343. Flynn, L.R. & Goldsmith, R.E. (1999). A Short, Reliable Measure of Subjective Knowledge. Journal of Business Research. 46(1), 57-66. Fullerton, T. (2014). Game Design Workshop: a Playcentric Approach to Creating Innovative Games. CRC Press. Grogorick, L. (2021). Design und Implementierung eines Serious Games zur Steigerung des Lernerfolgs am Beispiel der Informationskompetenz. Technische Universität Braunschweig. Häussler, P. (2007). Wie lässt sich der Lernerfolg messen? In E. Kircher, R. Griwdz, & P. Häußler (Eds.) Physikdidaktik (pp. 249-294). Springer. Hamari, J., Koivisto, J., & Sarsa, H. (2014). Does gamification work?--a literature review of empirical studies on gamification. Proceedings of the 47th Hawai international conference on system sciences, 3025-3034. Hoeken, H., & van Vliet, M. (2000). Suspense, curiosity, and surprise: How discourse structure influences the affective and cognitive processing of a story. Poetics, 27, 277–286. Homburg, C., Wieseke, J., & Hoyer, W. D. (2009). Social identity and the service–profit chain. Journal of Marketing, 73(2), 38-54. Johann, D. (2008). Probleme der befragungsbasierten Messung von Faktenwissen. Sozialwissenschaften und Berufspraxis, 31 (1), 53 65. Kapp, K. M. (2012). The Gamification of Learning and Instruction: Game-based Methods and Strategies for Training and Education. John Wiley & Sons. Kibler, S. & Eckardt, L. (2018). On the Role of Tasks in Virtual Game-based Learning: The Example of Lost in Antarctica. IATUL 2018 Proceedings, 2-13. Keller, J. M. (1987). Strategies for stimulating the motivation to learn. Performance+ Instruction, 26(8), 1-7. Kerres, M. (2001). Multimediale und telemediale Lernumgebungen: Konzeption und Entwicklung. De Gruyter Oldenbourg. Khosrawi-Rad, B., Grogorick, L., Robra-Bissantz, S. (2021) Zwischen Frust und Freude – Wie nehmen Lernende verschiedene Spielmechaniken beim Digital Game-based Learning wahr?. Proceedings der Gemeinschaft neuer Medien, 75-91. Kim, Y.-H., Kwon, H., Lee, J. & Chiu, C.-Y. (2016). Why Do People Overestimate or Underestimate Their Abilities? A Cross-Culturally Valid Model of Cognitive and Motivational Processes in Self-Assessment Biases. Journal of Cross-Cultural Psychology, 47(9), 1201-1216. L. Grogorick, M. Rohde , B. Khosrawi-Rad, S. Robra-Bissantz: Fiction or Reality – Which Game Story Promotes Learning Outcome More? 455 Kruse, V., Plicht, C., Spannagel, J., Wehrle, M., & Spannagel, C. (2014). Creatures of the Night: Konzeption und Evaluation einer Gamification-Plattform im Rahmen einer Ma-thematikvorlesung. DeLFI Workshop Proceedings, 246-253. Konijn, E., & Hoorn, J. (2005). Some like it bad: Testing a model for perceiving and experiencing fictional characters. Media Psychology, 7, 107–144 Landers, R. N., Armstrong, M. B., & Col mus, A. B. (2017). How to use game elements to enhance learning: Applications of the theory of gamified learning. Serious Games and Edutainment Applications, 2, 457-483. Laschke, M. & Hassenzahl, M. (2011). Mayor or Patron? The Difference Between a Badge and a Meaningful Story. Proceedings of the CHI 2011 Workshop on Gamification, 72-75. Lugmayr, A., Sutinen, E., Suhonen, J., Sedano, C. I., Hlavacs, H. & Montero, C. S. (2017). Serious Storytelling – a first definition and review. Multimedia Tools and Applications, 76, 15707-15733. Manzano-León, A., Camacho-Lazarraga, P., Guerrero, M. A., Guerrero-Puerta, L., Aguilar-Parra, J. M., Trigueros, R., & Alias, A. (2021). Between level up and game over: A systematic literature review of gamification in education. Sustainability, 13(4), 22-47. Mescheder, B. & Sal ach, C. (2012). Wettbewerbsvorteile durch Wissen. Springer. Nicholson, S. (2015). A recipe for meaningful gamification. In Reiners, T., Wood, L. (Eds.) Gamification in Education and Business. Springer McQuiggan, S., Rowe, J., Lee, S., & Lester, J. (2008). Story-based learning: The impact of narrative on learning experiences and outcomes. Proceedings oft he 9th International Conference on Intelligent Tutoring Systems, 530-539. Prensky, M. (2001). Digital Game-based Learning. Paragon House. Probst, G., Raub, S. & Romhardt, K. (2006). Wissen managen. Wie Unternehmen ihre wertvol ste Ressource optimal nutzen. Gabler. Reeves, B. & Read, J.L. (2009). Total engagement: Using games and virtual words to change the way people work and business complete. McGraw-Hill Professional. Ryan, R. M., Rigby, C. S. & Przybylski, A. (2006). The Motivational Pull of Video Games: A Self-Determination Theory Approach. Motivation and emotion, 30(4), 344-360. Sailer, M., & Homner, L. (2020). The gamification of learning: A meta-analysis. Educational Psychology Review, 32(1), 77-112. Schlimbach, R., and Robra-Bissantz, S. 2022. Adapt or Die! - A Board Game to Support Dynamic Business Model Creation in Digital Darwinism, Proceedings of the 15th International Conference on Information Systems. Siemon, D., & Eckardt, L. (2017). Gamification of Teaching in Higher Education. In S. Stieglitz, Lattemann, C. & Robra-Bissantz, S. (Eds.), Gamification (pp. 153–164). Springer. Tashiro, J., Parga, D., Pol ard, J., & Talanquer, V. (2021). Characterizing change in students' self-assessments of understanding when engaged in instructional activities. Chemistry Education Research and Practice, 22(3), 662-682. Oliveira, W., Hamari, J., Shi, L., Toda, A. M., Rodrigues, L., Palomino, P. T., & Isotani, S. (2023). Tailored Gamification in Education: A literature Review and Future Agenda. Education and Information Technologies, 28(1), 373–406. Olson, C. K. (2010). Childrenś Motivation for Video Game Play in the Context of normal Development. Review of General Psychology, 14, 180-187. van der Spek, E. D. (2011). Experiments in serious game design: A cognitive approach. Universiteit Utrecht. Van Roy, R., & Zaman, B. (2017). Why gamification fails in education and how to make it successful: Introducing nine gamification heuristics based on self-determination theory. Serious Games and Edutainment Applications, 2, 485-509. Wouters, P., van Nimwegen, C., van Oostendorp, H., & van der Spek, E. D. (2013). A meta-analysis of the cognitive and motivational effects of serious games. Journal of Educational Psychology, 105(2), 249–265. 456 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Zichermann, G. & Cunningham, C. (2011). Gamification by Design: Implementing Game Mechanics in Web and Mobile Apps. O'Reil y Media. IMPROVING COMPLETION RATE OF DIGITAL SKILL SELF-ASSESSMENT SURVEY: AN EMPIRICAL STUDY XIAO PENG, RAN ZHANG, TAMARA MARANTIKA, MACHIL DEINUM, JET VAN DER TOUW, XANDER LUB, KOEN VAN TURNHOUT University of Applied Sciences Utrecht, Utrecht, Netherland xiao.peng@hu.nl, ran.zhang@hu.nl, tamara.marantika@hu.nl, machil.deinum@hu.nl, jet.vandertouw@hu.nl, xander.lub@hu.nl, koen.vanturnhout@hu.nl This study examines completion rate for a self-assessment survey designed to assess employees' digital skills levels in the workplace. The aim is to improve data quality by investigating completion of the survey. The study reviews the theoretical background related to self-assessment surveys and completion rate, and explores the influence of survey length and format in survey design on completion rate. The research design and data analysis are described in detail, with a focus on identifying factors that may influence completion rate. Results suggest that survey Keywords: designers should consider using Likert scales to optimize self-assessment completion rate and completion time. However, this study did surveys, not find a significant increase in completion rate as a result of likert scale survey, motivation, which was claimed from the literature. The study slider concludes with implications for the design and implementation survey, of self-assessment surveys in the workplace, including the two-option survey, importance of reducing length and complexity of survey items completion and questions. rate DOI https://doi.org/10.18690/um.fov.4.2023.29 ISBN 978-961-286-751-5 458 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The use of self-assessment surveys has become increasingly popular in recent years as a means of evaluating an individual's skills, knowledge, and abilities (Andrade, 2019; Benraghda et al., 2022; Noorka & Sarwar, 2020). These surveys have proven to be particularly useful in the workplace, where they can help identify areas where employees may require additional training or support. Although self-assessment surveys have the potential to offer numerous benefits, gathering accurate data can be chal enging due to low response rates, low completion rate and early drop outs, particularly when the self-assessment survey results reflect one's own skil s level that he/she sees private. This study aims to investigate the factors that influence participants’ fatigue and completion rate for a self-assessment survey designed to test employees' digital skil s levels. The goal is to identify factors that can be used to increase participation and completion in the survey, thereby improving the quality of the data collected. Specifical y, this research wil explore the use of various perspectives in the survey design as a potential means of increasing completion rate. The digital skills survey has been chosen for testing in this study due to two principal reasons. Firstly, this assessment tool was conceptualized and developed by the authors themselves, who possess an excel ent grasp of its underlying theoretical constructs. Consequently, the authors are better equipped to effectively scrutinize and decipher the results yielded by this current study. Secondly, during the preliminary testing phase, the authors noted inconsistent completion rates, which served as a catalyst for devising the current study, which aims to delve deeper into the determinants of completion rates in self-assessment surveys employing diverse types of response scales. The study wil begin with a review of the theoretical background related to self-assessment surveys and completion rate. This will be followed by a detailed description of the research design, data analysis and results with a focus on identifying factors that may be influencing completion rate. Finally, conclusions will be drawn along with a discussion of their implications for the design and implementation of self-assessment surveys in the workplace. X. Peng, R. Zhang, T. Marantika, M. Deinum, J. van der Touw, X. Lub, K. van Turnhout: Improving Completion Rate of Digital Skil Self-assessment Survey: an Empirical Study 459 2 Research background Self-assessment survey for monitoring skil s costs less resources to administer than other methods and requires less time than a testing (Allen & van der Velden, 2005), and web-based survey as a survey instrument is developed as an easy and economical method of collecting data (Parsons, 2007). An advantage of web-based survey is that it’s possible to track the respondent behavior, namely “complete responders (respond to the entire survey), unit non-responders or refusals (do not respond to the survey at al ), and item non-responders (respond to only some of the questions)” (Parsons, 2007). This study primarily focuses on the impact of length and formats/types of a survey on the respondent behavior by examining the completion rate of the survey, as these factors are often discussed as significant determinants (Beebe et al., 2010a; Buskirk et al., 2015; Dobronte, 2015; Galesic & Bosnjak, 2009; Hoerger, 2010; Nestler et al., 2015; Revilla & Ochoa, 2017; Roster et al., 2015). Even though no significant relationship was found between survey length and quality (Beebe et al., 2010b), longer the survey, more fatigue and burden of the participants, especial y web-paged surveys (Galesic, 2006; Savage & Waldman, 2008). Both Deutskens et al. (2004) and Marcus et al. (2007) found that participants were willing to complete the survey if it lasted shorter (15-30 minutes and 10-20 minutes), compared to those longer than 30 minutes. Galesic and Bosnjak (2009) tested survey length with 10/20/30 minutes, and found that more respondents started and completed the survey with 10 minutes than the rest. Brace (2013) suggested the length of an online survey as 15 mins and around 40-45 questions. Moreover, Revilla & Ochoa (2017) tested and concluded that ideal survey length is a median of 10 minutes and that the maximum survey length is 20 minutes. Roster et al. (2015) compared traditional radio-button Likert scale and slider bar formats for five-point Likert scale questions and argued that “a more interactive experience may reduce survey fatigue and nonresponse, and potential y, lead to higher quality data”. Stanley & Jenkins (2007) claimed that the “fun factor” of slider scales can engage respondents, especial y for young respondents. Buskirk et al. (2015) found from experiment that even though slider starting position could lead to different responses, slider is more preferred than radio buttoned and more from Pc than smart phone users. However, systematic studies by academic researchers 460 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY have investigated data quality in multiple ways with mixed empirical results (Roster et al., 2015). Besides the design of the survey namely length and format, motivation of participants is an intrinsic factor in determining the completion rate of a survey, particularly when it comes to self-assessments of skil s by employees. If employees do not see the value in it for their own job or career development, they are not wil ing to spend effort to help achieving the goal and need of the organization (Govender & Parumasur, 2010). They may not be motivated to fil out the survey and they may and may not take it seriously. This can lead to incomplete or inaccurate responses, which can in turn affect the usefulness of the data col ected. 3 Method The current study employs a design used by Galesic (2006) and Nestler et al. (2015), which involved dividing the survey into blocks of questions and soliciting participant evaluation of the previous blocks. However, this method was adapted with several modifications. Specifically, a button with the option to stop the survey was included at approximately one-third and two-thirds of the survey length. When clicking on the “stop” button, a message box appears showing how much of the survey was finished but that they could stop or continue if they wished. This button feature was used instead of immediate message box to prevent overt intervention of participants into ending the survey early. This feature aimed to measure the rate of survey dropouts and provide immediate feedback of the questionnaire. The feedback questions assessed the motivation level of the participants filling in the survey adopted from motivation questionnaire developed by Fiorella et al. (2021). Specifical y three feedback questions were asked with five-point Likert scale: how interesting the questions are, how wel you have answered the previous questions, are the questions relevant for your career; and one open question of why you (may) choose to stop the survey. In addition to the feedback opportunity provided if the participants choose to stop at designed places, the participants are also presented with the same feedback questions at the end of the survey if they complete the whole survey. X. Peng, R. Zhang, T. Marantika, M. Deinum, J. van der Touw, X. Lub, K. van Turnhout: Improving Completion Rate of Digital Skil Self-assessment Survey: an Empirical Study 461 This study employs three survey formats: two-option choices, a slider bar with five scales, and a five-point Likert scale radio-button survey. The latter two formats have been discussed and compared in the literature. The reason why the two-option choices format is incorporated into this study is that, from the perspective of user experience, presenting participants with scenarios, particularly those related to topic of technology, which may be unfamiliar and complex, can be difficult to answer. The survey is divided into three blocks of questions based on three themes: Resilience (attitude toward technology), Analytical (information handling skills), and Technology (technology skil s). Participants receive a random survey form and a random sequence of the three themes. For example, one participant may receive a slider form of survey starting with questions from the Analytical theme. Each type of survey takes approximately 15 minutes, with 32 questions for the two-option choice survey and 35 questions for both the slider and five-point Likert scale surveys. The reason for this difference is to account for the fact that the two-option survey included more text and involved more reading. An example of the three formats of surveys and three themes of questions is provided in Figure 1. Figure 1: Example of survey formats 462 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY As an incentive for colecting more data (Nestler et al., 2015), participants are offered a chance to win a €40 gift card, and they are required to leave an email address at the end of the survey for registration in the lottery. To ensure compliance with GDPR regulations that prohibit the misuse of participant data, a privacy statement is presented at the start of the survey assuring participants that their email address wil only be used for registration purposes, and al collected information wil be used for research purposes only. 4 Analysis and results After removing non-valued and duplicated data, the survey yielded a total of 121 responses out of the initial 161 data collected. Descriptions of respondents' age and gender are shown in Table 1. Morover, completetion rates for 1/3, 2/3 or the whole surevy are reported in the table. Specifically, among the 121 respondents, 20 individuals or 17% of the sample completed less than 1/3 of all questions, 10 individuals or 8% of the sample completed between 1/3 and 2/3 or all questions, while the majority (75%) of respondents completed more than 2/3 of al survey questions. This entails that most respondents complete the survey with length of around 15 mins. Table 1: Sample profile Number of respondents % Gender Male 42 34.71% Female 50 41.32% Other genders 4 3.31% Missing 25 20.66% Age Younger than 20 25 20.66% Between 20 and 22 33 27.27% Between 23 and 25 21 17.36% Older than 25 16 13.22% Missing 26 21.49% Survey completion Less than 1/3 20 16.53% Between 1/3 and 2/3 10 8.26% More than 2/3 91 75.21% Note: N = 121, age (M = 23.41, SD = 7.61). X. Peng, R. Zhang, T. Marantika, M. Deinum, J. van der Touw, X. Lub, K. van Turnhout: Improving Completion Rate of Digital Skil Self-assessment Survey: an Empirical Study 463 Completion rate1 (calculated by dividing number of answered questions by total number of questions), completion time (calculated as sum of seconds used per question), and average time per answered question (calculated by completion time divided by number of questions answered) were compared among the three different survey formats (Likert scale, slider, and two options). ANOVAs show significant differences for all three variables (see Table 2). Likert-scale surveys were completed more (94%) than slider surveys (70%) and two-option surveys (72%). In terms of completion time, two-option surveys took the longest time (480 seconds), followed by Likert-scale surveys (366 seconds), and then slider surveys (225 seconds). Average time per answered question was the longest for two-option surveys (23 seconds per answered question), fol owed by Likert scale (11 seconds per answered question), and then slider (10 seconds per answered question). Table 2: Analysis of variance on survey completion among three survey formats2 Response type N M SD F Completion Likert scale 35 93.88% 11.68% 7.13** rate Slider 51 70.42% 37.70% Two options 35 72.06% 30.52% Completion Likert scale 35 365.74 161.51 12.77** time Slider 51 225.39 153.40 Two options 35 480.06 356.17 Average time Likert scale 35 11.29 4.91 22.14** per answered Slider 51 10.14 6.11 question Two options 35 23.37 15.51 Note: Completion time and average time per answered question measured in seconds; ** p < .01 Moreover, the differences identified in completion rate and average time per answered question are visualized in Figures 2 and 3. Also noteworthy is that two options and slider items show similar completion rates, but two option items took much longer to complete. This can be indicative that two option items, although 1 Completion rate is chosen instead of response rate in this study as it represents a meaningful indicator of the data quality. Given the research design of using a student sample, response rates were high across all surveys sent. 2 Further Bonferroni post-hoc analyses reveal that for completion rate: Likert scale is significantly higher than both slider and two-option; for average time per answered question: two-option is significantly longer than both Likert scale and slider; and for overall completion time: two-option is significantly longer than slider but not Likert scale while Likert scale is significantly longer than slider. It is also noted that overall completion time should be viewed in the context of the other two variables as completion time itself is contingent upon the number of questions each respondent chose to answer. 464 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY time-consuming, are associated with less survey fatigue compared to the slider format. Taken together, the findings shown in the Figures as wel as Table 2 indicate that for the administered survey among this sample, Likert-scale items led to the highest level of completion as respondents who were assigned this version of the survey answered the most questions and used relatively little time per answered question. For completion time, Likert scale response type lies in the middle. For both completion rate and average time per answered question, Likert scale is also associated with the lowest standard deviation (see Table 2), indicating a more consistent level of performance among respondents who were assigned surveys of this response type. Figure 2: Completion rates among surveys of three survey formats Figure 3: Average time in seconds per answered question among surveys of three survey formats X. Peng, R. Zhang, T. Marantika, M. Deinum, J. van der Touw, X. Lub, K. van Turnhout: Improving Completion Rate of Digital Skil Self-assessment Survey: an Empirical Study 465 Additionaly, average time per answered question and completion rate are significantly and negatively correlated ( r = -.29, p < .01), indicating those respondents who completed more questions used relatively less time per question; another interpretation of this correlational finding is that those who were able to spend less time per question tended to complete the survey more ful y. Motivation of the participants is checked, especial y because this survey is about employees' digital skills self-assessment, it is difficult to get response or honest response, may due to the reason that employees are not motivated. This is why the feedback questions regarding motivation were asked, the aim is to understand the relation between motivation of the participants and response, therefore motivate participation. Correlation from these three feedback questions with significant results are presented. First, the three feedback questions (questions being interesting, previous questions well answered, and questions being relevant for future career) are significantly and positively correlated among each other (see Table 3). This confirmed that the items chosen for measuring motivation level are reliable factors. Table 3: Correlations 1 2 3 1 Questions interesting 2 Previous questions wel answered .61** 3 Questions relevant for career .41** .37** Note: ** p < .01 A regression analysis is performed, using the three theoretical questions as predictors and completion rate as the dependent variable. It should be noted that not al respondents chose to answer the three theoretical control questions. Among the 121 respondents in this sample, 64 responded to these three questions. The outcome of this regression analysis shows no significant relationship between motivation questions and completion of the survey among participants in this study (see Table 4). Despite the inclusion of an open-ended question inquiring about the reasons for survey discontinuation, only a limited number of respondents, specifical y 22 out of 121, provided a response. This lack of responses can be attributed to the fact that most respondents did complete the survey to a large extent. Of those who did respond, the most commonly given reasons for discontinuation included: (1) 466 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY perceived irrelevance of the questions, (2) excessive number of questions, and (3) complexity and length of the survey items, which posed difficulties in reading and comprehending the survey content. These findings highlight the significance of designing surveys that are concise, relevant, and easy to navigate, to ensure optimal participation and minimize attrition. Table 4: Regression analysis with completion rate as dependent variable Predictor B SE Beta t p Intercept (constant) .82 .06 14.23 .00 Questions interesting .02 .02 .15 .92 .36 Previous questions wel answered .02 .02 .09 .55 .58 Questions relevant for career -.01 .01 -.11 -.81 .42 Note: R = .20, R2 = .04, F = .81 ( ns). 5 Conclusions and implications The results indicate that more than 75% participants complete the survey with the estimated length of 15 mins, this confirms the literature that the average length of 15 mins is appropriate supporting high completion rate. However, the incentive given in this study may contribute to the high completion rate as wel . In the real self-assessment survey, a relevant reward could be considered for improving completion rate. For example, the employees would receive free career consult and training for self-development. This study also shows that the formats of surveys significantly impact completion rates and completion time. Specifical y, Likert scale surveys are completed more than two-option surveys and slider surveys. In terms of completion time per answered question, Likert-scale surveys involve less time than two-option surveys and only marginally more time than slider surveys. Likert scale is also associated with the lowest standard deviation among the three survey formats, indicating a more consistent level of performance among respondents who were assigned surveys of this response type both in the amount of time needed to complete the survey and the extent to which they completed the survey. This suggests that survey designers should consider using Likert scale to optimize completion rates and completion time. However, this conclusion is only based on the completion rates of this 121-student sample, it does not refer to any other factors. One concern that some X. Peng, R. Zhang, T. Marantika, M. Deinum, J. van der Touw, X. Lub, K. van Turnhout: Improving Completion Rate of Digital Skil Self-assessment Survey: an Empirical Study 467 participants tend to skip the later part of the survey by choosing middle (Likert scale) or 4 (in slider), whether they are not engaged in the survey, influenced by the incentive or it has to do with less relevance because they are students rather than employees, should be investigated further in future study. In addition, given the longer time per answered question for two-option surveys, survey designers ought to consider using this survey format only when it is required (e.g.: complex reasoning is involved or when respondents must give a dichotomy answer). Even then, it is important to note that two-option format may still result in lower completion rates and longer overal completion time. Moreover, the negative correlation between completion rate and average time per answered question suggests that reducing the time spent per question can increase completion rates. To improve completion rates, survey designers may consider simplifying survey items and questions, thereby reducing their length and complexity. However, this study does not offer conclusive evidence of a relationship between motivation and completion rate. To increase motivation among employees, it can be helpful to clearly communicate the purpose and benefits of the survey, such as how the data wil be used to improve employee development programs or support career growth. It can also be useful to emphasize the confidentiality and anonymity of responses to encourage employees to provide honest feedback. Additional y, offering incentives such as a free career consult and training can provide additional motivation for employees to participate. Overal , by emphasizing the importance and benefits of the survey, employers can help increase employee motivation and participation, resulting in more useful data and better outcomes for all involved. Finally, the findings and conclusions of this study should be viewed in light of two potential limitations. The first limitation concerns the sample group and sample size. While the sample of students provides insightful findings, it may only partially reflect the behaviors of employees. Therefore, future research utilizing a larger employee sample is recommended to validate the generalizability of the present study's findings. Moreover, with a sample size of only 121, divided into three response type groups, the current study's sample size is limited, which further emphasizes the need for larger samples in future research. However, it is worth noting that the current 468 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY study still produced statistically significant results, which mitigates some of the concerns associated with the sample size limitation. The second limitation pertains to the potential presence of common method bias. As al variables used in the correlation and regression analyses are self-reported by the sample, there exists a possibility of common method bias. To combat this issue, future studies could incorporate research designs that utilize multiple data sources. It is essential to note, though, that the correlation and regression analyses served primarily as supplementary analysis, and the key conclusions of this study are based primarily on the ANOVA results. Despite the above mentioned two potential limitations, this study provides valuable insights into completion rates of self-assessment surveys. The findings indicate that survey designers should give careful consideration to survey length and format to optimize completion rates and completion time, with the use of Likert scales showing promise in enhancing survey completion. While the study did not find evidence of a significant relationship between motivation and completion rates, the results underscore the importance of reducing survey complexity and length to promote engagement and participation. These findings have implications for the design and implementation of self-assessment surveys in the workplace, highlighting the need to prioritize user experience and streamline survey design to maximize data quality and promote workforce development. Overall, this study contributes to the growing body of literature on survey design and completion rates, offering practical insights for HR managers and researchers seeking to improve the use of self-assessment surveys as a tool for evaluating digital skills in the workplace. References Allen, J., & van der Velden, R. (2005). The Role of Self-Assessment in Measuring Skil s. http://www.reflexproject.org.2 Andrade, H. L. (2019). A Critical Review of Research on Student Self-Assessment. Frontiers in Education, 4, 87. https://doi.org/10.3389/FEDUC.2019.00087/BIBTEX Beebe, T. J., Rey, E., Ziegenfuss, J. Y., Jenkins, S., Lackore, K., Tal ey, N. J., & Locke III, R. G. (2010a). Shortening a survey and using alternative forms of prenotification: Impact on response rate and quality. In BMC Medical Research Methodology (Vol. 10). http://www.biomedcentral.com/1471-2288/10/50 Beebe, T. J., Rey, E., Ziegenfuss, J. Y., Jenkins, S., Lackore, K., Tal ey, N. J., & Locke III, R. G. (2010b). Shortening a survey and using alternative forms of prenotification: Impact on response rate and quality. In BMC Medical Research Methodology (Vol. 10). http://www.biomedcentral.com/1471-2288/10/50 X. Peng, R. Zhang, T. Marantika, M. Deinum, J. van der Touw, X. Lub, K. van Turnhout: Improving Completion Rate of Digital Skil Self-assessment Survey: an Empirical Study 469 Benraghda, A., Mohd Radzuan, N. R., & Lardhi, F. A. S. (2022). Self-assessment as a self-regulated learning approach in English oral presentations: Col ege students’ choices and perceptions. Http://Www.Editorialmanager.Com/Cogentedu, 9(1). https://doi.org/10.1080/2331186X.2022.2123472 Brace, I. (2013). Questionnaire design : how to plan, structure and write survey material for effective market research. Buskirk, T. D., Saunders, T., & Michaud, J. (2015). Are Sliders Too Slick for Surveys? An Experiment Comparing Slider and Radio Button Scales for Smartphone, Tablet and Computer Based Surveys. Methods, Data, Analyses, 9(2), 32. https://doi.org/10.12758/mda.2015.013 Deutskens, E., De, K. O., Nl, R. K. D. U., Oosterveld, P., & Nl, P. O. (2004). Response Rate and Response Quality of Internet-Based Surveys: An Experimental Study. In Marketing Letters (Vol. 15, Issue 1). Dobronte, A. (2015, June 27). Likert scales vs. slider Scales in commercial market research. Fiorel a, L., Yoon, S. Y., Atit, K., Power, J. R., Panther, G., Sorby, S., Uttal, D. H., & Veurink, N. (2021). Validation of the Mathematics Motivation Questionnaire (MMQ) for secondary school students. International Journal of STEM Education, 8(52). https://doi.org/10.1186/s40594-021-00307-x Galesic, M. (2006). Dropouts on the Web: Effects of Interest and Burden Experienced During an Online Survey. Journal of Official Statistics, 22, 313–328. Galesic, M., & Bosnjak, M. (2009). Effects of questionnaire length on participation and indicators of response quality in a web survey. Public Opinion Quarterly, 73(2), 349–360. https://doi.org/10.1093/poq/nfp031 Govender, S., & Parumasur, S. (2010). The Relationship between Employee Motivation and Job Involvement. SAJEMS NS, 13(3), 237–253. Hoerger, M. (2010). Participant Dropout as a Function of Survey Length in Internet-Mediated University Studies: Implications for Study Design and Voluntary Participation in Psychological Research. Cyberpsychology, Behavior, and Social Networking, 13(6), 697–700. Marcus, B., Bosnjak, M., Lindner, S., Pilischenko, S., & Schütz, A. (2007). Compensating for Low Topic Interest and Long Surveys A Field Experiment on Nonresponse in Web Surveys. Social Science Computer Review, 25, 372–383. https://doi.org/10.1177/0894439307297606 Nestler, S., Thielsch, M., Vasilev, E., & Back, M. D. (2015). Will They Stay or Will They Go? Personality Predictors of Dropout in an Online Study. International Journal of Internet Science, 10(1), 37– 48. Noorka, I. R., & Sarwar, M. (2020). Self-Assessment for Students. 760–769. https://doi.org/10.1007/978-3-319-95870-5_90 Parsons, C. (2007). Web-Based surveys: Best practices based on the research literature. Visitor Studies, 10(1), 13–33. https://doi.org/10.1080/10645570701263404 Revilla, M., & Ochoa, C. (2017). FORUM Ideal and maximum length for a web survey. International Journal of Market Research, 59, 557. https://doi.org/10.2501/IJMR-2017-039 Roster, C. A., Lucianetti, L., & Albaum, G. (2015). Exploring Slider vs. Categorical Response Formats in Web-Based Surveys. Journal of Research Practice, 11(1), 1–19. http://jrp.icaap.org/index.php/jrp/article/view/509/413 Savage, S. J., & Waldman, D. M. (2008). Learning and fatigue during choice experiments: A comparison of online and mail survey modes. Journal of Applied Econometrics, 23(3), 351–371. https://doi.org/10.1002/jae.984 Stanley, N., & Jenkins, S. (2007). Watch What I Do! Using Graphic Input Controls in Web Surveys. Fifth International Conference of the Association for Survey Computing. http://www.asc.org.uk 470 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY USING SIMULATORS TO ASSIST WITH HEALTHCARE ISSUES: THE IMPACT OF A SAILING SIMULATOR ON PEOPLE WITH ADHD GURDEEP SARAI,1 OREN TIROSH,1 PREM PRAKASH JAYARAMAN,2 PETER BROOKS,3 NORMAN SAUNDERS,4 NILMINI WICKRAMASINGHE5 1 Swinburne University of Technology, Faculty of, Health Science, Melbourne, Australia gurdeep.sarai@gmail.com, otirosh@swin.edu.au 2 Swinburne University of Technology, School of Science, Computing and Engineering Technologies, Melbourne, Australia pjayaraman@swin.edu.au 3 University of Melbourne, Faculty of Medicine, Dentistry and Health Sciences, Melbourne, Australia brooksp@unimelb.edu.au 4 University of Melbourne, Faculty of Neuroscience, Melbourne, Australia n.saunders@unimelb.edu.au 5 Swinburne University of Technology, Faculty of Health Science/Digital Health, Melbourne, Australia nwickramasinghe@swin.edu.au The use of simulators has grown a vast amount in the past few years, with various research reports on the advantages that come with the use of this technology. However, the use of simulators as a form of treatment/aids have yet to be ful y embraced. Given the growing diffusion of technology, many people can benefit from simulators that are more accessible and improve specific Keywords: qualities to help them cope with their deficiencies. To examine simulators, this current void, this research in progress examines the benefits ADHD, physical of physical activity for people with attention deficit hyperactivity activity, disorder (ADHD) through the use of a VSail Sailing Simulator. intervention, Specifical y, we contend that there are benefits to addressing sailing simulation, strength and balance. Thus, this paper analyzes the ramifications motor of the findings, along with knowledge gaps and a research plan. function DOI https://doi.org/10.18690/um.fov.4.2023.30 ISBN 978-961-286-751-5 472 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The COVID-19 pandemic has caused a significant increase in mental health issues, leading to a surge in the number of adults being prescribed ADHD medication in Australia. According to the Australian Department of Health and Aged Care, the number of patients receiving ADHD medication has grown at an annual rate of 12.43%, with a higher growth rate of 16% observed from 2018-2020. This growth rate is almost double that of the growth rate from 2014-2017 (9%) and serves to highlight the importance to address this issue. The number of prescriptions has also increased, with an average annual growth rate of 10.25% and the highest growth rate of 17.67% observed from 2018-2020. As a result, more adults are now receiving ADHD medication than children. Attention-deficit hyperactivity disorder (ADHD) is a neurodevelopmental disorder that affects individuals across the lifespan. The global prevalence of ADHD in children is estimated to be around 5%, while in adults, it is around 3-4% (Polanczyk and Rohde, 2007, Shaw et al. 2007, AADPA). In early childhood, symptoms of ADHD may include impulsiveness, hyperactive behaviour, and low frustration tolerance. Longitudinal neuroimaging studies indicate that children with ADHD may be 2-3 years behind their peers in development. In adulthood, individuals with ADHD may experience memory problems, restlessness, and difficulty with mental focus. Notably, 4-11% of university students exhibit symptoms of ADHD (AADPA). Separate from this, yet contemporaneous, have been the advances in technology and simulation solutions. In this paper, we outline research in progress work that aims to address the larger key research question of “What is the degree of fidelity with which a sailing simulator can reproduce the health benefits associated with real-life sailing?”. This research in progress paper contribution is towards the feasibility of using inertial measuring units (IMU) to identify hand movement patterns and motor assessments that are beneficial in real-life sailing. The primary objectives of the research-in-progress paper are: 1. Objective 1: To measure the effect of sailing simulation on grip strength and postural balance in non-ADHD and ADHD participants G. Sarai, O. Tirosh, P. Prakash Jayaraman, P. Brooks, N. Saunders, N. Wickramasinghe: Using Simulators to Assist with Healthcare Issues: The Impact of a Sailing Simulator on People with Adhd 473 2. Objective 2: To measure the acceleration patterns of both hands in a sailing simulation between ADHD and non-ADHD students. 3. Objective 3: To understand the participant's perception of ADHD therapies and simulators as a part of their therapy. The research design employs a grounded theory mixed methods approach, incorporating both quantitative and qualitative methodologies. The quantitative aspect focuses on examining the effectiveness of IMUs in measuring various aspects of sailing simulation, including motor assessments. Whilst the qualitative perspective delves into exploring the perceptions and attitudes of individuals with ADHD towards ADHD therapy and the use of simulation as therapy. By combining both quantitative and qualitative data, we can understand how sailing simulation can influence grip strength and postural balance, along with the perception and acceptance of sports simulators as a therapeutic intervention for individuals with ADHD. Preliminary testing has been conducted to consider al critical IMU aspects of the proposed research plan. 2 Related Work 2.1 Physical Activity and ADHD Regulating the course of neural development has the potential to improve ADHD symptoms, suggesting that interventions targeting neural growth and development wil be more effective than other treatments. Non-pharmacological treatments such as cognitive-based therapies or physical activity have demonstrated effectiveness. Engaging in physical exercise is one of the most advantageous strategies for treatment, providing numerous benefits and reducing symptoms (Xie et al. 2021). Physical therapy leads to enhanced coordination and motor function, while mental health improvements include processing speed, selective attention, and cognitive flexibility (Montalva-Valenzuela et al. 2022, Watemberg et al. 2007). However, chal enges exist in introducing physical activity as a treatment for ADHD, with dropout rates of 17.5% (Vancampfort et al, 2016). This suggests a need for better engagement strategies and further studies to explore other forms of physical activity (Carta et al, 2014). 474 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 2.2 Physical Activity based Interventions Physical exercise has a positive impact on health and wel -being, including improving mood, quality of life, and mitigating stress. Research on physical activity and ADHD supports using exercise as an intervention strategy. Aquatics exercise programs and racquet sports have shown efficacy in improving cognitive, behavioural, and motor functions in children with ADHD. Sailing also offers numerous benefits for mental and physical health, including enhancing concentration, communication, mental wellness, endurance, and muscle strength. It has shown positive outcomes in improving quality of life, global functioning, social skills, outlook on life, and mental and physical health for individuals with disabilities and severe mental illnesses. Sailing may be a promising intervention for enhancing wel -being, particularly for those with mental il nesses, and may be more engaging than standard rehabilitation methods. 2.3 Simulator Based Interventions Simulators are effective tools for adaptive training, which adjusts the task's complexity to the user's skill level. They can break down cognitive activities into components that are difficult to train individually. For example, simulators can introduce factors involved in sailing gradual y, preventing the operator from becoming overwhelmed. Studies using a horse-riding simulator have yielded positive results in various populations. A study by Borges et al. (2011) revealed that children with cerebral palsy spastic diplegia showed greater emotional happiness and overal satisfaction with the simulator than with conventional physical therapy. The simulator was also statistically superior in increasing postural control in a seated position compared to conventional therapy. Other studies have demonstrated the simulator's potential to enhance physiological functioning in elderly individuals, such as postural control, muscle activation, and dynamic stability (Mitani et al. 2008, Kim and Lee, 2015). G. Sarai, O. Tirosh, P. Prakash Jayaraman, P. Brooks, N. Saunders, N. Wickramasinghe: Using Simulators to Assist with Healthcare Issues: The Impact of a Sailing Simulator on People with Adhd 475 3 Research Design 3.1 Participants Study participants are divided into two groups: "non-ADHD" students and "ADHD" students aged 18-30. Both groups are screened for physical conditions and injuries. The ADHD group includes students with self-diagnosed ADHD. Participants complete the APSS Screening Tool to determine those more vulnerable to adverse events caused by exercise. The ASRS-V1.1 Symptom Checklist is used for self-diagnosis of ADHD symptoms in the ADHD group. 3.2 Apparatus Sailing Simulator: The simulator itself is a VSail-Trainer®, designed by the company Virtual Sailing Pty Ltd. It comprises one boat hull (size length: 230 cm, breadth: 150 cm). The simulator al ows sailors to control the course and speed of the boat using a joystick and mainsheet, whilst being in a seated setup suitable for disabled individuals. The simulator has a range of tools and functionalities for researchers to design experiments around, such as the ability to adjust environmental conditions, boat characteristics, and sailor behaviours. The Inertial Measuring Unit (IMU) is a form of accelerometer, composed of an electromechanical sensor that is designed to measure dynamic acceleration. In this case, two IMUs are placed on the back of each hand of the participant. This wil look at the change in velocity for the rudder movement and mainsheet. The Jamar hand dynamometer wil be used to measure isometric grip force in the participants pre- and post-trial as part of their motor assessment. 3.3 Movement Assessment Studies suggest that postural instability in adults with ADHD may contribute to difficulties with motor coordination and everyday activities related to balance and postural control. Grip strength differences in individuals with ADHD may be influenced by a combination of motor coordination deficits, attentional deficits, and medication use. The relevance of grip strength and postural balance in real-life sailing 476 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY has been shown to improve both areas. Participants wil undergo a pre and post- simulator training movement assessment. The first motor assessment test wil be the strength component (Hove, et al. 2015, Jansen, et al.2019, Jeoung, et al. 2014, Neely, et al. 2017, Clarke, et al. 2020). The Jamar hand dynamometer will be used to assess grip strength before and after each scenario. Balance wil be assessed through a force plate or through a Balance Error Scoring System to assess overal static balance before and after each scenario. 3.4 Interviews Semi-structured interviews can provide valuable insights into the perception and acceptance of sports simulators as a therapeutic intervention for individuals with ADHD. Including whether individuals with ADHD would be willing to use sports simulators regularly, and investigating potential benefits such as improved motor skills and ADHD symptoms, along with potential drawbacks. This can help with the development of more effective interventions for individuals with ADHD. The interviews are conducted with a consistent set of questions, while the order of the questions may be adjusted based on responses. 4 Data Collection 4.1 Participant setup Before testing begins each participant is briefed on the simulator, equipment, and safety features. Participants sign a consent form and wear two IMUs on their hands. After being seated in the V-Sail simulator, participants are given an adaptation period to learn how to sail using the information provided in the HUD. Steering is introduced first, followed by the main sheet, teaching participants how to adjust the main sail and coordinate the boat's speed based on wind direction. Participants also learn maneuvers such as tacking and aligning the twin tails on the mainsail to use the wind to their advantage. The final step is to introduce the pneumatic rams, which allow participants to understand how the boat handles at certain angles relative to the wind. G. Sarai, O. Tirosh, P. Prakash Jayaraman, P. Brooks, N. Saunders, N. Wickramasinghe: Using Simulators to Assist with Healthcare Issues: The Impact of a Sailing Simulator on People with Adhd 477 4.2 Sailing Program Participants will train on a Sydney 2000 Olympic Games Trapezoid course in the Sydney harbour. The course involves sailing upwind to the first buoy, then downwind past three buoys, and back upwind. The training program consists of 6 sprints per week for three weeks, with each sprint lasting 3-4 minutes on average. The sprints progressively increase in difficulty and chal enge the participants' decision-making and technique accuracy. The wind speeds wil be implemented through a blocked and serial schedule in the first and final weeks. 4.3 Data Analysis Pre and Post Movement Assessment: The measurements obtained for both sets of pre and post scenarios are evaluated. To evaluate whether the mean difference between the two sets of data varies, a paired sample t-test is performed. IMUS: 1. Average acceleration magnitude (AAM): The standard deviation of acceleration from the mean. More task stability is indicated by smal er values. 2. Root mean square (RMS): An evaluation of the fluctuating signal strength. Significant stability is indicated by smaller values. Interviews: Thematic Analysis 5 Preliminary Findings The pilot study aimed to determine if IMU placement and hand movement are appropriate measures for testing and comparing real-life sailing. The study collected data from a healthy participant using two IMUs on the back of both hands in three different wind conditions (8, 12, and 16 knots). The approach was based on a previous study by Mackie and Legg (1999) that focused on how force output varies with wind speed, finding that force on the mainsheet increases with higher winds. Mackie and Legg (1999) also found a trend between force and wind through experience level in sailing. Specifically, the mainsheet force increased for club sailors 478 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY with less experience, while the opposite trend was found for pro sailors. This is due to differences in sailing techniques found with experience as conditions get more difficult. Table 1: Differences between both hands and the wind speed, expressed as an average acceleration magnitude (AAM) and root-mean-squared (RMS) Hand 8 Knots Wind 12 Knots wind 16 Speed speed Knots wind speed Average Left (Tiller) 9.52 9.53 9.56 acceleration magnitude Right (Mainsheet) 9.76 9.76 9.8 (AAM) (m/s2) Root mean Left (Tiller) 5.51 5.52 5.54 square (RMS) Right (Mainsheet) 5.66 5.67 5.68 (m/s2) The study aims to quantify the difference in acceleration for hand movement between ADHD and healthy students. The preliminary study showed variability in both AAM and RMS parameters with increasing wind speed. At 8 and 12 knots wind speed, there were similarities in the acceleration of both hands, while 16 knots showed less stability during the sprint. The results are consistent with Mackie and Legg's study, which found higher force in the mainsheet between wind speeds of 15-20 knots. The findings of the IMU showed an increase in acceleration for both hands as wind speed increased, similar to real-life sailing. The methodology allowed for finding variability in the set conditions in the simulator. Through the research, we can observe how the movement patterns of students with ADHD differ from those of healthy students in the sailing simulator, and how these patterns progress with experience. G. Sarai, O. Tirosh, P. Prakash Jayaraman, P. Brooks, N. Saunders, N. Wickramasinghe: Using Simulators to Assist with Healthcare Issues: The Impact of a Sailing Simulator on People with Adhd 479 This study has a few parts worth considering and should help inform future research directions. With this research looking at the feasibility of IMUs future research will be focusing on using the IMUs towards the comparison of movement in real-life sailing. Future Research In this current generation of human and computer interaction, simulator use will also provide insight into values, interests, and user experience giving a good basis for how simulators are taken as a form of therapy in managing ADHD. 6 Conclusion This research in progress study has served to highlight a potential role for the use of simulators to address mental health/mental wellness issues. Specifically, we have focused on the benefits of a sailing simulator to assist students with ADHD. To date, the application of simulators to assist with addressing mental health/ mental wel ness issues is embryonic at best; but we contend that the potential benefit of this approach justifies more research in this area. We believe our study will serve as one of the first to shed light on this major area. Acknowledgements This work was supported by Swinburne University of Technology and Research under the SUPRA program, in partnership with Northern Health and Virtual Sailing. This endeavor would not have been possible without the support of the Virtual Sailing team Norman Saunders, Mark Habgood and Jonathan Binns. Along with the added support by Northern Health from Peter Brooks. References BARKLEY, R. A., FISCHER, M., SMALLISH, L. & FLETCHER, K. 2002. The persistence of attention-deficit/hyperactivity disorder into young adulthood as a function of reporting source and definition of disorder. Journal of Abnormal Psychology, 111, 279-289. BEHRMANN, J. T., BLAABJERG, J., JORDANSEN, J. & JENSEN DE LÓPEZ, K. M. 2022. Systematic Review: Investigating the Impact of COVID-19 on Mental Health Outcomes of Individuals With ADHD. Journal of Attention Disorders, 26, 959-975. BELL, D. R., GUSKIEWICZ, K. M., CLARK, M. A. & PADUA, D. A. 2011. Systematic review of the balance error scoring system. Sports Health, 3, 287-95. BORGES, M. B. S., WERNECK, M. J. D., DA SILVA, M. D., GANDOLFI, L. & PRATESI, R. 2011. Therapeutic effects of a horse riding simulator in children with cerebral palsy. Arquivos De Neuro-Psiquiatria, 69, 799-804. CARTA, M. G., MAGGIANI, F., PILUTZU, L., MORO, M. F., MURA, G., CADONI, F., SANCASSIANI, F., VELLANTE, M., MACHADO, S. & PRETI, A. 2014. Sailing for 480 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY rehabilitation of patients with severe mental disorders: results of a cross over randomized controlled trial. Clin Pract Epidemiol Ment Health, 10, 73-9. CHANG, Y. K., HUNG, C. L., HUANG, C. J., HATFIELD, B. D. & HUNG, T. M. 2014. Effects of an aquatic exercise program on inhibitory control in children with ADHD: a preliminary study. Arch Clin Neuropsychol, 29, 217-23. CLARKE, M. L., CLAPHAM, E. D. & SHIM, M. 2020. Sailing as Therapy: Adapted Sailing on Children with Disabilities. Palaestra, 34, 37-43. COTTERILL, S. T. & BROWN, H. 2018. An exploration of the perceived health, life skil and academic benefits of dinghy sailing for 9-13-year-old school children. Journal of Adventure Education and Outdoor Learning, 18, 227-241. COUSINS, M. & SMYTH, M. M. 2003. Developmental coordination impairments in adulthood. Human Movement Science, 22, 433-459. DAVOODY, S., GOESCHL, S., DOLATSHAHI, M., DAVARI-ASHTIANI, R., SAFFARPOUR, R., SODEIFIAN, F. & BRAND, S. 2022. Relation between ADHD and COVID-19: A Narrative Review to Guide Advancing Clinical Research and Therapy. Iran J Psychiatry, 17, 110-117. D. U. S. C., (DUSC). 2021. Attention Deficit Hyperactivity Disorder: Utilisation Analysis. In: DHAC (ed.). Pharmaceutical Benefits Scheme (PBS): Drug utilisation sub-committee (DUSC). ECONOMICS, D. A. 2019. The social and economic costs of ADHD in Australia. Australian ADHD Professionals Association (AADPA) GIEDD, J. N. & RAPOPORT, J. L. 2010. Structural MRI of pediatric brain development: what have we learned and where are we going? Neuron, 67, 728-34. GUADAGNOLI, M. A. & LEE, T. D. 2004. Chal enge point: a framework for conceptualizing the effects of various practice conditions in motor learning. J Mot Behav, 36, 212-24. HOVE, M. J., ZEFFIRO, T. A., BIEDERMAN, J., LI, Z., SCHMAHMANN, J. & VALERA, E. M. 2015. Postural sway and regional cerebellar volume in adults with attention-deficit/hyperactivity disorder. Neuroimage-Clinical, 8, 422-428. HOZA, B., SMITH, A. L., SHOULBERG, E. K., LINNEA, K. S., DORSCH, T. E., BLAZO, J. A., ALERDING, C. M. & MCCABE, G. P. 2015. A randomized trial examining the effects of aerobic physical activity on attention-deficit/hyperactivity disorder symptoms in young children. J Abnorm Child Psychol, 43, 655-67. JANSEN, I., PHILIPSEN, A., DALIN, D., WIESMEIER, I. K. & MAURER, C. 2019. Postural instability in adult ADHD - A pilot study. Gait & Posture, 67, 284-289. JEOUNG, B. J. 2014. The relationship between attention deficit hyperactivity disorder and health-related physical fitness in university students. Journal of Exercise Rehabilitation, 10, 367-371. KIM, S. G. & LEE, J. H. 2015. The effects of horse riding simulation exercise on muscle activation and limits of stability in the elderly. Archives of Gerontology and Geriatrics, 60, 62-65. KIRBY, A., EDWARDS, L., SUGDEN, D. & ROSENBLUM, S. 2010. The development and standardization of the Adult Developmental Co-ordination Disorders/Dyspraxia Checklist (ADC). Res Dev Disabil, 31, 131-9. LATHAN, C. E., TRACEY, M. R., SEBRECHTS, M. M., CLAWSON, D. M. & HIGGINS, G. A. 2002. Using virtual environments as training simulators: Measuring transfer. Handbook of virtual environments: Design, implementation, and applications, 403-414. MACKIE, H. W. & LEGG, S. J. 1999. Preliminary assessment of force demands in laser racing. J Sci Med Sport, 2, 78-85. MARTIN, N. C., PIEK, J. P. & HAY, D. 2006. DCD and ADHD: A genetic study of their shared aetiology. Human Movement Science, 25, 110-124. MATHIOWETZ, V., WEBER, K., VOLLAND, G. & KASHMAN, N. 1984. Reliability and Validity of Grip and Pinch Strength Evaluations. Journal of Hand Surgery-American Volume, 9a, 222-226. MERCURIO, L. Y., AMANULLAH, S., GILL, N. & GJELSVIK, A. 2021. Children With ADHD Engage in Less Physical Activity. Journal of Attention Disorders, 25, 1187-1195. G. Sarai, O. Tirosh, P. Prakash Jayaraman, P. Brooks, N. Saunders, N. Wickramasinghe: Using Simulators to Assist with Healthcare Issues: The Impact of a Sailing Simulator on People with Adhd 481 MITANI, Y., DOI, K., YANO, T., SAKAMAKI, E., MUKAI, K., SHINOMIYA, Y. & KIMURA, T. 2008. Effect of exercise using a horse-riding simulator on physical ability of frail seniors. Journal of Physical Therapy Science, 20, 177-183. MONTALVA-VALENZUELA, F., ANDRADES-RAMIREZ, O. & CASTILLO-PAREDES, A. 2022. Effects of Physical Activity, Exercise and Sport on Executive Function in Young People with Attention Deficit Hyperactivity Disorder: A Systematic Review. European Journal of Investigation in Health Psychology and Education, 12, 61-76. NEELY, K. A., WANG, P., CHENNAVASIN, A. P., SAMIMY, S., TUCKER, J., MERIDA, A., PEREZ-EDGAR, K. & HUANG-POLLOCK, C. 2017. Deficits in inhibitory force control in young adults with ADHD. Neuropsychologia, 99, 172-178. PAN, C. Y., CHU, C. H., TSAI, C. L., LO, S. Y., CHENG, Y. W. & LIU, Y. J. 2016. A racket-sport intervention improves behavioral and cognitive performance in children with attention-deficit/hyperactivity disorder. Research in Developmental Disabilities, 57, 1-10. POLANCZYK, G. & ROHDE, L. A. 2007. Epidemiology of attention-deficit/hyperactivity disorder across the lifespan. Current Opinion in Psychiatry, 20, 386-392. RECIO, A. C., BECKER, D., MORGAN, M., SAUNDERS, N. R., SCHRAMM, L. P. & MCDONALD, J. W. 2013. Use of a virtual reality physical ride-on sailing simulator as a rehabilitation tool for recreational sports and community reintegration: A pilot study. American Journal of Physical Medicine and Rehabilitation, 92, 1104-1109. SHAW, P., ECKSTRAND, K., SHARP, W., BLUMENTHAL, J., LERCH, J. P., GREENSTEIN, D., CLASEN, L., EVANS, A., GIEDD, J. & RAPOPORT, J. L. 2007. Attention-deficit/hyperactivity disorder is characterized by a delay in cortical maturation. Proceedings of the National Academy of Sciences of the United States of America, 104, 19649-19654. SHAW, P., GILLIAM, M., LIVERPOOL, M., WEDDLE, C., MALEK, M., SHARP, W., GREENSTEIN, D., EVANS, A., RAPOPORT, J. & GIEDD, J. 2011. Cortical Development in Typically Developing Children With Symptoms of Hyperactivity and Impulsivity: Support for a Dimensional View of Attention Deficit Hyperactivity Disorder. American Journal of Psychiatry, 168, 143-151. TENBRINK, T. & DYLLA, F. 2017. Sailing: Cognition, action, communication. Journal of Spatial Information Science, 3-33. VANCAMPFORT, D., FIRTH, J., SCHUCH, F. B., ROSENBAUM, S., PROBST, M., WARD, P. B., VAN DAMME, T., DE HERT, M. & STUBBS, B. 2016. Dropout from physical activity interventions in children and adolescents with attention deficit hyperactivity disorder: A systematic review and meta-analysis. Mental Health and Physical Activity, 11, 46-52. WATEMBERG, N., WAISERBERG, N., ZUK, L. & LERMAN-SAGIE, T. 2007. Developmental coordination disorder in children with attention-deficit-hyperactivity disorder and physical therapy intervention. Dev Med Child Neurol, 49, 920-5. WILENS, T. E., BIEDERMAN, J., FARAONE, S. V., MARTELON, M., WESTERBERG, D. & SPENCER, T. J. 2009. Presenting ADHD Symptoms, Subtypes, and Comorbid Disorders in Clinically Referred Adults With ADHD. Journal of Clinical Psychiatry, 70, 1557-1562. XIE, Y. T., GAO, X. P., SONG, Y. L., ZHU, X. T., CHEN, M. G., YANG, L. & REN, Y. C. 2021. Effectiveness of Physical Activity Intervention on ADHD Symptoms: A Systematic Review and Meta-Analysis (vol 12, 706625, 2021). Frontiers in Psychiatry, 12. 482 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY THE ANATOMY OF A PERSONAL SERVICE: THE EIGHT DIMENSIONS OF 'PERSONAL' JEROEN VAN GRONDELLE,1 MARLIES VAN STEENBERGEN,1 ALETTA SMITS, TIJS TIMMERMAN,3 KOEN VAN TURNHOUT,2 HARALD POL3 1 HU University of Applied Sciences Utrecht, Digital Ethics, Ultrecht, Netherland jeroen.vangrondel e@hu.nl, marlies.vansteenbergen@hu.nl, 2 HU University of Applied Sciences Utrecht, Human Experience & Media Design, Ultrecht, Netherland aletta.smits@hu.nl, koen.vanturnhout@hu.nl 3 HU University of Applied Sciences Utrecht, Marketing & Customer Experience, Ultrecht, Netherland tijs.timmerman@hu.nl, harald.pol@hu.nl Digitalization enables public organizations to personalize their services, tuning them to the specific situation, abilities, and preferences of the citizens. At the same time, digital services can be experienced as being less personal than face-to-face contact by citizens. The large existing volume of academic literature on personalization mainly represents the service provider Keywords: perspective. In contrast, in this paper we investigate what makes personalisation, citizens experience a service as personal. The result are eight personal dimensions that capture the full range of individual experiences service, public and expectations that citizens expressed in focus groups. These services, dimensions can serve as a framework for public sector digitalisation organizations to explore the expectations of citizens of their own customer experience, services and identify the areas in which they can improve the Bled personal experiences they offer. eConference DOI https://doi.org/10.18690/um.fov.4.2023.31 ISBN 978-961-286-751-5 484 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction One of the promises of digitalization of data-intensive services is the ability to deliver personalized services that lead to personal customer experiences, catering to the unique needs, preferences, and situations of individual consumers. The expectation is that such personalized services wil result in enhanced customer experiences and yield better outcomes for service providers (Chandra et al., 2021). However, since the advent of digital services, private and public sector organizations have been struggling to offer their customers an authentic personal experience in the digital age. It has been shown that personalization does not always lead to personal service (Pol et al., 2020). Ironically, excessive personalization in digital services often misses the mark by failing to make customers feel valued and, instead, causing alienation (White et al., 2008; Teeny et al., 2021; Riegger et al., 2021; Kim et al., 2022). Apparently, organizations have yet to grasp what constitutes a genuinely personal service to customers. Designing personalized digital services that are truly personal is a complex chal enge, with at its heart the question of what customers or citizens perceive as ‘personal’ in a service interaction. Although there is a proliferation of literature on personalization, by far most studies are concerned with the technical aspects or, to a lesser extent, with the impact of personalization on the effects aimed for by the organization (Zanker et al., 2010; Sunikka & Bragge, 2008). The voice of the citizen in terms of what makes a service personal to them is hardly heard, let alone taken as starting point. This lacking perspective represents a significant gap in existing literature. This paper aims to address this gap within the context of public organizations. To explore experiences with and expectations of public services being personal, we analysed 12 interviews with managers of digital services provided by public organizations (i.e. tax services, utility companies, municipalities) and 16 focus group sessions with citizens who had interacted with those public organizations. The resulting analysis led to eight dimensions that describe the anatomy of a truly personal service from a citizen’s perspective rather than an organizational perspective. We consider these eight dimensions a starting point for a framework that organizations can employ to assess the personal quality of their services, and to drive potential redesign of their personalized services. J. van Grondel e, M. van Steenbergen, A. Smits, T. Timmerman, K. van Turnhout, H. Pol: The Anatomy of a Personal Service: the Eight Dimensions of 'Personal' 485 In this paper, section 2 delves into the history and current usage of the concepts of personalized versus personal service, while section 3 outlines the context and research methodology of our study. The results are presented in section 4, followed by a discussion of the findings and limitations in section 5. Section 6 presents the conclusions. 2 Personalized and personal The concept of personalization was original y introduced by Surprenant & Solomon (1987) as a means of describing the transformation of formal customer-organization interactions into more intimate and personal ones, which recognize the customer as a unique individual. In their seminal work, Surprenant & Solomon differentiated between outcome personalization, which is focused on the end result for the customer, and process personalization, which is concerned with the customer journey. Process personalization, in turn, can be classified into two types: programmed personalization, which is intended to convey the appearance of personalized service (e.g., a “Dear {First name}-email”), and customised personalization, which involves making genuine efforts to find ways to best help each individual customer. Today, "personalization" typical y denotes interactions in digital contexts that use data to tailor the services to the characteristics of the consumption experience (e.g., the consumer, context, and history) (Chandra et al., 2022). The majority of the literature on personalization of services focuses on the technical interventions that make a service more personalized (Fan & Poole, 2006; Morana et al., 2017). Tuning a service to the needs of the service consumer requires knowledge. Various types of knowledge can be used to improve the experience. Examples are knowledge of the situational circumstances in which interaction takes place, knowledge of the service consumer's goals and state of mind, and knowledge of their personal characteristics (Fischer, 2012; Barwitz & Maas, 2018). Systems possessing this kind of information, however, also entails potential risks. Firstly, in research carried out in commercial contexts, when automatic processes present a consumer with a supposedly relevant selection of all content based on available data, this selection inherently reflects some degree of subjectivity as well as nudging, which impairs consumer autonomy (Mittelstadt et al., 2016). Secondly, the consumer can develop 486 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY personal reactance, a situation where excessive personalization can lead to psychological resistance to subjectively inappropriate personalization (White et al., 2008). In such cases, consumers experience too much familiarity with their preferences and behaviours. Utilizing consumer data alone is, therefore, insufficient to create true personal services. A different approach to assessing the personal quality of a service is to consider it in terms of relational models. Fiske (1991) proposes the Relational Models Theory, which identifies four fundamental relational models that people use unconsciously in their interactions with each other, and with organizations (McGraw & Tetlock, 2005): communal sharing, market pricing, authority ranking, and equality matching. Communal sharing represents a common interest and a desire to connect, and it is the most intimate of the four models, thus closest to the concept of personal service. Pol et al. (2020) demonstrated that the presence of communal sharing has a positive effect on customer experience and satisfaction. As communal sharing increases, positive emotions are amplified, and negative emotions are mitigated. In our paper, we adopt the term "personal" to refer to how individuals perceive a service, recognizing that this perception is subjective and unique to each person. This has important implications. Firstly, it underscores that there is no universal solution for personalization. Secondly, it implies that conventional metrics for personalization success, such as engagement, conversion, and convenience, may not capture the essence of a personal experience, as they fail to account for the citizen’s true experience (Sunikka & Bragge, 2008; Zanker et al., 2010). Thirdly, it suggests that the only reliable way to determine whether a service is personal is to ask the consumer of the service. Accordingly, this study does exactly that. 3 Research Method This study was performed within the context of a research project conducted in the Netherlands on the effects of digital transformation on public services. For an earlier study within this project both public service providers and citizens, as the users of their services, were involved in interviews and focus group sessions about digital service experiences and expectations (see table 1 for specifics on the participants). For the current study we returned to the data col ection of this study, using the transcripts of the recordings as a secondary data source. We identified passages in J. van Grondel e, M. van Steenbergen, A. Smits, T. Timmerman, K. van Turnhout, H. Pol: The Anatomy of a Personal Service: the Eight Dimensions of 'Personal' 487 the transcripts of both interviews and focus groups in which participants speak about personal services and what aspects makes these services personal to them. The research design included two stages. In the first stage, 12 transcripts of organizational interviews were subjected to open coding, focusing on those parts where the organizational representatives discussed the expectations of their citizens with regard to the personal quality of their services. We focused our coding on fragments where the organizational representatives discussed actual feedback from and experiences with their customers, for instance through customer panels. The saturation point was reached after analysing 12 organizational interviews: no more new themes or dimensions were identified. Table 1: Participants on organizational and individual level Number of Number of organizational citizen focus interviews groups (number of (number of participants) participants) National government services (e.g. DMV, 5 (6) 4 (11) tax, foreign affairs, student loans) Municipalities (large and smaller counties) 3 (3) 4 (11) Public transport companies 1 (2) 4 (11) Insurances and pension funds 3 (5) Utility companies 1 (2) 4 (11) The second stage focused on validating these dimensions with relevant passages in the focus groups, using template coding. As the original study was exploratory in nature, the interviews were semi-structured according to a topical interview guide. Participants were invited to bring up experiences and observations they deemed relevant and react to each other’s contributions. To ensure the quality and consistency of our data set, we devised the following protocol on identifying relevant passages. The first step consisted of identifying parts of the interviews that explicitly talked about or asked for qualifications and assessments of personal. Those parts were then divided into passages to be coded. The first four transcripts were processed by two researchers independently, to validate identification and 488 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY partitioning processes, and the criteria used for both were calibrated. The other transcripts were then divided between two researchers and processed individual y. As the interviews were wide ranging, strict criteria were used for establishing whether an individual passage was to be coded: 1) al passages that were direct responses to questions that pertained explicitly to the personal qualities of services were considered for coding; 2) other passages in the interview were only considered if direct references were made to a service being personal; 3) only direct responses to questions of the interviewer and/or passages concerning a participant’s own experience or expectations were considered. Passages where participants reflected on each other’s experiences or in which they hypothesized on what people in general might prefer were excluded. All transcripts were first coded by two researchers independently and then compared. Coding differences were discussed and resolved. 12 of the available 16 focus group transcripts were analysed, after which saturation occurred. No new dimensions were found. 4 Results The analysis of the interviews with organizations and citizen focus groups led to eight dimensions that capture the ful range of citizen’s experiences and expectations of personal services. These dimensions are time, location, language, complexity, individual relevance, confidence to act, empathy and autonomy (see Table 2; the examples in this table are translated from Dutch for the benefit of this paper). Under time we classified al statements that pertained to aspects such as opening hours, flexibility in when to get in touch, and the amount of time needed for a service. Location pertains to statements concerning the locations where the service can be received and for instance the amount of travel time needed to get to a physical location of the service provider. Language refers to inclusiveness of the verbal statements (use of jargon, available languages), but also mode (spoken, written, visual, etc.) and tone of voice as well as form of address. Complexity of the task was mentioned, for instance, in relation to situations where the process was more complex than the underlying task or issue (an official document that could only be sent via physical mail and not digital y, felt cumbersome to one of the respondents). Individual relevance refers to al situations where processes, answers or suggestions J. van Grondel e, M. van Steenbergen, A. Smits, T. Timmerman, K. van Turnhout, H. Pol: The Anatomy of a Personal Service: the Eight Dimensions of 'Personal' 489 (generated for instance on the citizen’s known history) were felt to be truly geared to the situation. Table 2: Eight dimensions of personal services Dimension Definition Examples of aspects Quotes Time The extent to which the • The times of the day the service “And make moments when and the is available and whether it is just a little duration in which the offered at a time that suits more time service (can) take place citizens for that. • are experienced as being The freedom of choice that citizens are given (“fixed cal And don't appropriate. time, schedule an appointment rush.” yourself or just walk in”) • Sufficient time is taken for the service • The service is fast/on time (relative to expectations) Location The extent to which the • The location from which the “So the first location where the service can be used time you service (can) be • The distance one must travel to had to go to delivered is experienced purchase the service the town as being appropriate. hal after birth, now it could all be done digitally.” Language The extent to which the • The language spoken “I think form in which the • The tone of voice and manner being polite, communication about of address is very the service (can) take • Visual, spoken, written important to place is experienced as • Use of jargon me.” being appropriate. Complexity The extent to which the • Whether the complexity of the “If it is mental effort required to process is in proportion to the complicated do or understand complexity of the underlying then it is something is task or issue fine, they • experienced as being Whether the time it takes to acquire a service is in should take appropriate. proportion to the complexity of their time” the task. • Individual The extent to which the • Solutions, alternatives, help, “I don't Relevance service is experienced as tips, that match the situation want a being in line with the general current context. answer, but 490 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY • Known history (contact an answer moments, usage and specific to consumption behaviour) my • Proactive suggestions that meet situation.” the needs of citizens Confidence The extent to which • Feeling confident that past “ So, then to act someone experiences actions or actions to be taken you can say that the service provides led or will lead to a solution or again: ‘oh an action perspective. to next steps towards a solution. that means • The felt need for confirmation • Interim feedback that I that and that. ’, and you don't have that option if you look at something on the website.” Empathy The extent to which • Providing space to tell the story “Think someone experiences • Showing interest along with that they are being seen • Summarizing and checking someone. and heard and that their • Thinking along and providing Try to concerns are taken solutions and alternatives empathize a seriously. • Expressing understanding little bit and that has to do with caring.” Autonomy The extent to which • Outcome is not predetermined “In different someone experiences • Not being forced into an il - ways, eh, that they can shape or fitting pattern there are influence the service in • Being able to act in accordance different terms of process and with yourself options. outcome. You can also make an appointment without a DigiD” J. van Grondel e, M. van Steenbergen, A. Smits, T. Timmerman, K. van Turnhout, H. Pol: The Anatomy of a Personal Service: the Eight Dimensions of 'Personal' 491 Confidence to act is about how much someone feels that the service gives them enough information to proceed (clear feedback, transparent information, wel -written answers). Empathy captures al statements where citizens felt the organization truly cared, and autonomy, final y, is about whether someone feels that the outcome of an interaction is predetermined or that they are being forced into an inflexible process. Al eight dimensions were strongly represented in both data sets (organizational and citizen focus groups), with language, individual relevance and empathy being coded relatively most frequently, and location and confidence to act less frequently. Passages that led to discussions fal into three clusters. The first cluster contains statements by respondents who, as a matter of definition, seem to equate personal services with services offered by a human. Or conversely, some respondents seem to classify services offered digital y as not personal by definition. Their statements do not fit any of the eight dimensions wel , but also do not contribute new dimensions on what makes a service personal. In a similar fashion, the second cluster contains statement by respondents who equate personal services with services that deal with sensitive aspects of a citizen’s life, equivalent to the classification of medical or financial information as personal information. This, however, does not attribute any properties to the way a service is offered, but more to the area or domain the service is in. These statements, too, were not coded. Finally, the last cluster contains statements that begged reflection on the completeness of the eight dimensions. The first type were statements that refer to the importance of two-way communication, including physical and non-verbal aspects of communication. An example is the respondent who stated: “when I am in a face-to-face conversation, I can see by the way they react that something is wrong”. Careful consideration of these type of statements led to them begin mapped to the language dimension (i.e. being able to interpret the sincerity of a response). The second type included statements on the types of feedback that are needed to feel 'noticed’, which is part of the definition of the empathy dimension. The final set of statements in this cluster concerned experiencing relatedness to the service provider. This is in line with the conclusion by Pol et al. (2020) that communal sharing has a positive effect on customer experience. As it addresses another level than individual services, we did not translate these into an additional dimension. 492 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 5 Discussion The eight dimensions that emerged from the interviews and focus groups seem to have an interesting relation to existing areas of academic work. From the beginnings of digital service provisions, attention has been given to accessibility aspects. Being able to interact with a service provider at one’s own time and from any location has been one of the sel ing points of digitalized services. Interface design and tone of voice of digital conversations has also received plenty of attention from the start. We recognize this in the dimensions of time, location and language. With the increasing availability of data, attention turned to the cognitive aspects of digital services, creating personal relevance based on situational data. Large amounts of data enable context-awareness in service provision, providing the ‘right’ information, at the ‘right’ time, in the ‘right’ place, in the ‘right’ way, to the 'right’ person (Fischer, 2012). Morana et al. (2017) discusses the influence of task complexity. Complexity has a bi-directional relation with the way service interactions are designed: interaction design may be aimed at reducing complexity, but the effect of design or execution choices is also influenced by the complexity of the task (Morana et al., 2017). Users benefit more from suggestive guidance if the task is simple while the benefits of informative guidance increase as the complexity of the task increases. In the focus groups, various citizens expressed varying expectations that were specific for either high or low complexity tasks. We recognize these cognitive aspects in the dimensions of complexity and individual relevance. Research in artificial intelligence is fostering a large volume of literature on the ethical impact of digitalization, with attention for personal values and beliefs. The emerging literature on social AI deals with social relations including showing empathy (Kim et al., 2021). Confidence to act is much discussed in relation to the use of algorithms (Lee & See, 2004; Gaertig & Simmons, 2018; Longoni et al., 2019; Dietvorst & Bartels, 2020). Users may have the impression that algorithms are less able to respond to their unique, individual situation than a human advisor. Longoni et al. (2019) introduce the construct Uniqueness neglect for this. Users may also have the impression that algorithms always “just” maximize a one-dimensional utility function and therefore trust algorithms less in ethically difficult domains (Dietvorst J. van Grondel e, M. van Steenbergen, A. Smits, T. Timmerman, K. van Turnhout, H. Pol: The Anatomy of a Personal Service: the Eight Dimensions of 'Personal' 493 & Bartels, 2020). In the algorithmic decision making, autonomy plays an important role, and as a result the concept of algorithmic affordances (options for the user to actively influence the outcome of the algorithm, by for instance tweaking the weight of parameters) receives more and more attention (Hekman et al., 2022, Smits et al., 2019, Dietvorst et al., 2018). Having some control over what to do with the outcomes of AI can increase adoption rates (Dietvorst et al., 2018), and can make the recommendations feel more personal (Hekman et al., 2022). The importance of maintaining control is also found in broader research into human-machine interaction. Personal values and beliefs we recognize in the dimensions of confidence to act, empathy and autonomy. This same progress from attention for accessibility to cognitive aspects to values and beliefs in literature is partly mirrored in service provider strategies and development plans found in the professional field. Time, place and language are mentioned in any omnichannel strategy. Content personalization and advice robots are becoming increasingly common. However, providers are currently discovering that other, softer, dimensions remain that determine whether citizens experience their service as being personal. A final observation that results from this study is that, interestingly enough, none of the dimensions are, by their definition, inherently bound to either digital or human services. For each dimension, both personal and non-personal experiences were mentioned, and those experiences have taken place in the context of human-to- human services and digital services. That leads to the conclusion that the eight dimensions may also have descriptive value for organizations that offer mixed portfolios of human and digital services and are striving for a more personal service experience. 6 Conclusions The main contribution of this paper is capturing and organizing the language that citizens use when they assess the personal quality of a service. Although literature on personalization of services abound, this explicit perspective was still missing. The eight dimensions (time, location, language, complexity, individual relevance, confidence to act, empathy and autonomy) that citizens use to describe their 494 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY experiences with and expectations of personal services is a first step towards a framework of personal service. A limitation of the study presented here is that it is grounded in the context of public services in The Netherlands. Similar studies will have to be conducted in other domains and other cultural situations to assess the extent to which these dimensions and, in its wake, the preliminary clustering into three categories, can be generalized. For now, both the professional practice and the academic research community seem to recognize the dimensions. They are al regularly encountered in both domains, al be it in a fragmentary manner, distributed over various disciplines, and rarely discussed from the perspective of the citizen who uses the service. Further validation is needed to determine the completeness of our list of dimensions, and the extent to which a potential resulting framework of dimensions can, in fact, be used to evaluate and improve services. Currently, this preliminary framework does provide a structure to reflect on the personal quality of services (digital and human-to-human). The main contribution of this paper, therefore, is its insights on what makes a public service personal from the perspective of the service consumer, and its operationalization of the personal quality of services in terms of eight dimensions. Acknowledgements We would like to thank al participants and interviewers of the interviews and focus groups. This research is co-financed by the Taskforce for Applied Research SIA, part of the Netherlands Organization for Scientific Research (NWO). References Barwitz, N., Maas, P. (2018). Understanding the omnichannel customer journey: determinants of interaction choice. Journal of interactive marketing, 43(1), 116-133. Chandra, S., Verma, S., Lim, W. M., Kumar, S., Donthu, N. (2022). Personalization in personalized marketing: Trends and ways forward. Psychology & Marketing, 39(8), 1529-1562. Dietvorst, B. J., Simmons, J. P., Massey, C. (2018). Overcoming algorithm aversion: People will use imperfect algorithms if they can (even slightly) modify them. Management Science, 64(3), 1155-1170. Dietvorst, B. J., Bartels, D. M. (2020). Consumers object to algorithms making morally relevant tradeoffs because of algorithms’ consequentialist decision strategies. Journal of Consumer Psychology. J. van Grondel e, M. van Steenbergen, A. Smits, T. Timmerman, K. van Turnhout, H. Pol: The Anatomy of a Personal Service: the Eight Dimensions of 'Personal' 495 Fan, H., Poole, M. S. (2006). What is personalization? Perspectives on the design and implementation of personalization in information systems. Journal of Organizational Computing and Electronic Commerce, 16(3-4), 179-202. Fischer, G. (2012, May). Context-aware systems: the 'right' information, at the 'right' time, in the 'right' place, in the 'right' way, to the 'right' person. In Proceedings of the international working conference on advanced visual interfaces (pp. 287-294). Fiske, A. P. (1991). Structures of social life: The four elementary forms of human relations: Communal sharing, authority ranking, equality matching, market pricing. Free Press. Gaertig, C., Simmons, J. P. (2018). Do people inherently dislike uncertain advice? Psychological Science, 29(4), 504-520. Hekman, E., Nguyen, D., Stalenhoef, M., Van Turnhout, K. (2022). Towards a Pattern Library for Algorithmic Affordances. Joint Proceedings of the IUI 2022 Workshops, vol. 3124, 24-33. Kim, J., Merril Jr, K., Col ins, C. (2021). AI as a friend or assistant: The mediating role of perceived usefulness in social AI vs. functional AI. Telematics and Informatics, 64, 101694. Kim, J. J., Kim, T., Wojdynski, B. W., Jun, H. (2022). Getting a little too personal? Positive and negative effects of personalized advertising on online multitaskers. Telematics and Informatics, 101831. Lee, J. D., See, K. A. (2004). Trust in automation: Designing for appropriate reliance. Human factors, 46(1), 50-80. Longoni, C., Bonezzi, A., Morewedge, C. K. (2019). Resistance to medical artificial intel igence. Journal of Consumer Research, 46(4), 629-650. McGraw, A. P., Tetlock, P. E. (2005). Taboo trade‐offs, relational framing, and the acceptability of exchanges. Journal of Consumer psychology, 15(1), 2-15. Mittelstadt, B. D., Al o, P., Taddeo, M., Wachter, S., Floridi, L. (2016). The ethics of algorithms: Mapping the debate. Big Data & Society, 3(2), 2053951716679679. Morana, S., Schacht, S., Scherp, A., Maedche, A. (2017). A review of the nature and effects of guidance design features. Decision Support Systems, 97, 31-42. Pol, H., Galetzka, M., & Pruyn, A. (2020). New perspectives on customer relationships: how relational models influence customer experience and how they are activated. Journal of relationship marketing, 19(1), 29-51. Riegger, A. S., Klein, J. F., Merfeld, K., Henkel, S. (2021). Technology-enabled personalization in retail stores: Understanding drivers and barriers. Journal of Business Research, 123, 140-155. Smits, A., Van Turnhout, K., Hekman, E., Nguyen, D.: Data-driven design. Proceedings of the 22nd International Conference on Engineering and Product Design Education. (2020). https://doi.org/10.35199/EPDE.2020.10 Sunikka, A., & Bragge, J. (2008, January). What, who and where: insights into personalization. In Proceedings of the 41st annual Hawai international conference on system sciences (HICSS 2008) (pp. 283-283). IEEE. Surprenant, C. F., Solomon, M. R. (1987). Predictability and personalization in the service encounter. Journal of marketing, 51(2), 86-96. Teeny, J. D., Siev, J. J., Briñol, P., Petty, R. E. (2021). A review and conceptual framework for understanding personalized matching effects in persuasion. Journal of Consumer Psychology, 31(2), 382-414. White, T. B., Zahay, D. L., Thorbjørnsen, H., Shavitt, S. (2008). Getting too personal: Reactance to highly personalized email solicitations. Marketing Letters, 19, 39-50. Zanker, M., Ricci, F., Jannach, D., Terveen, L. (2010). Measuring the impact of personalization and recommendation on user behaviour. International Journal of Human-Computer Studies, 68(8), 469-471. 496 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY DERIVING DECISION MINING SYSTEM CAPABILITIES: A RESEARCH AGENDA KOEN SMIT, SAM LEEWIS, MATTHIJS BERKHOUT, JOHN VAN MEERTEN, CHAIM DE GELDER, SUSAN BRUGGELING, HANNE DE DECKERE, ANNEMAE VAN DE HOEF HU University of Applied Sciences Utrecht, Digital Ethics, Ultrecht, Netherland koen.smit@hu.nl, sam.leewis@hu.nl, matthijs.berkhout@hu.nl, john.vanmeerten@hu.nl, chaim.degelder@hu.nl, susan.bruggeling@hu.nl, hanne.dedeckere@hu.nl, annemae.vandehoef@hu.nl Decision Mining (DM) is increasingly gaining attention from academia and slowly progressing towards instrumental application in practice by leveraging decision logs to automatical y discover, check for conformance and improve derivation patterns for operational decision-making. This study aims to further operationalize DM by identifying capabilities in the form of functional and non-functional requirements that are posed in the current body of knowledge. By identifying and analysing DM contributions with a focus on derivation patterns Keywords: we were able to point out the aspects of DM getting attention as decision mining, well as which did not, e.g., a strong focus on input data and capabilities, algorithms regarding the discovery phase while the output (data) systems, of the improvement phase seems to be detailed insignificantly. requirements, research Based on this we formulated a research agenda in which five key agenda, points for future research studies are presented. literature DOI https://doi.org/10.18690/um.fov.4.2023.32 ISBN 978-961-286-751-5 498 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Proper operational decision-making is one of the most important capabilities of an organization (Mircea et al., 2012). Especial y when organizations designed products and services that focus on high volumes of data processing for operational decision-making (Chalvatzis et al., 2019; Rula et al., 2016), such as governmental services focusing on calculation and application of benefits or financial services for opening accounts at a bank. Adequately managing these operational decisions is becoming increasingly difficult as the digitization of products and services as well as the transition towards ful y automated operational decision-making becomes more prevalent. One way of reducing complexity in IT is to separate concerns such as 'data', 'user interfacing', and 'processes' from each other (Dijkstra, 1974; Ossher & Tarr, 2001), which resulted in separate systems to do so as wel . A whole research field now focuses on separating decisions, sometimes also referred to as 'the logic', so that its management can be simplified and made more explicit (Bajec & Krisper, 2005; Boyer & Mili, 2011; De Smedt et al., 2017; Graham, 2006; Schlosser et al., 2014; Smit, 2018). The next step in properly managing decisions is to optimal y use data to improve decisions that are explicitly managed, which is similar to what happened in the field of business process management, referred to as Process Mining (van der Aalst, 2011). In the context of separating the concern of 'decisions' and 'logic', a similar approach is referred to as Decision Mining (DM), which is defined as: “the method of extracting and analyzing decision logs with the aim to extract information from such decision logs for the creation of business rules, to check compliance to business rules and regulations, and to present performance information” (Leewis, Smit, et al., 2020). DM comprises 1) Discovery, 2) Conformance Checking, and 3) Improvement of decisions and underlying logic (Leewis, Smit, et al., 2020). These phases are very similar to Process Mining. The goal of process mining is to discover process models, check their conformance to theoretical process models and improve process models based on the outcome of the first two phases. The present state of the research field and the knowledge base on Data Mining (DM) is currently deemed inadequate by scholars (Leewis, Smit, et al., 2020; Vanthienen, 2021). However, there is an increasing interest in DM, particularly in academic circles (Goossens et al., 2023; Scheibel & Rinderle-Ma, 2022), where there is a significant emphasis on quality metrics such as the accuracy of the mining algorithm. However, the comprehensibility of the application of DM algorithms to stakeholders like analysts or end-users is often ignored (Vanthienen, 2021). Therefore, it would be worthwhile to investigate the capabilities of a K. Smit, S. Leewis, M. Berkhout, J. van Meerten, C. de Gelder, S. Bruggeling, H. de Deckere, A. van de Hoef: Deriving Decision Mining System Capabilities: A Research Agenda 499 DM system, as this would help to advance the research field by putting into practice the theoretical y proposed stages that are currently presented in a high level of abstraction in the existing body of knowledge. Within this context, a capability is defined as " an ability that an organization, person, or system possesses" (The Open Group, 2011). In this study, a first step is made towards operationalizing capabilities for discovery, conformance checking, and improvement as part of DM. This paper comprises the exploration of capabilities derived from the current body of knowledge on DM. In a follow-up study we aim to derive capabilities using an empirical approach so that both inputs can be compared and a final set of validated capabilities can be presented so that organizations are able to more swiftly experiment with DM. Therefore, in this paper, we answer the following research question: ‘ What capabilities can be identified from DM literature focused on the decision viewpoint for the development of DM systems and how should future research into DM systems be conducted?' The remainder of the paper is structured as follows. In the fol owing section, section two, we discuss the phases of DM in more detail and explore the current state-of-the-art on DM. Then, in section three, we present the research method used in this study to derive capabilities for the DM system, which we use as a basis to formulate a research agenda. This is followed by the data collection and data analysis in section four. Then, in section five, the identified DM system capabilities are presented. The limitations of the study and its results are presented in section six. Lastly, the paper is concluded in section seven by presenting a research agenda. 2 Background DM is also referred to as decision point analysis, which “aims at the detection of data dependencies that affect the routing of a case”(Rozinat & van der Aalst, 2006). The difference between the two is that decision point analysis mines ‘sequencing patterns’ from a process viewpoint, while DM, in line with the definition presented in the previous section, mines ‘derivation patterns’ from a decision viewpoint (Leewis, Smit, et al., 2020). Previous studies indicate that the focus towards DM and mining on derivation patterns is necessary to take steps forward in DM (De Smedt et al., 2017; Leewis, Smit, et al., 2020; Sarno et al., 2013). DM consists of three phases (Leewis, Smit, et al., 2020): discovery of decisions, conformance checking of decisions, and the improvement of decisions, see Figure 1. These phases comprise extracting information from decision logs (discovery), checking this information for compliance with business rules and regulations 500 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY (conformance), and presenting possible performance information (improvement) (Leewis, Berkhout, et al., 2020). Figure 1: DM phases with corresponding inputs and outputs As depicted in Figure 1, the input data necessary to leverage DM is the ‘decision log’ (De Smedt et al., 2017; Leewis, Smit, et al., 2020). The structure of this specific log file is essential for the success of DM. De Smedt et. al. (2017) introduced the necessity of decision logs in order to properly mine decisions and Leewis et. al. (2022)described the importance of a decision log for mining decisions with a decision viewpoint in mind. Additionally, DM aims at mining decisions from structured data. Structured data has a predefined data model compared to unstructured data. Therefore, a structure needs to be defined in the form of a decision log. The structure of a decision log consists of a minimum of four attributes. The first attribute is a unique identifier, for example, a Case ID. Moreover, the condition(s) (second attribute) and conclusion(s) (third attribute) connected to that Case ID are part of the decision log, for example, al the data that is necessary for acquiring a loan and the decisions made with that data. The last required attribute is a timestamp. For example, the time certain input data was received or a decision was made by an actor. An example of a decision log is shown in Figure 2. K. Smit, S. Leewis, M. Berkhout, J. van Meerten, C. de Gelder, S. Bruggeling, H. de Deckere, A. van de Hoef: Deriving Decision Mining System Capabilities: A Research Agenda 501 Input values Output value Conditions Conclusion ID Timestamp Gender Age Temperature Treatment 1 2021-22-10 Male 30 37,4 Treatment A 2 2021-26-10 Female 25 39,1 Treatment B 5 2021-30-10 Female 50 38,7 Treatment A 7 2021-21-10 Male 19 37,2 Treatment A 8 2021-22-10 Male 18 36,8 Treatment A 10 2021-26-10 Female 24 37,1 Treatment B 11 2021-28-10 Female 20 37,6 Treatment B Figure 2: Example of a decision log with the required attributes Despite the current focus on DM with structured data, more and more research is being conducted on extracting and discovering decision logic and decision dependencies in (semi) unstructured data such as laws and regulations or protocols and guidelines (Etikala et al., 2020; Goossens et al., 2022; Vanthienen, 2021). While mining decision logic from text using text mining algorithms and remodelling them to a decision table is a complex task in itself, finding relations between the individual decision tables by mining dependencies, among the other elements on the requirements level, is even more complex and is currently being explored (Vanthienen, 2021). Recent studies also focus on the conformance-checking phase of DM where anomalies in individual decision tables, as well as decision requirements diagrams, are found (Batoulis & Weske, 2018; Corea & Delfmann, 2018; Smit et al., 2017). The techniques described in these studies deal with the completeness and consistency of the individual rules, the decision table as a whole, and the dependencies between decisions. All models can be completely checked and audited, and anomalies can be found. Additionally, they can even be (semi) automatically altered (Corea & Delfmann, 2018). We argue that the current body of knowledge, although with a limited focus on the decision viewpoint, contains useful pointers for the operationalization towards capabilities. One way to define capabilities is by formulating functionalities that a system supporting the DM phases must possess. In software engineering, requirements are used to express functionalities (Kotonya & Sommerville, 1998). Different types of requirements exist, for example, functional requirements, non-functional requirements, business requirements, user requirements, and constraints 502 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY (Sommerville & Sawyer, 1997). In this paper, we solely focus on functional and non-functional requirements as these are recognizable archetypes (Kotonya & Sommerville, 1998). Furthermore, a functional requirement emphasizes what is required and a non-functional requirement emphasizes the general properties of a system. This does not define the how, which is in line with the notion of a capability, that also focuses on what (value) an organization can deliver, but not how the value is delivered. In this study, we use thematic coding to give meaning to a wide range of possible functional requirements and are domain-specific. For non-functional requirements we use the ISO25010 standard comprising system and software quality clustering, which is a thoroughly validated framework (ISO, 2014). Examples of clusters are Reliablity (e.g., availability, fault tolerance) and operability (e.g., learnability, technical accessibility). 3 Research method In order to identify functional and non-functional requirements for the development of DM systems we start, in this study, with a thorough analysis of the scientific literature. As described earlier, the focus lies on requirements which are defined for solutions specified from a decision viewpoint. Therefore, these sources need to be identified and coded with regard to functional and non-functional requirements. Because the field of DM is in a nascent state (Leewis, Smit, et al., 2020) the body of knowledge is limited. Even more so when limiting contributions to the decision viewpoint. Therefore, the study is considered an explorative one. 4 Data Collection & Analysis The data collection of relevant contributions focussed on DM from a decision viewpoint focusing on deriving derivation patterns took place during November 2022. For this search, two criteria were adhered to exclude non-relevant contributions; 1) contributions referring to the utilization of event logs only to derive sequence data, and 2) papers referring to decision-point analysis as a form of DM. The first criterion was used because it excludes contributions that focus (predominantly) on process mining, which is not the focus of this study. The second criterion was used because decision-point analysis is considered a form of DM, which however, aims at deriving sequencing patterns in the context of business processes and business process management, i.e., how a sequence of a business K. Smit, S. Leewis, M. Berkhout, J. van Meerten, C. de Gelder, S. Bruggeling, H. de Deckere, A. van de Hoef: Deriving Decision Mining System Capabilities: A Research Agenda 503 process is routed based on one or more conditions. Google Scholar has been used as our search engine. The exclusion criteria were as fol ows: 1) the contribution focuses on decision-point analysis (thus using a process instead of a decision perspective), 2) the contribution is accessible for the research team, and 3) the paper is written in English. The search resulted in eight relevant contributions, presented in Table 1. The analysis of the identified contributions started by identifying functional and non-functional requirements through thematic coding as specified by Gibbs (Gibbs, 2007). We employed four individual coders that used a pre-defined coding scheme, based on the definitions of functional and non-functional requirements from (Sommerville & Sawyer, 1997): '' functional requirements describe what the system should do and non-functional requirements place constraints on how these functional requirements are implemented. '' Table 1: Identified contributions focusing on the decision viewpoint ID: Title: Reference: (Berkhout & Smit, 1 Utilizing Algorithms for Decision Mining Discovery 2022) Decision Mining versus Process Mining: a Comparison 2 (de Jong et al., 2021) of Mining Methods Decision Mining in a Broader Context: An Overview (De Smedt et al., 3 of the Current Landscape and Future Directions 2017) Business Rules Management and Decision Mining- 4 (Leewis et al., 2022) Filling in the Gaps Deep Learning for the Identification of Decision (Goossens et al., 5 Model ing Components from Text 2021) Putting Decision Mining into Context: A Literature (Leewis, Smit, et al., 6 Study 2020) Extracting Decision Model and Notation models from (Goossens et al., 7 text using deep learning techniques 2023) Future challenges in decision mining at governmental (Leewis, Berkhout, 8 institutions et al., 2020) 504 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Three coding rounds were conducted in order to reliably identify functional (F) and non-functional (NF) requirements from the identified contributions, depicted in Figure 3. During the first round of coding, al coders coded all contributions with regard to functional and non-functional requirements separately. An example of a functional requirement code is: '' al main functionalities: text classification, decision dependency extraction and decision logic extraction. '' An example of a non-functional requirement code is: ' The output of an algorithm must be explainable and comprehensible by Subject Matter Experts. '' Figure 3: Coding process and results per step A total of 590 functional requirements and 157 non-functional requirements were coded during coding round one. Table 2 shows the coding results of each coder, for each contribution, regarding both functional and non-functional requirements. The second round of coding is used as a consolidation round. During this round, all coders discussed the individual coding results from the first round. Consensus or disagreement was focused on keeping or removing individual codes. The second coding round resulted in 173 functional requirements (417 were identical or removed) and 34 non-functional requirements (123 were identical or removed). K. Smit, S. Leewis, M. Berkhout, J. van Meerten, C. de Gelder, S. Bruggeling, H. de Deckere, A. van de Hoef: Deriving Decision Mining System Capabilities: A Research Agenda 505 Table 2: Coding totals Coder 1: Coder 2: Coder 3: Coder 4: Article ID: F NF F NF F NF F NF 1 10 1 8 2 6 1 0 0 2 60 2 34 24 25 0 14 5 3 31 4 2 9 6 0 6 4 4 13 7 12 2 15 0 1 0 5 62 0 13 10 5 0 1 0 6 25 0 1 11 5 0 3 0 7 173 4 8 15 6 0 1 0 8 26 9 10 23 7 8 1 16 Then, a third coding round was used to further validate the results of the first two coding rounds. This was done by one of the original coders that worked on coding during the first two rounds but was accompanied by two senior researchers with more experience in the field of study. This resulted in a total of 143 functional requirements (30 were removed) and 25 non-functional requirements (9 were removed). In this last round, we used an additional coding scheme presented in Table 3. The coding scheme is based on the three DM phases on the x-axis, while the y-axis comprises three attention areas concerning DM that we chose to further explore concerning the research agenda for further operationalisation of DM. The attention areas are selected based on the fact that, for each of the DM phases on the x-axis, the algorithm needs specific 1) input data, characterized by requirements (e.g., a minimum of one condition and one conclusion). Furthermore, there are requirements for what the 2) algorithm itself should be able to do with the input data (e.g., which transformative steps have to be taken). The transformation of data then results in certain 3) output data, which is characterized by requirements as wel (e.g., which types of data need to be presented in what manner). Table 3: Round 3 functional requirements coding scheme Discovery Conformance Improvement Checking Input data - Code 1 ... ... - Code 2 ... ... - Code n ... ... Algorithm ... ... ... Output data ... ... ... 506 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 5 Results Based on the coding of our data we describe the functional requirements in this section using the three phases of DM. Additional y, we describe the non-functional requirements according to the ISO25010 categories. Due to space limitations, we summarize the results for each category. 5.1 Discovery phase Input data – 26 requirements As the second-largest category by requirements identified, many assumptions and requirements were posed in the literature about how the input data should (ideally) be structured so that it can be used for the discovery of decisions and underlying logic. Al contributions seem to focus on the need for a decision log and refer to the same composition of a decision log used to generate output, including implicit or explicit conditions, conclusions, timestamps, and dependencies. An example of a coded fragment referring to this category is: ' The DMN model extraction tool takes as input a decision description and automatical y classifies sentences into irrelevant(for the decision model), dependency or logic.' Algorithm – 51 requirements The predominant category of requirements applies to the algorithmic technique used for decision discovery. The found literature revealed specific specifications that decision discovery algorithms must satisfy. These specifications apply to both structured data, which utilizes event or decision logs, and (semi) unstructured data, such as laws and regulations or textual descriptions. While there is significant overlap in the requirements presented, minor differences are noticed. For instance, some algorithms must create a decision requirements diagram (DRD) with supporting elements prior to discovering the decision logic, while others first discover the decision logic and then create a DRD based on the decision logic. An example of a coded fragment referring to this category is: ' The construction of a decision model from text requires a sequence of steps, each with their own chal enges, regardless of whether a human or a machine is performing it: coreference resolution (where al expressions that refer to the same entity are resolved); preprocessing (preparing the data for analysis); text classification (identifying the K. Smit, S. Leewis, M. Berkhout, J. van Meerten, C. de Gelder, S. Bruggeling, H. de Deckere, A. van de Hoef: Deriving Decision Mining System Capabilities: A Research Agenda 507 relevant sentences for a modeling problem) and decision dependency and logic extraction (identifying the relevant elements needed for the construction of the model). '' Output data – 13 requirements The requirements identified in this category entail the transformation of data into various representations, guided by the construction of a decision model. The resulting output comprises a DRD, decision tables, and/or business rules. Additionally, a crucial aspect of the output data is that the algorithms should indicate the under or overfitting of data. An example of a coded fragment referring to this category is: ' The output of the Discovery phase is a business decision architecture, e.g., a DRD, as wel as decision tables, and business rules.' 5.2 Conformance Checking phase Input data – 8 requirements The requirements identified in the literature highlighted the relevant components necessary for conformance checking. DM utilizes decision logs and models, such as DMN models, during the conformance-checking phase. Therefore, the input for the conformance-checking activity consists of the log file and the discovered model, which usually emerges from the output of the discovery activity. An example of a coded fragment referring to this category is: ' Decision conformance checking has the same purpose and consists of the same input components as described for Process mining (log file and discovered model). '' Algorithm – 10 requirements Requirements in this category emphasize the diagnostic aspect of conformance checking. The algorithms or techniques used must possess the capability to identify and quantify discrepancies between the model and associated log files. Conformance checking should function to identify, locate, and detect deviations. Furthermore, by analyzing the conditions present in the decision log, the model can be utilized to detect bottlenecks and incorrect dependencies on both the DRD level as wel as decision logic level. An example of a coded fragment referring to this category is: 508 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 'Conformance checking analyzes whether the reality, as recorded in a logbook, corresponds to the model and vice versa. The aim is to measure their severity and to detect abnormalities. '' Output data – 5 requirements As the smalest category of identified requirements, the focus of the discovered requirements for output data was diagnostic in nature. For example, the diagnosis, in the form of output data, examines whether decisions are executed as intended, providing diagnostic information that highlights differences and similarities between the model and the input data. An example of a coded fragment referring to this category is: '' The output consists of diagnostic information that shows differences and similarities between the model and the input data (log files). '' 5.3 Improvement phase Input data – 10 requirements The majority of the found assumptions and requirements in this category overlap with conformance checking. The focus is on the utilization of structured data within an algorithm to enhance or improve discovered models in DM activities. A decision log is employed as input but the improvement phase also needs the inclusion of both a decision log and a decision model, which is the same as conformance checking. An example of a coded fragment referring to this category is: ' The enhancement/improvement activity needs an event or decision log and a model as input. '' Algorithm – 8 requirements As the second smalest category coded, most requirements found in the literature were about providing potential improvements to an existing model ultimately resulting in a new model. The algorithm must support the changing or extending of a decision model, based on a decision log and the theoretical model. An example of a coded fragment referring to this category is: ' enhancement/improvement activity aims at changing or extending the model.' K. Smit, S. Leewis, M. Berkhout, J. van Meerten, C. de Gelder, S. Bruggeling, H. de Deckere, A. van de Hoef: Deriving Decision Mining System Capabilities: A Research Agenda 509 Output data – 12 requirements The last category coded is output data during the improvement phase. The requirements coded in this category were mainly about a new model that is created using the input of a discovered model and the decision log. This phase must not only identify and present improvements but also outputs a new model with the improvements, which acts as a basis for revision during the improvement phase. An example of a coded fragment referring to this category is: ' In the enhancement/ improvement activity a new process/ decision model is created or an existing process or decision model is adapted.' 5.4 Non-functional requirements In terms of non-functional requirements, most codes referred to the usability (8 codes) and reliability (12 codes). Regarding usability, we see that contributions mention different stakeholder groups from different domains that should be able to work with DM systems. Therefore, the output of the algorithm should be explainable and understandable. For example, one of the coded fragments referring to this is: ' Further focus seems required to ensure a user-friendly interface where non-experts could use the capabilities of DM and thereby not confronted with algorithms where expert interpretation is needed. '' Another contribution mentioned that offering too much transparency can pose a risk to the accuracy of the DM algorithm used, which should be further explored. Regarding reliability, many generic data science constraints are identified, such as data quality (comprising the decision log), avoidance of data contamination, over and underfitting of DM algorithms, DM algorithm accuracy levels, (sample) data representativeness, and detection of outliers. An example of a coded fragment referring to this is: ' High validity can be ensured by utilizing accurate and reliable DM techniques (ensuring internal validity), and utilizing a data sample representative towards the population when decision mining (ensuring external validity). '' 6 Discussion This study and its results have some limitations that should be discussed. Firstly, this study only aimed to identify capabilities from a theoretical perspective. Although the eight sources have some empirical basis, most of them are conceptual in nature and should be further supplemented and validated using empirical observations in a 510 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY realistic context, i.e., using case studies. This also helps integrate DM processes into other IT processes as the current contributions on DM with a decision viewpoint mostly examine DM systems as a standalone phenomenon. A second limitation is the low number of contributions in the current body of knowledge with a focus on the decision viewpoint that aims to identify derivation patterns. This severely limits the theoretical richness of capabilities identified in this paper, but also points out a gap in the current body of knowledge that should be covered by future research. Another direction for future research would be to include contributions from the process mining research field that focus on the operationalisation of the three phases and look at systems for process mining specifical y, a perspective that has not been included in this study. A third limitation would be the overestimation of depth and richness we expected the current body of knowledge to have in terms of potential for operationalisation of capabilities into requirements. If we look at the results from the first coding round alone, we identified many coded fragments of concepts that are described at a high level of abstraction. This also limits the value of the (patterns of) requirements we unearthed in this study, although we argue that the results are still a useful starting point for the development of tooling to support DM systems as well as that it serves as a research agenda for what needs further consideration in future research. The high level of abstraction adhered to in the identified contributions is presumably also caused by the focus of these studies as they are not meant to operationalize capabilities to the level we pursued to do so in this study. Thus, this observation also cal s for future research focused on applying DM capabilities and their operationalization in DM systems in practice. 7 Conclusion To conclude this paper we revisit the research question posed in the introduction section: ‘ What capabilities can be identified from Decision Mining literature focused on the decision viewpoint for the development of decision mining systems and how should future research into DM systems be conducted?' Based on a thorough analysis of the current body of knowledge on DM focusing on the decision viewpoint we derived requirements on nine areas on a functional level and used the ISO 25010 software quality standard to identify a predominant focus on usability and reliability non-functional requirements for DM systems. From a theoretical perspective, our results point out directions for future research. Additional y, we demonstrated that our theoretical model of DM phases and focus areas for the operationalisation of DM system capabilities could be a K. Smit, S. Leewis, M. Berkhout, J. van Meerten, C. de Gelder, S. Bruggeling, H. de Deckere, A. van de Hoef: Deriving Decision Mining System Capabilities: A Research Agenda 511 useful approach, which can be used in future studies as suggested further in this section. From a practical perspective, our results help practitioners to 'unbox' the conceptual level of DM that the current body of knowledge on DM comprises. Doing so, DM capabilities can be further exploited in practice as proper tooling requires attention from both researchers and practitioners in the coming years, similar to the development of process mining systems in the last decade. Based on the results we formulate the fol owing research agenda for the operationalisation of DM capabilities: 1. Analysis of the current body of knowledge on DM reveals that most contributions focus on deriving sequencing patterns and are also referred to as decision-point analyses based on event data from business processes. In general, we argue that (more) research attention should be directed towards DM from a decision point of view focused on discovering, conformance checking and improvement of derivation patterns in data captured in decision logs. This also further helps in maturing the research field of DM and provides a basis for further operationalisation towards proper software systems to support businesses in leveraging the power of their operational decision-making. 2. Although the identified contributions did not primarily focus on presenting requirements for DM systems we argue that many of them describe DM capabilities from a high level of abstraction that should be further explored in the future, e.g., by describing how to process open norms in decision logs, how outliers in decision logs can be detected and managed, or what changing or extending of decision models exactly entails. Doing so helps future research in becoming more practical y applicable as the current contributions are limited in their use for practitioners. Also, future studies can focus on validating capabilities that are detailed wel enough. 3. Our findings show that most contributions focus on exploring the discovery phase of DM, while the conformance checking and improvement phases seem to be defined in less detail. Future studies should secure that both the conformance checking and improvement phases are considered and explored further. This is important to close the feedback loop of decision- making using DM capabilities as only discovery only gets organisations so far. Furthermore, we see that the identified contributions primarily focus on 512 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY detailing the input data, while less attention is focused on the algorithm and especial y the output (data) of the conformance checking and improvement phases. This observation also cal s for further exploration and definition regarding these aspects. 4. The current body of knowledge offers a narrow glimpse into non-functional requirements specifical y relevant to DM systems. From the contributions, we identified a focus on usability and reliability, but how the other software quality aspects come into play is yet to be discovered. Exploration of the non-functional requirements (also referred to as constraints by some) is important as these are contextual boundaries that should be taken into account by design as much as possible. Future studies that explore the use of DM systems in practice could benefit from these findings. 5. We argue that it is important to conceptual y ground DM phases and concepts. The identified contributions focus on theoretical y proving that DM based on deriving derivation patterns is a technique that should be further explored. However, the current body of knowledge seems to lack a strong empirical basis, which should be taken into account in future studies, e.g., involving subject-matter experts and experts from outside academia as wel as setting up case studies in which DM phases and systems are integral y evaluated and lessons learned formulated. References Bajec, M., & Krisper, M. (2005). A methodology and tool support for managing business rules in organisations. Information Systems, 30(6), 423–443. https://doi.org/10.1016/j.is.2004.05.003 Batoulis, K., & Weske, M. (2018). Disambiguation of DMN Decision Tables. In Lecture Notes in Business Information Processing (Vol. 320, Issue March, pp. 236–249). https://doi.org/10.1007/978-3-319-93931-5_17 Berkhout, M., & Smit, K. (2022). Utilizing Algorithms for Decision Mining Discovery. 35 Th Bled EConference Digital Restructuring and Human (Re)Action, 343–358. https://doi.org/10.18690/um.fov.4.2022.21 Boyer, J., & Mili, H. (2011). Agile business rule development: Process, Architecture and JRules Examples. Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-642-19041-4 Chalvatzis, K. J., Malekpoor, H., Mishra, N., Lettice, F., & Choudhary, S. (2019). Sustainable resource al ocation for power generation: The role of big data in enabling interindustry architectural innovation. Technological Forecasting and Social Change, 144(March 2018), 381–393. https://doi.org/10.1016/j.techfore.2018.04.031 Corea, C., & Delfmann, P. (2018). A tool to monitor consistent decision-making in business process execution. CEUR Workshop Proceedings, 2196(September), 76–80. K. Smit, S. Leewis, M. Berkhout, J. van Meerten, C. de Gelder, S. Bruggeling, H. de Deckere, A. van de Hoef: Deriving Decision Mining System Capabilities: A Research Agenda 513 de Jong, R., Leewis, S., & Berkhout, M. (2021). Decision Mining versus Process Mining: a Comparison of Mining Methods. 2021 5th International Conference on Software and E-Business (ICSEB), 28–32. https://doi.org/10.1145/3507485.3507490 De Smedt, J., vanden Broucke, S. K. L. M., Obregon, J., Kim, A., Jung, J.-Y., Vanthienen, J., B, J. D. S., Broucke, S. K. L. M., & Obregon, J. (2017). Decision Mining in a Broader Context: An Overview of the Current Landscape and Future Directions. In Lecture Notes in Business Information Processing (Vol. 281, pp. 197–207). Springer International Publishing. https://doi.org/10.1007/978-3-319-58457-7_15 Dijkstra, E. W. (1974). On the Role of Scientific Thought. In Selected Writings on Computing: A personal Perspective (pp. 60–66). Springer New York. https://doi.org/10.1007/978-1-4612-5695-3_12 Etikala, V., Veldhoven, Z. Van, & Vanthienen, J. (2020). Text2Dec: extracting decision dependencies from natural language text for automated DMN decision model ing. … Management Workshops: BPM …. https://doi.org/10.1007/978-3-030-66498-5_27 Gibbs, G. (2007). Thematic Coding and Categorizing. In Analyzing Qualitative Data (pp. 38–55). SAGE Publications, Ltd. https://doi.org/10.4135/9781849208574.n4 Goossens, A., Claessens, M., Parthoens, C., & Vanthienen, J. (2021). Deep Learning for the Identification of Decision Modelling Components from Text (pp. 158–171). https://doi.org/10.1007/978-3-030-91167-6_11 Goossens, A., Claessens, M., Parthoens, C., & Vanthienen, J. (2022). Extracting Decision Dependencies and Decision Logic from Text Using Deep Learning Techniques (pp. 349–361). https://doi.org/10.1007/978-3-030-94343-1_27 Goossens, A., De Smedt, J., & Vanthienen, J. (2023). Extracting Decision Model and Notation models from text using deep learning techniques. Expert Systems with Applications, 211(August 2022), 118667. https://doi.org/10.1016/j.eswa.2022.118667 Graham, I. (2006). Business rules management and service oriented architecture a pattern language (1st ed.). John Wiley & Sons. ISO. (2014). ISO/IEC 25000:2014. https://www.iso.org/standard/64764.html Kotonya, G., & Sommervil e, I. (1998). Requirements engineering: processes and techniques. Wiley Publishing. Leewis, S., Berkhout, M., & Smit, K. (2020). Future Challenges in Decision Mining at Governmental Institutions. AMCIS 2020 Proceedings., 6. https://doi.org/https://aisel.aisnet.org/amcis2020/adv_info_systems_research/adv_info_sy stems_research/6 Leewis, S., Smit, K., & Berkhout, M. (2022). Business Rules Management and Decision Mining - Filling in the Gaps. Proceedings of the 55th Hawai International Conference on System Sciences, 6229–6238. https://doi.org/10.24251/HICSS.2022.755 Leewis, S., Smit, K., & Zoet, M. (2020). Putting Decision Mining into Context: A Literature Study. In Lecture Notes in Information Systems and Organisation (Vol. 38, Issue September, pp. 31– 46). https://doi.org/10.1007/978-3-030-47355-6_3 Mircea, M., Ghilic-Micu, B., & Stoic, M. (2012). An Agile Architecture Framework that Leverages the Strengths of Business Intel igence, Decision Management and Service Orientation. In Business Intel igence - Solution for Business Development. Ossher, H., & Tarr, P. (2001). Using multidimensional separation of concerns to (re)shape evolving software. Communications of the ACM, 44(10), 43–50. https://doi.org/10.1145/383845.383856 Rozinat, A., & van der Aalst, W. M. P. (2006). Decision Mining in ProM. In S. Dustdar, J. L. Fiadeiro, & A. P. Sheth (Eds.), Business Process Management: 4th International Conference, BPM 2006, Vienna, Austria, September 5-7, 2006. Proceedings (pp. 420–425). Springer Berlin Heidelberg. https://doi.org/10.1007/11841760_33 Rula, A., Maurino, A., & Batini, C. (2016). Data and Information Quality: Dimensions, Principles and Techniques (1st ed.). Springer. https://doi.org/10.1007/978-3-319-24106-7 514 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Sarno, R., Sari, P. L. I., Ginardi, H., Sunaryono, D., & Mukhlash, I. (2013). Decision mining for multi choice workflow patterns. Proceeding - 2013 International Conference on Computer, Control, Informatics and Its Applications: “Recent Chal enges in Computer, Control and Informatics”, IC3INA 2013, 2007, 337–342. https://doi.org/10.1109/IC3INA.2013.6819197 Scheibel, B., & Rinderle-Ma, S. (2022). Decision Mining with Time Series Data Based on Automatic Feature Generation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics): Vol. 13295 LNCS (Issue c, pp. 3–18). https://doi.org/10.1007/978-3-031-07472-1_1 Schlosser, S., Baghi, E., Otto, B., & Oesterle, H. (2014). Toward a Functional Reference Model for Business Rules Management. 2014 47th Hawai International Conference on System Sciences, 3837–3846. https://doi.org/10.1109/HICSS.2014.476 Smit, K. (2018). Organization and Governance of Business Rules Management Capabilities. Open University. https://research.ou.nl/en/publications/organization-and-governance-of-business-rules-management-capabili Smit, K., Zoet, M., & Berkhout, M. (2017). Verification capabilities for business rules management in the Dutch governmental context. 2017 International Conference on Research and Innovation in Information Systems (ICRIIS), 1–6. https://doi.org/10.1109/ICRIIS.2017.8002499 Sommervil e, I., & Sawyer, P. (1997). Requirements engineering: a good practice guide. John Wil ey & Son Ltd. The Open Group. (2011). TOGAF v9.1 standard. http://pubs.opengroup.org/architecture/togaf9-doc/arch/ van der Aalst, W. M. P. (2011). Process Mining. Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-642-19345-3 Vanthienen, J. (2021). Decisions, advice and explanation: An overview and research agenda. In A Research Agenda for Knowledge Management and Analytics (pp. 149–170). https://doi.org/10.4337/9781800370623.00016 HOW TO DESIGN A PARTICIPATION COMPANION: A CONVERSATIONAL INTERFACE TO FOSTER MOTIVATION AND SUPPORT PARTICIPATION VARINIA WITTHOLZ, TIMO STROHMANN, SUSANNE ROBRA-BISSANTZ Technische Universität Braunschweig, Chair of Service Information Systems, Braunschweig, Germany v.wittholz@tu-bs.de, t.strohmann@tu-bs.de, s.robra-bissantz@tu-bs.de Smart cities are no longer based only on technologies because it is their citizens who significantly influence the development. Lack of motivation plays a major role in the area of bottom-up participation. The digital transformation is creating new opportunities to support people in their participation process and increase motivation. Through the use of artificial intelligence, systems such as virtual companions can be improved further to create a valuable relationship between the human and Keywords: the machine by incorporating interpersonal elements. A virtual conversational companion that supports people in their participation process agent, could be a solution to motivate people to participate. To provide participation, smart a basis for deriving design knowledge for a Participation city, Companion this research starts by identifying the stakeholders’ virtual needs and problems based on literature reviews and interviews. companion, human- It then follows an iterative, user centered prototype development computer and evaluation. interaction DOI https://doi.org/10.18690/um.fov.4.2023.33 ISBN 978-961-286-751-5 516 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The growth of urban population presents chal enges such as traffic congestion, waste management, resource access, and crime (Simonofski et al., 2021). Smart cities, initially based on IoT, cloud computing, and sensor networks, are considered as the answer to these chal enges (Harrison and Donnelly, 2011; Perera et al., 2014). However, recent research emphasizes that smart cities should be driven by citizens' needs and expectations (Becker et al., 2022; Hollands, 2008; Vácha et al., 2016). Information and communication technologies (ICT) are integral to smart city concepts and support participation (Becker et al., 2022; Viale Pereira et al., 2017). Currently, design knowledge on participation information systems is limited, focusing on top-down digital solutions (Becker et al., 2022). Our suggestion is a shift towards a bottom-up approach by designing a Participation Companion (PaCo) based on the knowledge of artificial intelligence (AI) and Virtual Companions (VC), that helps people to participate and motivates them to seize participation opportunities. VC work with users to achieve a common goal while fostering valuable, long-term human-machine relationships (Krämer et al., 2015; Strohmann et al., 2022). Though chatbots and virtual assistants are explored in healthcare (Ahmad et al., 2022; Müller and Reuter-Oppermann, 2022) and education (Gubareva and Lopes, 2020), there is a lack of scientifically based design knowledge for virtual companions and other AI-based technologies aiming at encouraging and supporting participation. For this reason, the following research question (RQ) is to be answered: How can a VC be designed to encourage and support good and purposeful participation? Our study adopts the design science research paradigm (Hevner et al., 2004) and, specifical y, employs the reflective approach for generating design knowledge, as proposed by Möller et al. (2020). Initially, we conduct a structured literature research and interviews to gain a comprehensive understanding of the problem space. Based on this, we initiate an iterative artifact design process. Following the reflective approach, we first instantiate an artifact as a prototype, designed as a conversational interface, and extract design principles from our design process. Final y, we perform an exploratory study to gather feedback and promote participatory design with stakeholders. The scope of our study is related to the organization "Sandkasten" at the Technische Universität (TU) Braunschweig, which aims to support students in creating a sustainable and livable campus. By examining participation dynamics V. Wit holz, T. Strohmann, S. Robra-Bissantz: How to Design a Participation Companion: A Conversational Interface to Foster Motivation and Support Participation 517 within this context, we seek to generate knowledge applicable to promoting engagement in smart cities. 2 Theoretical Background 2.1 Participation and motivation Simonefski et al. (2021) define a Smart City as “a city that provides innovative solutions, in collaboration with its citizens and with the support of technology, to solve the specific challenges of its territory” (Simonofski et al., 2021, p. 1). Therefore, citizen participation is crucial, and people require intrinsic or extrinsic motivation to engage. Intrinsical y motivated people engage in an activity because they find it interesting and inherently satisfying (Alamri et al., 2020; Di Domenico and Ryan, 2017). On the other hand, extrinsically motivated people engage in an activity to achieve a consequence e.g., a reward (Di Domenico and Ryan, 2017). Our study is grounded on two kernel theories (KT) (Kuechler and Vaishnavi, 2008), that describe motivation. KT1: The Self-Determination Theory (SDT) by Ryan and Deci (2000) is a motivation theory that identifies three basic psychological needs that influence motivation: Competence describes the need for mastery and control over the outcome of a chal enge. Autonomy means the need to overcome a chal enge on one's own power. Relatedness describes the need to feel connected (Birk et al., 2016; Ryan and Deci, 2000). KT2: The theory of self-efficacy by Bandura (1977) describes that a person's expectation of an ability influences their perception of performance and promotes motivation to perform. To perform tasks, intrinsic motivation and the perception that success has been/can be achieved are needed (Bandura, 1977). 2.2 Conversational Agents and Virtual Companions Conversational Agents (CA) are digital systems based on natural language (McTear et al., 2016), interacting with users by text (as a chatbot) or voice (as a virtual assistant) (Gnewuch et al., 2017). CA can have different characteristics in terms of the scope of the task and the intensity of the relationship between the agent and the user (Janssen et al., 2020; Strohmann et al., 2022). Virtual Companions (VC) are emotionally and socially acting virtual collaboration partners (Krämer et al., 2011; Strohmann et al., 2022) with autonomous, proactive interactions (Strohmann et al., 2022). To develop a VC supporting participation, another relevant KT is 518 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY introduced: KT3: The Computers Are Social Actors-Theory (CASA) states that people transfer social heuristics to computers by assigning social attributes to the system. Despite recognizing the system as non-human, users form attachments through the social behavior of the system. (Nass et al., 1994; Nass and Moon, 2000) 3 Methodology/Design Science Research The Design Science Research (DSR) paradigm enhances knowledge and understanding of a problem domain and potential solutions through designing and applying DSR artifacts (Hevner et al., 2004). Vom Brocke et al. (2020) assert that knowledge of a context-specific problem and knowledge about potential solutions can independently coexist. Therefore, design knowledge arises from the fact that the context and quality criteria dependent on it (problem space) as wel as the representations of possible solutions and their development process (solution space) are related to each other via the evaluation of artifacts. Design knowledge can be either in the form of theoretical y abstract knowledge (design theory), such as design principles, or in the form of instantiated artifacts (design entities) (Brocke et al., 2020). In the DSR, there are different strategic approaches to arrive at the design theoretical knowledge (Iivari, 2015; Möller et al., 2020). Möller et al. (2020) propose two possible approaches in their method for design principles development: reflective (instantiation first, then extract knowledge) and supportive (identify and synthesize knowledge, then instantiate). In our research approach, we want to use the reflective approach. In our design cycle, we focus on a very creative, participatory, and user-centered approach to instantiate a prototype in three stages and then extracting reflective design principles. Figure 1: Design Cycle (Reflective Approach) V. Wit holz, T. Strohmann, S. Robra-Bissantz: How to Design a Participation Companion: A Conversational Interface to Foster Motivation and Support Participation 519 3.1 Problem Space Understanding the problem space is crucial for generating design knowledge, comprising the four key components: needs, goals, requirements, and stakeholders (Maedche et al., 2019). In our approach we divide the process of understanding the problem space into the following three consecutive steps: (1) Identification of issues from kernel theories and literature review, (2) Deriving User Stories through stakeholder interviews and (3) Extracting possible solutions from literature review. In the first step, interviews were conducted to identify the basic problems and needs of the stakeholders. Two systematic literature reviews based on the PRISMA Statement of Moher et al. (2009) were used to frame the questions for the interview. The databases selected for the literature search were Scopus, IEEE Xplore Digital Library, and ACM Digital Library to identify literature from two important categories (1) participation and (2) motivation. Accordingly, the final search query was as follows: (participation OR e-participation OR col aboration) AND (service OR support OR experience OR benefit OR utility OR incentive OR gain OR inspiration OR encouragement OR interest). Furthermore, we aimed to gain insights into the general attitude of stakeholders towards VC. Using the same databases, we searched within the categories of (1) virtual companion and (2) service. The final search query was: (virtual assistant OR virtual companion OR companion OR virtual col aborator OR artificial intel igence OR conversational agents) AND (service OR support OR experience OR benefit OR utility OR incentive OR gain OR inspiration OR encouragement OR interest). To capture user stories on VC and participation, we conducted additional interviews using an open questionnaire based on the literature. In the interviews, we explored experiences with virtual assistants, differentiating between assistants and companions, and defined "participation" in the context of TU Braunschweig's “Sandkasten” examining experiences, motives, and restraints. Final y, we introduced the PaCo concept to explore potential motivation and support methods. 30-minutes-interviews with six students were conducted, recorded, transcribed and coded to ensure a complete evaluation and identification of their problems and needs. The age range of interviewees was 21-28 years, with 66.6% male and 33.3% female participants. All interviewees are currently studying at the TU Braunschweig. For the 520 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY evaluation of the interviews, we used the text analysis software MAXQDA1, that provides the option of assigning a code to individual text segments (coded segments). The codes were developed inductively and data-driven after reviewing the transcribed interviews, therefore the categories were created afterwards (Kuckartz and Rädiker, 2019). Final y, a literature review was conducted to identify approaches for motivating individuals in other areas, searching the same databases for the categories (1) motivation and (2) encouragement. The search query was: (develop OR increase OR attract OR motivate OR animate) AND ("intrinsic motivation" OR "intrinsic encouragement"). 3.2 Solution Space The resulting proposal features a prototype of a VC assisting students in participation activities and adapting to their needs. To ensure a user centered development and practical relevance, three students (1 female, 2 male) participated in shaping the solution space within the scope of an innovation project at the TU Braunschweig. In order to generate design knowledge, we fol owed two main steps: (1) Iterative artifact design based on the design sprint approach (Knapp et al., 2016) and (2) Explorative Evaluation with stakeholders. The first step included three substeps, inspired by the design sprint approach by Knapp et al. (2016): (a) Sketching and evaluating possible solutions within the problem space: We started by generating ideas for motivating individuals to participate by a VC. Every co-creator sketched eight individual ideas each, resulting in a total of 24 ideas. To prepare for the next substep, the ideas were then voted, to decide on the three best solutions. (b) Creating and evaluating final solutions and decide on the best: In the second substep, every co-creator created a storyboard using one of the top-three-solutions developed in substep a). The storyboards aimed to show the stakeholders' problem, the possible solutions and the desired goal in a comic-like way. In the end the three storyboards were evaluated by the co-creators and three additional students (2 female, 1 male) via dot voting, to decide on the final solution. (c) Development of a testable prototype: We started by developing a conversational path for an interaction 1 www.maxqda.com V. Wit holz, T. Strohmann, S. Robra-Bissantz: How to Design a Participation Companion: A Conversational Interface to Foster Motivation and Support Participation 521 between the PaCo and a potential user based on the final solution (see step b). Based on the conversational path we created a clickable prototype-chat using AdobeXD2. The second step involved an explorative evaluation with 26 students to collect and analyze early feedback. The user test took approximately 20 minutes, and the students were encouraged to speak their thoughts aloud while using the prototype. The thoughts were col ected in keywords, to create a relaxed scenario by not using an audio recording device and transcribing every spoken word. The age range of interviewees was 19-33 years, with 58% male and 42% female participants, with everyone currently studying at the TU Braunschweig. 4 Designing a Participation Companion 4.1 Deriving User Stories for Participation Companions In total, during the analysis of the transcribed interviews in MAXQDA we col ected and evaluated 104 passages of text, with 14 codes. The results were then classified into three main categories: (1) Participation, (2) Virtual Companion and (3) Participation Companion. In the next step, the codes were further sorted within these categories and assigned to experiences, chal enges, and value. The results represent three main reasons for the lack of motivation to participate among students: lack of interest, lack of knowledge and lack of time. First, the lack of interest was identified, which refers to both missing interest in participation and in the projects offered by the "Sandkasten". Furthermore, it was found that most of the students are unaware of the projects or lack knowledge on how to participate. Furthermore, the students cited their preference for other activities as the reason for their limited time and lack of participation. In addition to identifying the main reasons for the lack motivation to participate, we were able to develop 12 user stories (US) for the categories support, purpose, behaviour and safety from the interviews. In summary users desire the PaCo to introduce and connect them to projects, clarify participation goals and meaning, and suggest supportable activities. They expect the PaCo to understand them, display (or not display) emotions, and provide motivation. Users also prioritize establishing (or avoiding) a relationship with the PaCo and ensuring personal data protection. The fol owing literature research aimed to identify 2 www.adobe.com/products/xd 522 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY possible solutions, that can increase motivation and can be used for our next steps of developing a prototype. The identified literature is primarily from the fields of education, sports, and work. An excerpt about the main findings is given in Table 1. Table 1: Excerpt from motivating activities from literature review # Applicable motivating activities from literature review Notable KT The motivating person should be self-motivated and authentic S1 KT3 (Bejtic, 2021; Brophy, 1972; Hanke, 2019). Content and tasks should be manageable, but not too easy or S2 KT2 repetitive (Bejtic, 2021). Verbal praise and positive feedback increase motivation (Cameron S3 KT3 and Pierce, 1994; Taylor and Al a, 2019). By solving a task and/or improving social skil s, the possibility of KT1, S4 professional development arises, which promotes motivation (Chen KT2 et al., 2019). Gamification, i.e. the application of the game idea and its basic S5 mechanisms in non-game contexts can promote motivation (Aditya - et al., 2018; Birk et al., 2016; Helmefalk et al., 2020). KT1, S6 Task should have a clear goal and structure (Pange et al., 2018). KT2 Deciding on their own which task to fulfil , increases motivation S7 KT1 (Koskialho, 2017; Pange et al., 2018; Syahril et al., 2021). 4.2 Designing of the artifact The artifact design process was based on the design sprint approach (see Figure 2). We started by generating 24 ideas inspired by our findings from literature review and internet search. Afterwards, the co-creators voted for the 3 best solutions, which can be divided into 3 phases: (1) attract attention, (2) facilitate participation process and (3) stay involved. The first phase consisted of a push-notification combined with an extrinsic motivator, to attract attention. The second phase included the subdivision of the main goal into subgoals, to ease the participation process. The third phase targets the moment, when the participation is over and the competition among friends can keep people stay involved. V. Wit holz, T. Strohmann, S. Robra-Bissantz: How to Design a Participation Companion: A Conversational Interface to Foster Motivation and Support Participation 523 Figure 2: Process of the design sprint The ideas were then refined and drawn in five comic-style slides using storyboards that depicted the stakeholders' problem, the possible solution, and the desired goal. Subsequently, the storyboards were voted by the co-creators which led to the final choice: the facilitation of the participation process by subdividing the main goal into subgoals. For the final solution motivating activities from literature and internet searches were integrated into the storyboard, along with the matching task assignment based on the stakeholder interviews. To represent the assignment to a task, a fictional project was developed and divided into subtasks that were assigned to interests and time slots. By creating a conversational path, which outlines a rough flow for the interaction between the user and the PaCo, we specified the steps required for the prototype implementation. Using the Virtual Companion Canvas (Strohmann et al., 2019), we framed the PaCo's behavior and appearance as friendly and courteous, with a neutral humanlike avatar. The prototype was designed as a clickable interface in AdobeXD (see Figure 3). To simulate the assignment of the matching task, different click paths were created that lead to the correct task according to the previously selected interests. The participants had to choose from prefabricated answers as the prototype wasn’t able to provide individual messages. 524 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Figure 3: Conversation Flow of the Prototype 5 Explorative evaluation The explorative evaluation of the prototype had the goal of obtaining feedback for further developments as early as possible. While using the prototype the participants phrased their thoughts about the PaCo, which led to results outlining first impressions of the stakeholders. Participants expressed some thoughts on the language, particularly related to the design of the PaCo's text messages, that should be shorter, more structured, use emojis and hyperlinks. The appearance of the PaCo was partial y criticized, as it was perceived as strange and uncanny. But in general, the human-like appearance was perceived as positive. Regarding the capabilities, it was noted that the PaCo should communicate in a goal-oriented and open manner to ensure transparency. In the area of participation, the participants would like to see clear goals and motives, as these motivate them to participate. In addition to the collection of the participants thoughts, we developed a post-test questionnaire to assess the participants’ experience with the task assignment of the PaCo. The questionnaire covered questions about the task assignment, the purpose, the use and the sustainability. We rated all items on a five-point Likert-scale. A summary of the results of the post-test questionnaire can be found in Table 2. Looking at the descriptive data, the assignment of the task by the PaCo was general y rated as positive. Especially the low standard deviation (0,49) for the first question also indicates, that the participants agree that the task assignment facilitates the entry V. Wit holz, T. Strohmann, S. Robra-Bissantz: How to Design a Participation Companion: A Conversational Interface to Foster Motivation and Support Participation 525 into the project. Furthermore, it can be stated that the participants are aware of the goal and the benefit of the task and would recommend the PaCo to friends. Table 2: Results of the Post-test Questionnaire Category Question Mean SD Task- PaCo's assignment of a task made it easier for 4,60 0,49 assignment me to get started in the project. The structuring of the project into tasks by 4,48 0,57 PaCo makes it easier for me to get started with the project. I can identify with my assigned task. 4,28 0,78 Purpose It is clear to me what goal I am supposed to 4,56 0,70 achieve with my task. Use My activity in the project has a clear use. 4,40 0,80 Sustainability After my first experience, I would 4,16 0,83 recommend the participation opportunity to friends 6 Discussion Based on the findings of the explorative evaluation, there are some notes for further adjustments of the PaCo prototype: The study's participants highlighted the importance of clear, interactive, and engaging communication in the PaCo, ensuring that the language used by the PaCo facilitates understanding of the information being conveyed. The design of the PaCo should be non-prejudiced, customizable, and human-like to foster relatability and emphasize the importance of humanity in participation. This can enhance the user experience and creates a sense of connection between the user and the PaCo, thus encouraging participation. To foster transparency and trust, there is the need for openly communicating the benefits and capabilities of PaCo, clarifying goals and background of tasks early on, and providing precise details on time commitments. Clarifying goals, creating motives for participation, and providing value to users when compared to internet research is an essential feature. 526 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY According to the reflective approach (Möller et al., 2020) we first instantiated the prototype and then derived five reflective design principles (see Table 3). These design principles cover areas such as information transparency (DP1), motivating people to participate through matching algorithms and subtasks (DP2, DP3), implementing a systematic conversation structure (DP4), and providing valuable interactions compared to traditional internet research (DP5). Table 3: Derivation of Reflective Design Principles # Reflective Design Principle (DP) DP1 For designers and developers to design a Participation Companion (PaCo) that promotes transparency and purposefulness, clear information must be given about the goal, the time required, and the content of the participation opportunity. DP2 For designers and developers to design a PaCo that promotes autonomy, relatedness and competence, a matching algorithm can be integrated that suggests different participation opportunities based on interests, skil s, and available time. DP3 For designers and developers to design a PaCo that supports self-efficancy and autonomy, participation tasks need to be divided into sub-steps that are clearly defined and explained in order to make participation understandable and accessible. DP4 For designers and developers to design a PaCo that introduces the participation scenario to the user the conversation must be designed systematically, logically and intuitively by explaining the functions and contents of the PaCo, introducing the participation project and then asking about the user's interests and skil s. DP5 For designers and developers to design a PaCo that enhances the value and motivation for participation, goals must be articulated clearly, incentives (e.g., community, commitment, certificates) must be created and a unique value proposition compared to traditional internet research must be offered. Our study has some limitations that should be acknowledged. First the sample size of the user test was relatively small. Additionally, the participants in the study were mainly students, which may not ful y represent the diverse range of potential users of the PaCo. Second, the prototype's limited interactive capabilities, with users choosing prefabricated answers, may have affected feedback due to the lack of personalized conversations. V. Wit holz, T. Strohmann, S. Robra-Bissantz: How to Design a Participation Companion: A Conversational Interface to Foster Motivation and Support Participation 527 7 Conclusion To address the lack of design knowledge for VC supporting bottom-up participation, we identified stakeholder needs and motivational issues. Based on possible solutions from literature, a prototype, designed as a conversational interface, was developed and evaluated in an iterative, user-oriented process. In the final step five reflective DP were derived, that cover areas such as information transparency, motivating people to participate through matching algorithms and subtasks, implementing a systematic conversation structure, and providing valuable interactions compared to traditional internet research. The DP contribute to the to the body of knowledge on designing a VC supporting participation and provide guidance for designers in developing a PaCo. Although our research focuses on students, its findings on the potential of a PaCo in promoting engagement and motivation can contribute to municipal participation. More research is needed to determine how our DP can be extended to urban contexts. In addition, future research should refine the DP as wel as develop an advanced prototype, and investigate the long-term effects of the PaCo on user engagement and motivation for participation. References Aditya, Y., Kusumo, D., Nurjanah, D., 2018. Gamification for Learning Basic Algorithm. https://doi.org/10.1109/ICoICT.2018.8528723 Ahmad, R., Siemon, D., Gnewuch, U., Robra-Bissantz, S., 2022. Designing Personality-Adaptive Conversational Agents for Mental Health Care. Information Systems Frontiers 24. https://doi.org/10.1007/s10796-022-10254-9 Alamri, H., Lowel , V., Watson, W., Watson, S.L., 2020. Using personalized learning as an instructional approach to motivate learners in online higher education: Learner self-determination and intrinsic motivation. Journal of Research on Technology in Education 52, 322–352. https://doi.org/10.1080/15391523.2020.1728449 Bandura, A., 1977. Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review 84, 191–215. https://doi.org/10.1037/0033-295X.84.2.191 Becker, F., Siemon, D., Robra-Bissantz, S., 2022. Smart Participation Design: Prescriptive Knowledge for Bottom-Up Participation. Communications of the Association for Information Systems 51. Bejtic, Z., 2021. Instructor’s Personal Qualities and Meaningful Behaviors That Positively Impact Students’ Intrinsic Motivation in Studio-Based Disciplines. Journal of Higher Education Theory and Practice 21. https://doi.org/10.33423/jhetp.v21i3.4144 Birk, M.V., Atkins, C., Bowey, J.T., Mandryk, R.L., 2016. Fostering Intrinsic Motivation through Avatar Identification in Digital Games, in: Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems, CHI ’16. Association for Computing Machinery, New York, NY, USA, pp. 2982–2995. https://doi.org/10.1145/2858036.2858062 528 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Brocke, J. vom, Winter, R., University of St. Gallen, Switzerland, Hevner, A., University of South Florida, USA, Maedche, A., Karlsruhe Institute of Technology, Germany, 2020. Accumulation and Evolution of Design Knowledge in Design Science Research: A Journey Through Time and Space. JAIS 21, 520–544. https://doi.org/10.17705/1jais.00611 Brophy, J.E., 1972. The role of rewards and reinforcements in early education programs: II fostering intrinsic motivation to learn. Journal of School Psychology 10, 243–251. https://doi.org/10.1016/0022-4405(72)90059-3 Cameron, J., Pierce, W.D., 1994. Reinforcement, Reward, and Intrinsic Motivation: A Meta-Analysis. Review of Educational Research 64, 363–423. https://doi.org/10.3102/00346543064003363 Chen, P., Ma, T., Cheng, X., Liu, C., Gan, L., 2019. Participation motivations of product innovation in the crowdsourcing system, in: 2019 International Conference on Industrial Engineering and Systems Management (IESM). Presented at the 2019 International Conference on Industrial Engineering and Systems Management (IESM), pp. 1–6. https://doi.org/10.1109/IESM45758.2019.8948108 Di Domenico, S.I., Ryan, R.M., 2017. The Emerging Neuroscience of Intrinsic Motivation: A New Frontier in Self-Determination Research. Front Hum Neurosci 11, 145. https://doi.org/10.3389/fnhum.2017.00145 Gnewuch, U., Morana, S., Maedche, A., 2017. Towards Designing Cooperative and Social Conversational Agents for Customer Service. Gubareva, R., Lopes, R.P., 2020. Virtual assistants for learning: A systematic literature review. pp. 97– 103. Hanke, D., 2019. How to talk to an expatriate: The International Trade Journal: Vol 33, No 1 [WWW Document]. URL https://www.tandfonline.com/doi/abs/10.1080/08853908.2018.1524801 (accessed 11.16.22). Harrison, C., Donnel y, I.A., 2011. A Theory of Smart Cities. Proceedings of the 55th Annual Meeting of the ISSS - 2011, Hull, UK. Helmefalk, M., Marcusson, L., Sel , A., 2020. “Who cares about fireworks?” – A Study on Digital Coaching, Gamification and Exercise Motivation. Proceedings of the Annual Hawai International Conference on System Sciences. Hevner, A.R., March, S.T., Park, J., Ram, S., 2004. Design Science in Information Systems Research 32. Hollands, R., 2008. Will the Real Smart City Please Stand Up? City 12, 303–320. https://doi.org/10.1080/13604810802479126 Iivari, J., 2015. Distinguishing and contrasting two strategies for design science research. European Journal of Information Systems 24, 107–115. https://doi.org/10.1057/ejis.2013.35 Janssen, A., Passlick, J., Rodríguez Cardona, D., Breitner, M.H., 2020. Virtual Assistance in Any Context. https://doi.org/10.1007/s12599-020-00644-1 Knapp, J., Zeratsky, J., Kowitz, B., 2016. Sprint: How to Solve Big Problems and Test New Ideas in Just Five Days, Illustrated Edition. ed. Simon & Schuster, New York. Koskialho, J., 2017. A Manager’s Means to Motivate Experts at Work. pp. 1047–1054. https://doi.org/10.1007/978-3-319-42070-7_95 Krämer, N., Eimler, S., Rosenthal-von der Pütten, A.M., Payr, S., 2011. Theory of Companions: What Can Theoretical Models Contribute to Applications and Understanding of Human-Robot Interaction? Applied Artificial Intelligence 25, 474–502. https://doi.org/10.1080/08839514.2011.587153 Krämer, N.C., Rosenthal‐von der Pütten, A.M., Hoffmann, L., 2015. Social Effects of Virtual and Robot Companions, in: Sundar, S.S. (Ed.), The Handbook of the Psychology of Communication Technology. Wiley, pp. 137–159. https://doi.org/10.1002/9781118426456.ch6 Kuechler, B., Vaishnavi, V., 2008. On theory development in design science research: anatomy of a research project. European Journal of Information Systems 17, 489–504. https://doi.org/10.1057/ejis.2008.40 V. Wit holz, T. Strohmann, S. Robra-Bissantz: How to Design a Participation Companion: A Conversational Interface to Foster Motivation and Support Participation 529 Maedche, A., Gregor, S., Morana, S., Feine, J., 2019. Conceptualization of the Problem Space in Design Science Research, in: Tulu, B., Djamasbi, S., Leroy, G. (Eds.), Extending the Boundaries of Design Science Theory and Practice, Lecture Notes in Computer Science. Springer International Publishing, Cham, pp. 18–31. https://doi.org/10.1007/978-3-030-19504-5_2 McTear, M., Cal ejas, Z., Griol, D., 2016. Introducing the Conversational Interface, in: McTear, M., Cal ejas, Z., Griol, D. (Eds.), The Conversational Interface: Talking to Smart Devices. Springer International Publishing, Cham, pp. 1–7. https://doi.org/10.1007/978-3-319-32967-3_1 Möl er, F., Guggenberger, T.M., Otto, B., 2020. Towards a Method for Design Principle Development in Information Systems, in: Hofmann, S., Mül er, O., Rossi, M. (Eds.), Designing for Digital Transformation. Co-Creating Services with Citizens and Industry, Lecture Notes in Computer Science. Springer International Publishing, Cham, pp. 208–220. https://doi.org/10.1007/9783-030-64823-7_20 Mül er, H., Reuter-Oppermann, M., 2022. Chatblood - Towards designing chatbots for blood donors. Nass, C., Moon, Y., 2000. Machines and Mindlessness: Social Responses to Computers. Journal of Social Issues 56, 81–103. https://doi.org/10.1111/0022-4537.00153 Nass, C., Steuer, J., Siminoff, E., 1994. Computer are social actors. Presented at the Conference on Human Factors in Computing Systems - Proceedings, p. 204. https://doi.org/10.1145/259963.260288 Pange, J., Lekka, A., Katsigianni, S., 2018. Serious Games and Motivation. pp. 240–246. https://doi.org/10.1007/978-3-319-75175-7_25 Perera, C., Zaslavsky, A., Christen, P., Georgakopoulos, D., 2014. Sensing as a Service Model for Smart Cities Supported by Internet of Things. European Transactions on Telecommunications. https://doi.org/10.1002/ett.2704 Ryan, R.M., Deci, E.L., 2000. Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist 55, 68–78. https://doi.org/10.1037/0003-066X.55.1.68 Simonofski, A., Val é, T., Serral, E., Wautelet, Y., 2021. Investigating context factors in citizen participation strategies: A comparative analysis of Swedish and Belgian smart cities. International Journal of Information Management 56, 102011. https://doi.org/10.1016/j.ijinfomgt.2019.09.007 Strohmann, T., Siemon, D., Khosrawi-Rad, B., Robra-Bissantz, S., 2022. Toward a design theory for virtual companionship. Human–Computer Interaction 1–41. https://doi.org/10.1080/07370024.2022.2084620 Strohmann, T., Siemon, D., Robra-Bissantz, S., 2019. Introducing the Virtual Companion Canvas – Towards Designing Collaborative Agents. https://doi.org/10.5445/IR/1000095222 Syahril, S., Nabawi, R., Safitri, D., 2021. Students’ perceptions of the project based on the potential of their region: A Project-based learning implementation. Journal of Technology and Science Education 11, 295. https://doi.org/10.3926/jotse.1153 Taylor, A.K., Al a, S., 2019. Influence of Reward Systems on motivation: pros and cons based on current literature 8. Vácha, T., Přibyl, O., Lom, M., Bacúrová, M., 2016. Involving citizens in smart city projects: Systems engineering meets participation, in: 2016 Smart Cities Symposium Prague (SCSP). Presented at the 2016 Smart Cities Symposium Prague (SCSP), pp. 1–6. https://doi.org/10.1109/SCSP.2016.7501027 Viale Pereira, G., Cunha, M.A., Lampoltshammer, T.J., Parycek, P., Testa, M.G., 2017. Increasing col aboration and participation in smart city governance: a cross-case analysis of smart city initiatives. Information Technology for Development 23, 526–553. https://doi.org/10.1080/02681102.2017.1353946 530 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY COMPARATIVE ANALYSIS OF MARKET STRUCTURES OF P2P ENERGY TRADING IN A LOCAL ENERGY SYSTEM MENGQIU DENG,1 XIAO PENG,2 YANG ZHAO1 1 Zhejiang University, School of Energy Engineering, Hangzhou, China mengqiu_deng@zju.edu.cn, youngzhao@zju.edu.cn 2 Utrecht University of Applied Sciences, Organizations in Digital Transition, Utrecht, Netherland xiao.peng@hu.nl Peer-to-peer (P2P) energy trading has been recognized as an important technology to increase the local self-consumption of photovoltaics in the local energy system. Different auction mechanisms and bidding strategies haven been investigated in previous studies. However, there has been no comparatively analysis on how different market structures influence the local energy system’s overal performance. This paper presents and Keywords: compares two market structures, namely a centralized market P2P energy and a decentralized market. Two pricing mechanisms in the trading, centralized market and two bidding strategies in the decentralized local market are developed. The results show that the centralized energy system, market leads to higher overal system self-consumption and market profits. In the decentralized market, some electricity is directly structure, sold to the grid due to unmatchable bids and asks. Bidding bidding strategies, strategies based on the learning algorithm can achieve better Bled performance compared to the random method. eConference DOI https://doi.org/10.18690/um.fov.4.2023.34 ISBN 978-961-286-751-5 532 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction In order to reach carbon neutrality, the local energy system is undergoing a huge transformation (Salvia et al., 2021). Distributed generation resources, particularly photovoltaics (PVs) are becoming more and more popular at the demand side (B. Zhou et al., 2021). However, the large-scale penetration of PVs has a significant impact on the safe operation of the power system (Kumar et al., 2021). Therefore, to promote the self-consumption of PV within the local energy system is the focus of current research. Peer-to-peer (P2P) energy trading has been proposed as a critical technology to increase the PV consumption in recent years (Y. Zhou et al., 2020). P2P energy trading enables direct energy trading between prosumers and consumers within the local energy system (Zhang et al., 2018). A prosumer is defined as an entity that can produce and consume electricity, such as residential households with PVs (Iazzolino et al., 2022). In the P2P energy trading, prosumers can obtain additional benefits from selling their electricity to individual consumers (Zheng, 2022). Furthermore, P2P energy trading can facilitate power balance for the power system (Soto et al., 2021). The research on the P2P energy trading can be categorized into two main streams according to the market structure: the centralized market and the decentralized market (Muhsen et al., 2022). In the centralized market, the coordinator collects information on electricity production and consumption of al prosumers and consumers. After the trading, the coordinator al ocates the payoffs of the whole system to the participants according to a predefined rule. Some rules distribute costs or profits according to each participant’s contribution to the aggregate system net consumption or surplus generation (Reis et al., 2020). Some rules calculate the local market price based on the pricing mechanism, such as supply and demand ratio (SDR) (Liu et al., 2017), mid-market rate (MMR) (Long et al., 2017) and bill sharing (BS) (Y. Zhou et al., 2018). In this market, market participants are considered as price takers and they can only accept the price made by the coordinator. In the decentralized market, prosumers and consumers are able to make autonomous decisions about the amount and price of electricity to bid. These bids are submitted to a P2P trading platform and then cleared by a certain clearing approach. Different auction mechanisms, such as Discriminatory k-Double Auction (k-DA), Uniform k- M. Deng, X. Peng, Y. Zhao: How to Design a Participation Companion: Comparative Analysis of Market Structures of p2p Energy Trading in a Local Energy System 533 DA, Vickrey-Clark-Groves (VCG), and Trade Reduction (TR), have been proposed and compared (Lin et al., 2019). Several bidding strategies have also been introduced to investigate their influence on the conditions of the market (Yu et al., 2018). Although current studies have investigated auction mechanisms and bidding strategies, there has been no comparatively analysis on how different market structures influence the local energy system’s overal performance. This paper presents and compares two market structures, namely a centralized market and a decentralized market, with the aim of providing valuable insights into establishing a P2P energy market. Firstly, two pricing mechanisms in the centralized market and two bidding strategies in the decentralized market are developed. Secondly, a comprehensive assessment of the local energy system's overal performance including costs, profits and self-consumption, is analyzed. 2 Methodolodgy 2.1 Simulation model We simulate a local energy market with the P2P energy trading in a local energy system. The market participants include N residential consumers or prosumers ( I = 1, 2, 3, . ., N), and a coordinator. A coordinator plays a different role in different market structures. In a centralized market, they typical y act as the system operator responsible for managing the system's operations. In a decentralized market, they are general y the trading platforms where electricity order matching takes place. The simulation are conducted for the day-ahead at the time interval of 1 hour ( Δt = 1 h). The internal market price in the local market should lie between the electricity feedin price and the retail electricity price. Therefore, consumers and prosumers can benefit from participating in the energy market. This can increase the local PV consumption and reduce the amount of electricity sold from the local energy system to the higher-level power grid. 2.2 Centralized market In the centralized market, the information about the requested electricity from consumers and the available surplus PV generation from prosumers is transmitted to the coordinator. The coordinator calculates the total electricity demand and 534 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY electricity surplus of the local energy system. Then the coordinator trades with the grid to balance local supply and demand. After the trading, the coordinator decides how to distribute the system profits according to a predefined rule. This paper compares two rules: the costs and profits distribution rule, and the internal pricing mechanism. 2.2.1 Costs and profits distribution rule The costs and profits distribution rule represents a fair mechanism to directly distribute the system costs and profits. Different rules have been proposed to achieve fairness in the distribution. Al ocation based on the amount of each participant’s electricity consumption and electricity injected is the most basic rule. The costs and profits of participants are calculated by Eq. (1).  PE ⋅ f if NC ≥ in  i t , i t 0 , , Costs and profits distribution = (1)  local grid  SE ⋅ f + SE ⋅ f if NC  i t market i t out , i t <0 , , , NC is the net consumption of the paticipant i at time t [kWh]. If i, t NC ≥ , the i t 0 , paticipant i is a buyer; otherwise the paticipant i is a seller. f is the retail electricity in price from the grid [$/kWh]. f is the internal market price [$/kWh]. market f is the out electricity feed-in price [$/kWh]. PE is the purchased electricity of the paticipant i i, t at time t [kWh]. local SE is the locally sold electricity of the paticipant i at time t [kWh]. i, t grid SE is the sold electricity into the grid of the paticipant i at time t [kWh]. If i, t ∑ NC < , the system sel s the surplus electricity to the grid, and the sold electricity i t 0 , is distributed among sel ers according to the proportion of thier contribution to the system's net comsuption; otherwise, no surplus electricity is sold to the grid. 2.2.2 Internal pricing mechanism As the costs and profits distribution rule are unable to reflect real-time electricity prices in the current trading market, various internal pricing mechanisms are proposed. Supply and demand ratio (SDR), mid-market rate (MMR) and bill sharing (BS) are three typical pricing mechanisms. Taking the SDR mechanism as an example, an internal pricing model for energy sharing is established, where the M. Deng, X. Peng, Y. Zhao: How to Design a Participation Companion: Comparative Analysis of Market Structures of p2p Energy Trading in a Local Energy System 535 internal price is defined as a segmented function of the energy supply-demand ratio within the market. Specifically, the supply-demand relationship in the P2P energy market can be represented by the total surplus and demand electricity at each time interval. The supply-demand ratio of electricity is defined by Eq. (2). ∑ SEi, t i N ∈ S SDR = (2) t ∑ PEi, t i N ∈ B SDR is the supply-demand ratio at time t. SE is the sold electricity of the paticipant t i, t i at time t [kWh]. PE is the purchased electricity of the paticipant i at time t [kWh]. i, t N and N are the col ection of sel ers and buyers. S B Typical y, there is an inverse proportion relationship between the price and the supply-demand ratio. Therefore, the sel ing and purchasing prices within the market are calculated by Eq. (3) and Eq. (4), respectively.  f ⋅ f out in , 0 ≤ SDR <  t 1 p = ( f − f ⋅ SDR + f (3) sell t in out ) , t out  f SDR >  out , t 1  p ⋅ SDR + f − SDR ≤ SDR < sell t t in (1 t ), 0 t 1 , p = (4) buy, t  f SDR >  out , t 1 p and p is the sel ing and purchasing prices within the market at time t sell, t buy, t [$/kWh]. 2.3 Decentralized market In the decentralized market, P2P energy trading allows consumers and prosumers to directly buy and sel PV resources using a blockchain-based platform. Consumers and prosumers can submit their own electricity bidding to the trading platform, and the platform settles orders through a specific clearing algorithm. In this paper, the periodic double auction market that leads to a single clearing price for every trading 536 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY period is implemented (Zade et al., 2022). During the trading period, participants submit bids or asks according to their roles of buyers or sel ers. Once al bids and asks have been received, they are collected in an order book. At clearing time, bids are sorted in descending order and asks are sorted in ascending order by price. All bids where the buy price exceeds or equals the ask price are matched. At the end, the clearing prices are calculated based on the uniform price calculation method. Different bidding strategies have been proposed to test the feasibility of implementing a P2P trading market. This paper compares two bidding strategies: random bidding strategies and learning algorithm based bidding strategies. 2.3.1 Random bidding strategies During the trading period, each consumer or prosumer bids with a random buy or sel price without any strategic foresight. This means that it does not take into account historical or retail electricity costs on the market. Participants randomly bid in a certain price interval. Prosumers are unwil ing to accept a price below the electricity feed-in price, while consumers are unwil ing to pay a price above the retail electricity price. Therefore, the upper price limit is set to the retail electricity price, and the lower price limit is set to the electricity feed-in price. Bid and ask prices are randomly sampled from a uniform distribution between f and f . out in 2.3.2 Learning algorithm based bidding strategies In a real-world setting, consumers and prosumers are capable of learning from past decision-making experiences, which reflects their intelligent characteristics. Based on the income they earn as prosumers and the costs that incur to them consumers, they adjust their propensities to place specific orders. Different reinforcement learning algorithms have been proposed to simulate the learning ability of participants, such as Roth-Erev (RE) algorithm (Nicolaisen et al., 2001) and Q-learning algorithm (Chiu et al., 2022). This paper takes the RE algorithm as an example to illustrate the performance of reinforcement learning algorithms (Mengelkamp et al., 2017). The basic idea of the algorithm is to give priority to previous successful decisions and to learn from recent experiences. M. Deng, X. Peng, Y. Zhao: How to Design a Participation Companion: Comparative Analysis of Market Structures of p2p Energy Trading in a Local Energy System 537 Firstly, participants determine their own set of bidding strategies. In this paper, the market price range is discretized into an integer number of bid strategies according to its upper and lower bounds. Therefore, the set S = { f , .. , f } represents all out in possible bidding strategies of the participants. In the beginning, the participants have the same initial propensities for each strategy. Secondly, as the P2P energy trading clears, participants update their own propensities for each strategy, as shown in Eq. (5).  R s ⋅ −ε j = s t ( t ) (1 ), t pr λ  = − ⋅ + (5) + pr  ε i j t (1 ) , , 1 i, j, t pr ⋅ otherwise i j t , , ,  S −1  pr represents the propensity of the participant i for each strategy j at time t. The i, j, t parameter λ ∈[0,1] represents the participant's memory factor. The higher the value of λ is, the faster the participant forgets the past decision results. The parameter ε ∈[0,1] represents the participant's learning speed. As the value of ε decreases, the importance of the previous action in future decisions increases for the participant. R s is the achieved income or the saved costs of the participant when its chosen t ( t ) strategy is s at time t. For prosumers and consumers, R s is calculated by Eq. (6) t ( t ) and Eq. (7). R s = SE ⋅ f + SE ⋅ f (6) t ( t ) local grid prosumer, i, t market i, t out R ( s ) = PE ⋅( f − f ） (7) , , in consumer t t i t market Finally, the probabilities with which the participant i chooses strategy j are then derived from these propensities at time t +1, as shown in Eq. (8). The roulette method is employed to select the final bid strategy. pri, j, t 1 prob + = (8) i, j, t 1 + S ∑ pri, j, t 1 j 1 + = 538 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3 Case study description This study targets a hypothetical local energy system of 100 residential homes. In order to reflect the diversity of the houses, the size of the house is randomly sampled from 1,000 to 4,000 square feet. Figure 1 presents the hourly base load and PV generation profiles of a 2,546 square feet house during a summer month, which are obtained from (Lin et al., 2019). The load and PV generation profile of each house in the system is determined by scaling the base load and PV generation profile proportional y to the previously generated house size. The retail electricity price is $0.123/kWh and the electricity feed-in price is $0.033/kWh. For the system, two PV penetration levels are tested (40% and 60%) with two market structures (the centralized market with the distribution rule and SDR pricing mechanism, and the decentralized market with random and learning algorithm based bidding strategies). 7 6 ) 5 (kW put 4 3 PV out 2 Load/ 1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time (h) Base load profile Base PV generation profile Figure 1: A 24h base load and PV generation profile 4 Simulation results 4.1 Case I: 40% PV penetration The 24h load and PV generation profiles of the local energy system with 60 consumers and 40 prosumers are illustrated in Figure 2. It should be noted that the load profile for the 60 consumers is superimposed on the aggregate load profile for the 40 prosumers. By combining these two profiles, the total load of the system can be determined. As a result of the 40% ratio of prosumers to consumers, the total prosumer load is lower than the total consumer load. Between approximately 8:30 M. Deng, X. Peng, Y. Zhao: How to Design a Participation Companion: Comparative Analysis of Market Structures of p2p Energy Trading in a Local Energy System 539 am and 16:30 pm, prosumers in the system generate surplus PV energy that can be exchanged with their neighbors. 600 500 ) (kW 400 put 300 PV out 200 Load/ 100 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time (h) Prosumer demand PV generation Total demand Figure 2: System supply vs demand at 40% PV penetration Some key performance evaluation indicators regarding the overall performance of the system under different market structures are analyzed in Table 2. The centralized market achieved 100% local PV self-consumption, while in the decentralized market, some prosumers have to sel electricity directly to the grid due to the failure of some bidding orders. This leads to lower self-consumption in the decentralized market. In both pricing rules of the centralized market, the same total system profits are achieved. However, the benefits obtained by different consumers and consumers vary. Consumers earned revenue by sel ing electricity to consumers and the grid, while consumers saved costs through the P2P energy trading. Under the distribution rule, consumers could not benefit from energy trading. The SDR mechanism appears to be a fair mechanism that can benefit both consumers and consumers. In the decentralized market, bid strategies based on the learning algorithm achieves higher system self-consumption and total profits. However, the cost savings obtained by consumers actual y decrease. As shown in Figure 3, this is due to the higher clearing prices resulting from the learning algorithm. The increase of the average percentage traded through the learning algorithm is not significant compared to the random algorithm. This is because the key to achieving higher self-consumption lies in the bidding of a smal number of prosumers. 540 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 1: The system performance of different market structures at 40% PV penetration Self- Profits ($) Average consumpti Consum Prosume Percentag on Total er r e traded Centralized – Distribution 100% 0 95.29 95.29 / rule Centralized – SDR 100% 56.46 38.83 95.29 / mechanism Decentralized – Random 64.79% 16.72 45.02 61.74 46% bidding Decentralized – Learning 92.32% 10.06 77.91 87.97 48% algorithm From Figure 3, it can be seen that in the centralized market, the internal selling price and buying price determined by the SDR mechanism are highly correlated with the power supply-demand ratio. The greater the supply-demand ratio is, the lower the internal price is. In the decentralized market, the clearing price obtained through the learning algorithm gradual y increases because it is a sel er's market and the electricity cannot ful y meet the needs of al consumers. 0.14 0.12 0.1 h)w 0.08 0.06 Price ($/K 0.04 0.02 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time (h) SDR - Selling price SDR - Buying price Random - Clearing price Learning - Clearing price Figure 3: Internal market price of different market structures at 40% PV penetration M. Deng, X. Peng, Y. Zhao: How to Design a Participation Companion: Comparative Analysis of Market Structures of p2p Energy Trading in a Local Energy System 541 4.2 Case II: 60% PV penetration Figure 4 displays the load and PV generation profiles for the system at 60% PV penetration. Even after meeting the electricity demand of the system, there is an excess of PV output from the system between approximately 9:30 am and 14:30 pm due to a higher penetration of PV. 600 500 ) (kW 400 put 300 PV out 200 Load/ 100 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time (h) Prosumer demand PV generation Total demand Figure 4: System supply vs demand at 60% PV penetration According to Table 2, the centralized market can only achieve an ideal 75.43% local energy consumption. Higher total system profits are achieved at a 60% PV penetration rate compared to a 40% PV penetration rate. The conclusions drawn from Table 2 are similar to those in Table 1, where the centralized market leads to higher self-consumption and total profits compared to the decentralized market. In the decentralized market, bid strategies based on the learning algorithm can achieve better performance than the random method. In Figure 6, due to the surplus of electricity caused by PV generation being greater than consumers demand at noon, the internal electricity price in the centralized market equals the electricity feed-in price. Similarly, in the decentralized market, the clearing price resulting from bid strategies based on the learning algorithm first decreases and then increases, which is also a response to the surplus of electricity resources. 542 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 2: The system performance of different market structures at 60% PV penetration Self- Profits ($) Average consumpti Consum Prosume Percentag on Total er r e traded Centralized – Distribution 75.43% 0 118.68 118.68 / rule Centralized – SDR 75.43% 41.77 76.91 118.68 / mechanism Decentralized – Random 42.31% 23.91 59.71 83.62 46% bidding Decentralized – Learning 43.64% 24.54 73.96 98.50 42% algorithm 0.12 0.1 h) 0.08 w 0.06 Price ($/K 0.04 0.02 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time (h) SDR - Selling price SDR - Buying price Random - Clearing price Learning - Clearing price Figure 5: Internal market price of different market structures at 60% PV penetration 5 Conclusions P2P energy trading plays a significant role in increasing the self-consumption of the local energy system. It can also involve consumers and prosumers in the local energy market. In different market structures, namely centralized and decentralized markets, P2P energy trading provides great economic advantages to market participants to encourage their involvement. The centralized market-based approach seems to have greater advantages as it leads to higher overall system self-consumption and profits. However, participants in the centralized market are considered as price takers and may not be ful y incentivized to participate the market actively. Otherwise, M. Deng, X. Peng, Y. Zhao: How to Design a Participation Companion: Comparative Analysis of Market Structures of p2p Energy Trading in a Local Energy System 543 participants can make decisions and bid in the decentralized market. Bid strategies based on the learning algorithm in the decentralized market show better performance compared to the random method, but this relies on participants learning from their bidding history. Overall, this study provide insights for evaluating the impact of different P2P energy trading market structures on the performance of the local energy system. Acknowledgements The authors gratefuly acknowledge the support of the National Natural Science Foundation of China (No. 51978601 and No. 52161135202). References Chiu, W.-Y., Hu, C.-W., & Chiu, K.-Y. (2022). Renewable Energy Bidding Strategies Using Multiagent Q-Learning in Double-Sided Auctions. IEEE Systems Journal, 16(1), 985–996. Iazzolino, G., Sorrentino, N., Menniti, D., Pinnarelli, A., De Carolis, M., & Mendicino, L. (2022). Energy communities and key features emerged from business models review. Energy Policy, 165, 112929. Kumar, D. S., Quan, H., Wen, K. Y., & Srinivasan, D. (2021). Probabilistic risk and severity analysis of power systems with high penetration of photovoltaics. Solar Energy, 230, 1156–1164. Lin, J., Pipattanasomporn, M., & Rahman, S. (2019). Comparative analysis of auction mechanisms and bidding strategies for P2P solar transactive energy markets. Applied Energy, 255, 113687. Liu, N., Yu, X., Wang, C., Li, C., Ma, L., & Lei, J. (2017). Energy-Sharing Model With Price-Based Demand Response for Microgrids of Peer-to-Peer Prosumers. IEEE Transactions on Power Systems, 32(5), 3569–3583. Long, C., Wu, J., Zhang, C., Thomas, L., Cheng, M., & Jenkins, N. (2017). Peer-to-peer energy trading in a community microgrid. 2017 IEEE Power & Energy Society General Meeting, 1–5. Mengelkamp, E., Staudt, P., Garttner, J., & Weinhardt, C. (2017). Trading on local energy markets: A comparison of market designs and bidding strategies. 2017 14th International Conference on the European Energy Market (EEM), 1–6. Muhsen, H., Al ahham, A., Al-Halhouli, A., Al-Mahmodi, M., Alkhraibat, A., & Hamdan, M. (2022). Business Model of Peer-to-Peer Energy Trading: A Review of Literature. Sustainability, 14(3), 1616. Nicolaisen, J., Petrov, V., & Tesfatsion, L. (2001). Market power and efficiency in a computational electricity market with discriminatory double-auction pricing. IEEE Transactions on Evolutionary Computation, 5(5), 504–523. Reis, I. F. G., Gonçalves, I., Lopes, M. A. R., & Antunes, C. H. (2020). A multi-agent system approach to exploit demand-side flexibility in an energy community. Utilities Policy, 67, 101114. Salvia, M., Reckien, D., Pietrapertosa, F., Eckersley, P., Spyridaki, N.-A., Krook-Riekkola, A., Olazabal, M., De Gregorio Hurtado, S., Simoes, S. G., Geneletti, D., Viguié, V., Fokaides, P. A., Ioannou, B. I., Flamos, A., Csete, M. S., Buzasi, A., Orru, H., de Boer, C., Foley, A., … Heidrich, O. (2021). Will climate mitigation ambitions lead to carbon neutrality? An analysis of the local-level plans of 327 cities in the EU. Renewable and Sustainable Energy Reviews, 135, 110253. Soto, E. A., Bosman, L. B., Wol ega, E., & Leon-Salas, W. D. (2021). Peer-to-peer energy trading: A review of the literature. Applied Energy, 283, 116268. 544 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Yu, Q., Meeuw, A., & Wortmann, F. (2018). Design and implementation of a blockchain multi-energy system. Energy Informatics, 1(S1), 17. Zade, M., Lumpp, S. D., Tzscheutschler, P., & Wagner, U. (2022). Satisfying user preferences in community-based local energy markets—Auction-based clearing approaches. Applied Energy, 306, 118004. Zhang, C., Wu, J., Zhou, Y., Cheng, M., & Long, C. (2018). Peer-to-Peer energy trading in a Microgrid. Applied Energy, 220, 1–12. Zheng, B. (2022). A peer-to-peer energy trading market embedded with residential shared energy storage units. Applied Energy. Zhou, B., Meng, Y., Huang, W., Wang, H., Deng, L., Huang, S., & Wei, J. (2021). Multi-energy net load forecasting for integrated local energy systems with heterogeneous prosumers. International Journal of Electrical Power & Energy Systems, 126, 106542. Zhou, Y., Wu, J., & Long, C. (2018). Evaluation of peer-to-peer energy sharing mechanisms based on a multiagent simulation framework. Applied Energy, 222, 993–1022. Zhou, Y., Wu, J., Song, G., & Long, C. (2020). Framework design and optimal bidding strategy for ancil ary service provision from a peer-to-peer energy trading community. Applied Energy, 278, 115671. TOWARDS PRINCIPLES FOR A DATA-DRIVEN BUSINESS MODEL INNOVATION PROCESS – A DESIGN SCIENCE CASE STUDY MICHAEL FRUHWIRTH,1,3 VIKTORIA PAMMER-SCHINDLER2 1 Know-Center GmbH, Graz, Austria michael.fruhwirth@silicon-austria.com 2 Graz University of Technology, Graz, Austria & Know-Center GmbH, Graz, Austria vpammer@know-center.at 3 Silicon Austria Labs GmbH, Graz, Austria Transforming an existing business model into a data-driven one is chal enging. Tools, methods and processes can support organisations in that innovation. This paper presents a three-year interventionist case study with an automotive company, where we investigated how an innovation process towards data-driven Keywords: business models should be designed. We analysed data from business model interviews, notes from company meetings and workshops, as innovation well as learnings from supporting seven different data initiatives process, interventionist within the organisation. As a result, we present requirements that case decision-makers have regarding a process and principles that study, guide the process design. The principles are not specific to data- data-driven business driven business model innovation. However, at the level of models, operationalising the process, activities and actionable tools need design to be specific to the goal of a business model innovation: how science research, data and analytics can be used for new services and business Bled models. eConference DOI https://doi.org/10.18690/um.fov.4.2023.35 ISBN 978-961-286-751-5 546 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Developments in data-driven technologies, as well as the availability of large data sets, hold the opportunity for developing new products, services, and business models (Günther et al. 2017), so-called data-driven business models (DDBMs) (Hartmann et al. 2016). This transformation toward a DDBM is particularly chal enging for offline-established organisations (Schüritz et al. 2017b), i.e., organisations with an established business model that does not (yet) substantial y rely on data analytics-enabled services or products. Therefore, research has started to design tools and methods as support for developing DDBMs (Fruhwirth et al. 2020b), e.g., supporting idea generation (Kühne and Böhmann 2019), performing financial evaluations (Zolnowski et al. 2017) or identifying risks (Fruhwirth et al. 2021). While these approaches investigate specific aspects of DDBM innovation, such as idea generation, evaluation, or risk management, organisations also need support over the innovation activities via a structured management process (Terrenghi 2019). Further, the knowledge of such a holistic process is still fragmented, specifical y missing a sequence of activities and connection of tools (Fruhwirth et al. 2020b). Moreover, there is a lack of knowledge in designing such processes. Research has recently started to develop frameworks to guide the development of a DDBM (e.g., Rashed et al. 2022). Nevertheless, they need to be adapted to the organisational requirements, connected to innovation tools and converted into a structured process. Therefore, we answer the fol owing research question: What process allows us to develop data-driven business models in of line-established organisations systematical y? 2 Background Business models can be understood as “stories that explain how enterprises work” (Magretta 2002) and describe how organisations create, deliver and capture value (Osterwalder and Pigneur 2010). Business model innovation (BMI) appears “when two or more elements of a business model are reinvented to deliver value in a new way” (Lindgardt et al. 2009). BMI can be seen as a process, i.e., “the activity of designing – that is, creating, implementing and validating – a new BM [business model]” (Massa and Tucci 2013). Processes serve as a guideline to structure BMI activities in organisations (Wirtz and Daiser 2018). A process comprises idealised phases, such as idea generation or implementation (Wirtz 2011). Each phase is M. Fruhwirth, V. Pammer-Schindler: Towards Principles For a Data-Driven Business Model Innovation Process – A Design Science Case Study 547 associated with certain activities and generates distinct outcomes (Terrenghi 2019), and tools and methods support these activities (Bouwman et al. 2020). As research often tends to focus on parts of processes, e.g., providing single tools, the knowledge about how phases, activities, and tools are connected is stil fragmented (Fruhwirth et al. 2020b). Therefore, research on business model innovation processes increased recently (Andreini et al. 2022), such as a process model to align value creation and value capture in BMI. Nevertheless, little attention has been paid to how to desig such processes. Concretely, we know of only Geissdoerfer (2019) and Simmert et al. (2019), who designed a process for sustainable business model innovation and continuous business model improvement, respectively. Data-driven business models (DDBMs), in particular, have a conceptual focus on value creation from data (Guggenberger et al. 2020). In a DDBM, data is used as a key resource (Hartmann et al. 2016). Data analytics methods are applied to discover insights from data (Kühne and Böhmann 2019) that are delivered as data-based features, products, or services and support customers in decision-making (Schüritz et al. 2019) and enable the generation of new revenue streams (Schüritz et al. 2017a). Existing literature provides a comprehensive set of typologies of DDBMs (Dehnert et al. 2021), often based on the business models of start-ups (e.g., Hartmann et al. 2016; Schmidt et al. 2018), thus, neglecting offline-established organisations (Fruhwirth et al. 2020b). Further, academia has paid little attention to the dynamic aspects of DDBMs (Wiener et al. 2020), particularly their design and realisation (Rashed and Drews 2021). One exception is the study of Lange et al. (2021), who found that DDBMs are realised iteratively along four periods (experimentation, minimum viable product, minimum marketable product and scaling). Although tools and methods should support organisations along that process (Fruhwirth et al. 2020b), current research mainly focuses on supporting idea generation through canvases (e.g., Hunke et al. 2021; Kayser et al. 2019; Kühne and Böhmann 2019). Further, there is a need for repeatable processes and the connection of tools and methods (Fruhwirth et al. 2020b). Existing high-level process approaches are based on expert interviews (Hunke et al. 2017) or literature reviews (Lange and Drews 2020). Rashed et al. (2022) recently provided a reference framework with six enablers and related activities that guide the design and realisation of DDBMs. Such models guide 548 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY activities but are not yet embedded in a manageable process and do not connect tools to an overarching procedure (Fruhwirth et al. 2020b). 3 Research Design Our overarching research approach is an interventionist case study (Korhonen et al. 2021; Yin 2009) with one automotive company (masked due to confidentiality as Comp) fol owing principles of design science research. Comp is one of the world’s leading organisations in engineering and testing of automotive systems, operating in a B2B context, with more than 10.000 employees. Comp has a knowledge-intensive business, where innovations are often triggered bottom-up. The automotive industry is undergoing a significant transformation due to data-driven technologies like autonomous driving that offer opportunities for new revenue streams with DDBMs (Seiberth and Gründinger 2018). Thus, the question of Comp is how to evolve its profitable business model by leveraging new technologies such as big data analytics and artificial intelligence. We conducted this case study over three years, from 2018 to 2021. In this case study, we developed individual tools, methods, and an overall process to support Comp’s DDBM initiatives. Note that the scope of the presented paper is not on the individual stages, activities, or tools but on the overal process and structured support during a DDBM innovation. Further, the scope of this research is on DDBMs on a unit or service level of Comp that are proposed in addition to their existing business models. This research can be label ed as an interventionist case study since we actively col aborated with representatives of Comp and were involved in different stages of DDBM innovation initiatives. This approach allowed us to access meaningful research data (Korhonen et al. 2021). Aside from 28 semi-structured interviews, this study includes 97 documented meetings and workshops with 73 representatives of Comp. Further, one researcher actively participated in seven DDBM initiatives at Comp. As design outcomes, we derived design requirements, design principles and design features for such a process. Design requirements describe what users need and expect from a process. To identify requirements, we conducted 17 interviews with employees responsible for data-driven innovations and 11 interviews with employees responsible for BMI at Comp. Further, we col ected tacit knowledge about BMI practices and considerations M. Fruhwirth, V. Pammer-Schindler: Towards Principles For a Data-Driven Business Model Innovation Process – A Design Science Case Study 549 for DDBMs by participating in 97 meetings and workshops over three years. We took notes and had access to additional internal materials (e.g., presentations). We analysed our data following a Qualitative Content Analysis (Mayring 2015): we applied an open coding approach to identify relevant statements, grouped similar statements to codes, and structured the codes to requirements. Design principles capture the knowledge from the design process and describe salient characteristics of the design that are transferable to other solutions for the same problem (i.e., other business model innovation processes) (Sein et al. 2011). Design principles also show how the requirements link to the specific implementation, i.e., the design features (Meth et al. 2015). We extracted our design principles through reflection and abstraction from our design requirements and features (Gregor et al. 2013). Design features address specific aspects of the requirements (Maedche et al. 2021) and structure the description of the process design. We crafted design features by addressing the requirements and synthesising best practices at Comp. and grounded them in the BMI literature. Further, we conducted a structured literature review (Fruhwirth et al. 2020b), leading to an initial toolbox. One researcher actively participated in DDBM innovations at Comp, where we developed DDBM-specific tools. Supporting seven DDBM initiatives in specific activities enabled us to generate learnings on the activity and tool level. 4 Design Requirements DR1: A DDBM innovation process should increase the speed of innovation, i.e., the time to market from an idea to launching a DDBM. Increasing speed is especial y important for DDBMs, as they move faster with shorter life cycles compared to traditional product-oriented businesses. One manager at Comp mentioned: “Time to market will be quite important with data. We wil only be successful […] if we are real y fast in development.” (Manager Data Science). In contrast to traditional products with extensive release and approval processes, DDBMs must go to market with a semi-finished solution – a “minimum marketable product” (Lange et al. 2021). Customer (decision) problems that can be addressed by a data service emerge over time through customer interactions and insights from data analytics. 550 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY DR2: A DDBM innovation process should guide management (investment) decisions. A successful DDBM innovation requires commitment from management to provide sufficient resources (Rashed et al. 2022). As innovating a DDBM is associated with many uncertainties, resources must be al ocated reasonably. Therefore, criteria are needed to inform and objectify decisions, as one manager highlighted: “It [the process] must support decision-making, it must provide orientation and clear yes/no decisions, provide clear statements.” (Product Manager Data Solutions). One important aspect of decision-making is to identify risks (Tesch and Brillinger 2017), e.g., if critical information could be shared through a data-based value proposition (Fruhwirth et al. 2021). DR3: A DDBM innovation process should have an iterative character and fol ow an ef ectuation logic to address the uncertainties in innovating a DDBM . Effectuation focuses on taking action in the market to generate new insights by a trial and error logic (Sosna et al. 2010; Tesch et al. 2017). One common approach for early customer feedback is a Minimum Viable Product (MVP), as one manager reported: In digital innovations, you have to create an MVP and go into testing at about 50 per cent maturity. Before that, the substance for validation was missing. In the data-driven environment, MVP approaches are much more prominent.” (Manager Digital Services). DR4: A DDBM innovation process should be simple and adaptive. It should focus on the minimal necessary elements that also a prerequisite that the process wil be used by all target users, as one manager highlighted: “It must be simple to make a new topic understandable for a department. Everyone should have the know-how to use the process correctly.” (Project Manager). DDBM innovation requires adaptive approaches (Lange et al. 2021) in contrast to traditional structured processes, e.g., for product development. DR5: A DDBM innovation process should educate its users and establish a mindset (e.g., customer orientation and data thinking). Thus, a process should also provide guidance and how-to instructions and equip its users with the competencies to innovate DDBMs, as one manager exemplary mentioned: “A process can be supportive if you plan to establish the thinking that is inherent in the process anyway. Keyword: process plus education.” (Project Manager) One example of such educative topics was fostering customer-centricity, which is critical in DDBMs, as value is closely co-created with the customer (Schüritz et al. 2019). M. Fruhwirth, V. Pammer-Schindler: Towards Principles For a Data-Driven Business Model Innovation Process – A Design Science Case Study 551 DR6: A DDBM innovation process should provide actionable how-to instructions for its users. We found that the how-to of certain activities, such as defining a data product and its potential benefits, is often unclear for non-data domain experts. For instance, one manager mentioned that “a process should provide a clear roadmap from grasping first ideas up to calculating an ROI with checklists, best practices, examples, and suggestions for business model tools.” (Manager Software). 5 Design Principles and Features We describe our process design, as shown in Figure 1, guided by our three design principles: structure the process by investment decisions, support cyclic convergent and divergent thinking and enable organisational learning. We implemented the design principles via seven design features. Initiation and Ideation Analytical and Prototyping and Implementation and Feasibility Validation Execution Gate: Commitment to provide business Gate: Decision to test business Gate: Decision to implement resources for further elaborating BM idea model sketch with a PoC prototype business model in a (sub-) market Generating business model ideas. Rollout the business model in at least a s Describe the business model idea and Analytically analyze and evaluate the Testing the business model idea sub-market and scale the business show the potential and relevance for business case. This stage requires (assumptions, hypothesis) through model. oal investing time and resources to further resources for the business model team successful MVP prototypes and Controlling of the business model and G elaborate the idea. and supplementary functions customer interactions. ensuring its sustainability. s  Check technical feasibility  Implement and rollout the business  Decide in what direction you want  Create a business model design  develop a minimum viable product model to go with a data-driven innovation itie  Identify and test hypothesis in your  Create a financial model  Continuously monitor the success  Analyze your current situation (data business model design  Perform a risk evaluation of the business model sources, customer pains, ..)  Evaluate your business model  Test your MVP with customer  Continuously adopt the business  Describe your business model idea Activ analytically interactions model to a changing environment  Business Model Canvas  Business Model Canvas  BM Implementation Plan /  Value Proposition Canvas  Customer surveys / MVP field tests  SWOT, PESTEL Analysis Roadmap  Data Map  Business case calculation  Focus groups, customer or expert  Metrics (e.g., customer  Data Product Canvas Interviews  Business Model Risk Evaluation Tools satisfaction, market share,  Data Service Cards  Scenario Planning / Technique  Market and competitor analysis ROI, margin, …) e om Idea Description Business Model Design & Evaluated customer demand MVP Implemented business model utcO - Fit with strategic roadmap of - Possession of core competencies - Commitment of stakeholders and company (strategic fit) and dynamic capabilities key partners - Strategic importance Technical gap, technical complexity - Benchmark with competitors Required data sources - Alignment with market trends - Idea evaluation criteria - Successful prototyping and - Novelty of problem - Potential customer demand customer interaction, - Results from riteria - Novelty of solution customer surveys or field tests C Information Supporting AVL strategy, required - Rough cost and revenue - Viable Business Case Calculation from tools competencies estimation (ROI, Financial Plan) inform the - Technological Complexity - Technological Effort decision - Technical proof of concept Figure 1: Overview of our process design with instantiated design features (DFs). 5.1 DP1: Structuring the Process by Decision Points A process should be structured along with management investment decisions that are informed by a set of decision criteria. These criteria determine the information that needs to be collected in each phase to inform the decisions. Based on these 552 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY criteria, the precedent phase can be defined with its activities and supporting tools. Recommended tools guide and support the data collection process. The following features implement this principle. DF1.1 - Definition of phases and gates: The DDBM innovation process is structured along four phases and intermediate gates: initiation and ideation, analytical feasibility, prototyping and validation, implementation and execution. We based these phases on the review work of Wirtz and Daiser (2018), who identified seven generic phases of a business model innovation process based on a systematic literature review. We merged phases between two gates and dropped the decision- making phase due to our gate structure. DF1.2 - Support decision-making by actionable criteria: Decision criteria inform and objectify the management decision if they further invest in a business model initiative (Tesch et al. 2017). We identified decision criteria from six categories based on our case and the literature: customer demand, market and competition, organisation and strategy, data and technology, financial rationales, and risks. The criteria are operationalised via evaluation questions and a response scale for each criterion in a closed form (in terms of a binary “yes” or “no”, or in the form of Likert items) (Gilsing et al. 2020). Data-specific examples for such criteria are data ownership or risks associated with data sharing. DF1.3 – Define an outcome for each phase: From our case study, we learned that it is important to have a clearly defined outcome, documented in a coherent form, at the end of each phase. This is, in particular, important when a portfolio of DDBM innovations has to be managed. For instance, the goal for the idea generation phase is to have a ful y elaborated idea with a description of the key elements of a DDBM. We used this requirement to develop the Data Product Canvas (Fruhwirth et al. 2020a) as a template. The main elements are a description of the customer, benefits, problems addressed, a vision for the data analytics solution, required data sources and data analytics methods. M. Fruhwirth, V. Pammer-Schindler: Towards Principles For a Data-Driven Business Model Innovation Process – A Design Science Case Study 553 5.2 DP2: Support Cyclic Divergent and Convergent Thinking Every phase in BMI has alternating activities that require divergent (i.e., exploring multiple options) and convergent thinking (i.e., deciding and going for one option). These two types of thinking and related activities are iterated until a target outcome is achieved. For instance, in the idea generation phase, activities encompass generating multiple DDBM ideas (divergent thinking) and filtering and deciding on one promising opportunity (convergent thinking). We implemented this design principle via the fol owing two features. DF2.1 – Definition of iterative activities for each phase: The process suggests activities for each phase that lead to the defined outcomes. The activities are iterated and alternated until the target outcome is achieved (e.g., identifying and validating a meaningful customer need). In our interventionist case study with Comp, we ran through an iterative cycle for testing the hypothesis in a DDBM, as suggested, for instance, by Bland et al. (2020) and added it to the process. This involved alternating activities of identifying potential data sources, generating insights from the data via data analytics and exploring customer needs. DF2.2 – Suggestions from a toolbox: Each activity of a phase is supported by suggestions from a toolbox. In workshops of our case study, we combined several tools from the literature to support idea generation. First, we used a classification matrix (e.g., Breitfuß et al. 2019) to guide the direction of the ideation workshops (i.e., what type of DDBM should be investigated). A Data Map (e.g., Kayser et al. 2019) then supports identifying, structuring, and documenting data sources as input for idea generation workshops. A card deck (Breitfuß et al. 2023) provides basic information on DDBMs for non-data experts and supports the creative process in idea-generation workshops. Final y, the Data Product Canvas (Fruhwirth et al. 2020a) is used to structure idea generation workshops and describe and communicate a DDBM idea. 5.3 DP3: Enable Organisational Learning A process should enable organisational learning by providing best practices. A process is based on both generic knowledge from the literature (e.g., phases of BMI) and organisation-specific best practices and tacit knowledge. Further, a process is a 554 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY vehicle of change to establish desired procedures and ways of thinking. Organisations and their employees can learn by using and continuously updating the process. We implemented this principle based on two features. DF3.1 – Include best practices: Our process design for DDBM incorporates best practices and learnings from previous DDBM innovations (i.e., critical aspects to consider or how to execute an activity). For instance, for evaluating a DDBM with the help of a SWOT analysis, we added guiding questions that were asked in previous DDBM innovations, such as: Are we too dependent on certain external data sources/providers? What happens if we have no access to the data any more? Are we handling critical customer data where a data breach would have serious consequences (e.g., threatening our reputation)? DF3.2 – Provide a method of use for tools: For each tool, we defined a goal, an explanation, a template, and suggestions for a combination with other tools. Further we provided an illustrative example of how the tool could be used or was used in a previous initiative. As we found that it was often unclear how a tool should be used, we added step-by-step descriptions of how to use each tool. For instance, for the Data Product Canvas (Fruhwirth et al. 2020a), we found it useful to start with the customer problems, then think about a vision for the data service, consider required data sources and analytics activities, and then iterate. 6 Discussion In this study, we investigated a DDBM innovation process through a case study from two perspectives: First, what are the requirements and expectations of the management regarding a process? And second, what are the inherent principles that guide the design of such a process? We found three salient principles that map to the characteristics of our case study. Structuring a process along gates and investment decision points reflects the hierarchical control structures and the need for investment steering of traditional B2B organisations (Rummel et al. 2022). Support cyclic convergent and divergent thinking reflect the agile and iterative nature of digital innovations (Ghezzi and Cavallo 2020). Enabling organisational learning reflects the need for a knowledge-intensive business where innovations often happen decentralised and bottom-up (Burnes et al. 2003). Thus, our principles are not specific to DDBMs; they can be transferred to other types of business models M. Fruhwirth, V. Pammer-Schindler: Towards Principles For a Data-Driven Business Model Innovation Process – A Design Science Case Study 555 in similar contextual settings. Nevertheless, on the activities and tools level, the process is very specific for DDBMs, as the tools and activities bring in the necessary knowledge and specifics of DDBMs. We showed that a process design should be structured along with (investment) decision points and informing criteria. This principle relates to a linear BMI process approach (Andreini et al. 2022) and reflects traditional organisations’ management steering and hierarchical control structure (Rummel et al. 2022). However, specific decision points are missing in current BMI processes. Tesch et al. (2017) empirically investigated decision points and decision criteria, Lange et al. (2021) further add that incumbent organisations use Stage-Gates for a stop-or-go decision during DDBM innovation. Further, a process design should support cyclic convergent and divergent thinking within each phase. This principle relates to the recursive BMI process approach (Andreini et al. 2022) and reflects the iterative and agile nature of digital innovations in general (Ghezzi and Cavallo 2020; Rummel et al. 2022). This principle relates to topical approaches such as Design Thinking and Lean Start-Up for BMI (Brown 2008; Ries 2011; Rummel et al. 2022). Existing process designs from the literature do not explicitly differentiate between convergent and divergent thinking. However, Hunke et al. (2017) already visualise convergent and divergent thinking aspects in their process. Fruhwirth et al. (2020b) also suggest structuring BMI tools by convergent and divergent thinking. Nevertheless, there is a tension between these two design principles, i.e., the required iterative and flexible character (within each phase) with iterative divergent and convergent activities and the strict Stage-Gate logic (at the gates between the phases). Cooper and Sommer (2016) found that IT and manufacturing firms recently combined agile development and Stage-Gate approaches for product development, the so-called Agile-Stage-Gate hybrid model. Rummel et al. (2022) further found that manufacturing firms with a B2B business model use hybrid agile and Stage-Gate models for their BMI process. Thus, these empirical studies underpin the relevance of this two design principles. 556 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY A BMI process should also enable organisational learning. This principle reflects that innovations in knowledge-intensive organisations (such as Comp) often happen bottom-up and that knowledge about new business models is emerging over time in organisations. In our case study, we observed that DDBM innovations happen bottom-up in the units based on customer interactions, as domain experts are closer to the customer problem that can be addressed with data analytics. Therefore, generated learnings and insights about DDBMs need to be transferred to the organisational system. The domain experts also need the skil s and tools to develop DDBMs successfully. Thus, by incorporating best practices and learning, a process can be the vehicle of knowledge transfer from individuals to the organisation and vice-versa (Sosna et al. 2010). The need for a high level of learning is crucial in fast-moving environments based on digital technologies (Burnes et al. 2003), such as data analytics. Thus, this literature stream underpins the relevance of our third design principle. On a tools and activity level, our process is very specific to DDBMs. Tools and methods support the activities within each phase and bring in the knowledge and specifics for DDBMs. By suggesting DDBM-specific activities, tools and methods – and showing how they are interlinked - organisations and individuals can learn about the characteristics of this new type of business model. Recent literature has investigated DDBM-specific activities during BMI (Lange and Drews 2020; Rashed and Drews 2021). Nevertheless, there is a need for further research in several areas of tool support for DDBMs. First, further research should identify DDBM-specific decision criteria and evaluate DDBM-specific risks. Second, as research recently started to empirical y investigate the realisation of DDBMs (i.e., prototyping and implementation; e.g., Lange et al. 2021; Rashed and Drews 2021), these insights should be transferred to tools and processes. Finally, our research is not without limitations. All our design outcomes are based on a single case. While we aimed to generalise our results through design principles and reflected in the discussion how these appear in other, similar processes, future research should build on the principles and reflect on their usefulness in helping design a DDBM innovation process. Second, we did not rigorously evaluate the process. Future research could conduct interviews to evaluate our design principles. It should also investigate how comparative and experimental research complement case study work. This could show how our process improves the performance and M. Fruhwirth, V. Pammer-Schindler: Towards Principles For a Data-Driven Business Model Innovation Process – A Design Science Case Study 557 outcome of DDBM innovations in organisations. It could further measure the effectiveness of the process in terms of velocity (i.e., time from the first idea to the execution of the DDBM) and economic impact (i.e., the success rate of innovations). 7 Conclusion This paper provides three contributions to the literature: First , we showed that the design of a BMI process could be viewed from two perspectives: what the users expect from a process (requirements) and how to design such a process (principles and features). Our principles align with recent literature and point to other disciplines, such as psychology (with convergent and divergent thinking) and (organisational) learning. Second, we showed that a BMI process can be operationalised by defining outcomes, activities and tools for each phase, as shown in Figure 1. Third, our results show that the activity and tool level bring the specifics of DDBMs to a BMI process. Further, we provide an integrated perspective on how different tools and methods are interlinked. Acknowledgements This work was supported by the COMET — Competence Centers for Excel ent Technologies Program, funded by the Austrian Federal Ministry (BMK and BMDW), the Austrian Research Promotion Agency (FFG), the province of Styria (SFG) and partners from industry and academia. The COMET Program is managed by FFG. In addition, we thank Maximilian Ferstl for conducting part of the interviews. References Andreini, D., Bettinelli, C., Foss, N. J., and Mismetti, M. 2022. “Business model innovation: a review of the process-based literature,” Journal of Management and Governance (26:4), pp. 1089-1121 (doi: 10.1007/s10997-021-09590-w). Bland, D. J., Osterwalder, A., Smith, A., and Papadakos, T. 2020. Testing business ideas: Strategyzer.com/test, Hoboken, New Jersey: John Wiley & Sons, Inc. Bouwman, H., Reuver, M. de, Heikkilä, M., and Fielt, E. 2020. “Business model tooling: where research and practice meet,” Electronic Markets (30:3), pp. 413-419 (doi: 10.1007/s12525-020-00424-5). Breitfuß, G., Fruhwirth, M., Pammer-Schindler, V., Stern, H., and Dennerlein, S. 2019. “The DataDriven Business Value Matrix: A Classification Scheme for Data-Driven Business Models,” in 32nd Bled eConference Humanizing Technology for a Sustainable Society, A. Pucihar (ed.), pp. 803-820. Breitfuß, G., Santa-Maria, T., Fruhwirth, M., and Disch, L. 2023. “Use Your Data: Design and Evaluation of a Card-Based Ideation Tool for Data-Driven Services,” Proceedings of the 56th Hawai International Conference on System Sciences, pp. 5176-5185. Brown, T. 2008. “Design Thinking,” Harvard Business Review (56:4). 558 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Burnes, B., Cooper, C., and West, P. 2003. “Organisational learning: the new management paradigm?” Management Decision (41:5), pp. 452-464 (doi: 10.1108/00251740310479304]). Cooper, R. G., and Sommer, A. F. 2016. “Agile-Stage-Gate: New idea-to-launch method for manufactured new products is faster, more responsive,” Industrial Marketing Management (59), pp. 167-180 (doi: 10.1016/j.indmarman.2016.10.006). Dehnert, M., Gleiss, A., and Frederik, R. 2021. “What makes a data-driven business model? A consolidated taxonomy,” Proceedings of the Twenty-Ninth European Conference on Information Systems (ECIS 2021). Fruhwirth, M., Breitfuß, G., and Pammer-Schindler, V. 2020a. “The Data Product Canvas: A Visual Col aborative Tool for Designing Data-Driven Business Models,” in 33rd Bled eConference Enabling Technology for a Sustainable Society, A. Pucihar, M. K. Borštnar, R. Bons, H. Cripps, A. Sheombar and D. Vidmar (eds.), Online. June 28-29 2020, pp. 515-528. Fruhwirth, M., Pammer-Schindler, V., and Thalmann, S. 2021. “A Network-based Tool for Identifying Knowledge Risks in Data-Driven Business Models,” in Proceedings of the 54th Annual Hawai International Conference on System Sciences, T. Bui (ed.), pp. 5218-5227. Fruhwirth, M., Ropposch, C., and Pammer-Schindler, V. 2020b. “Supporting Data-Driven Business Model Innovations: A structured literature review on tools and methods,” Journal of Business Models (8:1), 7-25. Geissdoerfer, M. 2019. Sustainable business model innovation: Process, chal enges and implementation. Ghezzi, A., and Cavallo, A. 2020. “Agile Business Model Innovation in Digital Entrepreneurship: Lean Startup Approaches,” Journal of Business Research (110), pp. 519-537 (doi: 10.1016/j.jbusres.2018.06.013). Gilsing, R., Turetken, O., Ozkan, Baris, Adali, Onat Ege, and Grefen, P. 2020. “A method for qualitative evaluation of service-dominant business models,” Proceedings of the Twenty-Eight European Conference on Information Systems (ECIS) - A Virtual AIS Conference. Guggenberger, M. T., Möl er, F., Boualouch, K., and Otto, B. 2020. “Towards a Unifying Understanding of Digital Business Models,” Twenty-Third Pacific Asia Conference on Information Systems, Dubai, UAE, 2020. Günther, W. A., Rezazade Mehrizi, M. H., Huysman, M., and Feldberg, F. 2017. “Debating big data: A literature review on realizing value from big data,” The Journal of Strategic Information Systems (26:3), pp. 191-209 (doi: 10.1016/j.jsis.2017.07.003). Hartmann, P. M., Zaki, M., Feldmann, N., and Neely, A. 2016. “Capturing value from big data – a taxonomy of data-driven business models used by start-up firms,” International Journal of Operations & Production Management (36:10), pp. 1382-1406 (doi: 10.1108/IJOPM-02-2014-0098). Hunke, F., Seebacher, S., Schüritz, R., and Il i, A. 2017. “Towards a Process Model for Data-Driven Business Model Innovation,” Proceedings of the 19th IEEE Conference on Business Informatics (CBI), pp. 150-157. Hunke, F., Thomsen, H., and Satzger, G. 2021. “Investigating Modular Reuse as an Underlying Mechanism of Conceptualization during Service Design – the Case of Key Activity Orchestration,” in Proceedings of the 54th Annual Hawai International Conference on System Sciences, T. Bui (ed.). Kayser, L., Muel er, R. M., and Kronsbein, T. 2019. “Data Col ection Map: A Canvas for Shared Data Awareness in Data-Driven Innovation Projects,” 2019 Pre-ICIS SIGDSA Symposium on Inspiring mindset for Innovation with Business Analytics and Data Science, Munich 2019. Korhonen, T., Selos, E., Laine, T., and Suomala, P. 2021. “Exploring the programmability of management accounting work for increasing automation: an interventionist case study,” Accounting, Auditing & Accountability Journal (34:2), pp. 253-280 (doi: 10.1108/AAAJ-12-2016-2809). Kühne, B., and Böhmann, T. 2019. “Data-Driven Business Models: Building the Bridge Between Data and Value,” in Proceedings of the 27th European Conference on Information Systems - Information Systems for a Sharing Society, J. Vom Brocke, S. Gregor and O. Mül er (eds.), Stockholm and Uppsala, Sweden. M. Fruhwirth, V. Pammer-Schindler: Towards Principles For a Data-Driven Business Model Innovation Process – A Design Science Case Study 559 Lange, H. E., and Drews, P. 2020. “From Ideation to Realization: Essential Steps and Activities for Realizing Data-Driven Business Models,” 22nd IEEE International Conference on Business Informatics (CBI) Workshops, Antwerp, 2020. Lange, H. E., Drews, P., and Höft, M. 2021. “Realization of Data-Driven Business Models in Incumbent Companies: An Exploratory Study Based on the Resource-Based View,” Proceedings of the Forty-Second International Conference on Information Systems, Austin 2021. Lindgardt, Z., Reeves, M., Stalk, G., and Deimler, M. S. 2009. “Business Model Innovation: When the Game Gets Tough, Change the Game,” Boston Consulting Group. Maedche, A., Gregor, S., and Parsons, J. 2021. “Mapping Design Contributions in Information Systems Research: The Design Research Activity Framework.,” Communications of the Association for Information Systems ((forthcoming), in Press). Magretta, J. 2002. “Why Business Models Matter,” Harvard Business Review (80:5), pp. 86-92. Massa, L., and Tucci, C. L. 2013. “Business model innovation,” in The Oxford handbook of innovation management, M. Dodgson, D. M. Gann and N. Phil ips (eds.), Oxford: Oxford Univ. Press, pp. 420-441. Mayring, P. 2015. Qualitative Inhaltsanalyse: Grundlagen und Techniken, Weinheim: Beltz. Meth, H., Muel er, B., and Maedche, A. 2015. “Designing a Requirement Mining System,” Journal of the Association for Information Systems (16:9), pp. 799-837 (doi: 10.17705/1jais.00408). Osterwalder, A., and Pigneur, Y. 2010. Business Model Generation: A Handbook for Visionaries, Game Changers, and Chal engers, Hoboken, New Jersey: John Wiley & Sons. Rashed, F., and Drews, P. 2021. “Pathways of Data-driven Business Model Design and Realization: A Qualitative Research Study,” in Proceedings of the 54th Annual Hawai International Conference on System Sciences, T. Bui (ed.), 5676-5685 (doi: 10.24251/HICSS.2021.689). Rashed, F., Drews, P., and Zaki, M. 2022. “A Reference Model for Data-Driven Business Model Innovation Initiatives in Incumbent Firms,” Proceedings of the Thirtieth European Conference on Information Systems (ECIS 2022), Timișoara, Romania. Ries. 2011. The lean startup: How today’s entrepreneurs use continuous innovation to create radical y successful business, New York: Currency. Rummel, F., Hüsig, S., and Steinhauser, S. 2022. “Two archetypes of business model innovation processes for manufacturing firms in the context of digital transformation,” R&D Management (52:4), pp. 685-703 (doi: 10.1111/radm.12514). Schmidt, J., Drews, P., and Schirmer, I. 2018. “Charting the Emerging Financial Services Ecosystem of Fintechs and Banks: Six Types of Data-Driven Business Models in the Fintech Sector,” Proceedings of the 51st Hawaii International Conference on System Sciences, pp. 5004-5013. Schüritz, R., Farrel , K., Wixom, B. H., and Satzger, G. 2019. “Value Co-Creation in Data-Driven Services: Towards a Deeper Understanding of the Joint Sphere,” Proceedings of the Fortieth International Conference on Information Systems (ICIS), Munich. Schüritz, R., Seebacher, S., and Dorner, R. 2017a. “Capturing Value from Data: Revenue Models for Data-Driven Services,” in Proceedings of the 50th Hawai International Conference on System Sciences, T. Bui (ed.), Waikoloa Village, Hawaii, USA, pp. 5348-5357. Schüritz, R., Seebacher, S., Satzger, G., and Schwarz, L. 2017b. “Datatization as the Next Frontier of Servitization: Understanding the Chal enges for Transforming Organizations,” Proceedings of the Thirty-Eighth International Conference on Information Systems (ICIS), Seoul. Seiberth, G., and Gründinger, W. 2018. “Data-driven business models in connected cars, mobility services and beyond,” BVDW Research, available at https://bvdw.org/datadrivenbusinessmodels/. Sein, M. K., Henfridsson, O., Purao, S., Rossi, M., Lindgren, R., Sein, Henfridsson, Purao, Rossi, and Lindgren. 2011. “Action Design Research,” Management Information Systems Quarterly (35:1), pp. 37-56 (doi: 10.2307/23043488). Simmert, B., Ebel, P. A., Peters, C., Bittner, E. A. C., and Leimeister, J. M. 2019. “Conquering the Chal enge of Continuous Business Model Improvement,” Business & Information Systems Engineering (61:4), pp. 451-468 (doi: 10.1007/s12599-018-0556-y). 560 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Sosna, M., Trevinyo-Rodríguez, R. N., and Velamuri, S. R. 2010. “Business Model Innovation through Trial-and-Error Learning,” Long Range Planning (43:2-3), pp. 383-407 (doi: 10.1016/j.lrp.2010.02.003). Terrenghi, N. 2019. Design, Implement, Repeat: Essays on Business Model Management in Of line-Born Organizations. Tesch, J., and Bril inger, A.-S. 2017. “The Evaluation Aspect of Digital Business Model Innovation: A Literature Review on Tools and Methodologies,” in Proceedings of the 25th European Conference on Information Systems, I. Ramos, V. Tuunainen and H. Krcmar (eds.), Guimarães, Portugal, pp. 2250-2268. Tesch, J., Bril inger, A.-S., and Bilgeri, D. 2017. “Internet of Things Business Model Innovation and the Stage-Gate Process: An Exploratory Analysis,” International Journal of Innovation Management (21:5), 1740002-1 - 1740002-17 (doi: 10.1142/S1363919617400023). Wiener, M., Saunders, C., and Marabel i, M. 2020. “Big-data business models: A critical literature review and multiperspective research framework: A Critical Literature Review and Multi-Perspective Research Framework,” Journal of Information Technology (35:1), pp. 66-91 (doi: 10.1177/0268396219896811). Winterhalter, S., Weiblen, T., Wecht, C. H., and Gassmann, O. 2017. “Business model innovation processes in large corporations: Insights from BASF,” Journal of Business Strategy (38:2), pp. 62-75 (doi: 10.1108/JBS-10-2016-0116). Wirtz, B. W. 2011. Business Model Management: Design - Instruments - Success Factors, Gabler. Wirtz, B. W., and Daiser, P. 2018. “Business Model Innovation Processes: A Systematic Literature Review,” Journal of Business Models (6:1), pp. 40-58. Yin, R. K. 2009. Case study research: Design and methods, Los Angeles: SAGE. Zolnowski, A., Anke, J., and Gudat, J. 2017. “Towards a Cost-Benefit-Analysis of Data-Driven Business Models,” in Proceedings der 13. Internationalen Tagung Wirtschafstinformatik (WI 2017), J. M. Leimeister and W. Brenner (eds.), pp. 181-195. FUTURE COMPETENCES FOR THE EUROPEAN SOFTWARE SECTOR: A MIXED-METHOD APPROACH PAUL MORSCH, WILLEMIJN VAN HAEFTEN, PASCAL RAVESTEYN, GUIDO ONGENA HU University of Applied Sciences, Research Group Process Innovation & Information Systems, Utrecht, Netherland paul.morsch@hu.nl, wil emijn.vanhaeften @hu.nl, pascal.ravesteijn@hu.nl, guido.ongena@hu.nl The shortage for ICT personal in the EU is large and expected to increase. The aim of this research is to contribute to a better understanding of the roles and competences needed, so that education curricula can be better aligned to evolving market demand by answering the research question: Which competence gaps do we need to bridge in order to meet the future need for suf iciently qualified personnel in the EU Software sector? In this research, a mixed method approach was executed in twelve European countries, to map the current and future needs for competences in the EU. The analyses shows changes in demand regarding technical skil s, e.g. Keywords: software low-code and a stronger focus on soft skil s like communication sector, and critical thinking. Besides this, the research showed skil s, educational institutes would do well to develop their curricula in future, roles, a practical way by integration of real live cases and work together competences, with organizations. curricula DOI https://doi.org/10.18690/um.fov.4.2023.36 ISBN 978-961-286-751-5 562 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Current education and training programs are unable to meet the rising demand for properly skilled staff in the European software industry. Several studies highlight the alarming figures concerning the increasing shortages of ICT professionals in a growing number of different roles with related skills requirements (European Commission, 2019; European Commission, 2020, Beckett & Daberkow (2019)). To close this growing skil s gap, Europe needs a new strategy for software skills that can accelerate workforce upskilling and reskilling (European Commission 2017). Technological developments have changed the way software and ICT infrastructure are being designed, delivered, and managed. This enables the development of more services using new types of technologies such as cloud computing, AI, Machine Learning and Blockchain. Moreover, technological developments have increased the automation of aspects such as testing, deployment, management of new releases and introduced new approaches to working in teams to create software and services, like DevOps (ESSA, 2021). This cal s for shorter and more efficient education cycles. However, the pace of technological change makes it difficult for education and training providers to keep up and adapt their curriculums to the changing software skills demand of the market. Vocational education is seen as one solution to this problem as it provides an appropriate format for more flexible learning pathways and better connection with industry and employer needs. This research aims to contribute to the timely anticipation of the expected future scarcity of sufficient and properly trained staff in the EU software sector. We thus postulate the following research question: • Which competence gaps do we need to bridge in order to meet the future need for suf iciently qualified personnel in the EU Software sector? To be able to answer this research question, the researchers answered two sub-questions: • What are the current profiles and competences needed in the European software industry? • What are the future profiles and competences needed in the European software industry? P. Morsch, W. van Haeften, P. Ravesteyn, G. Ongena: Future Competences for the European Software Sector: A Mixed-Method Approach 563 The structure of this paper is as follows: In section 2 the methodology and its substantiation are described. In section 3, the findings of the research are presented. The 4th section provides a description of the conclusions and final y in section 5 the limitations of this research and recommendations for further research are discussed. 2 Methodology This chapter outlines the methodology used to investigate the current and future demand for ICT software development skil s. To better understand the fast-changing software services environment and competences needed both quantitative and qualitative data is collected. Specifical y, as “different methods have the potential to enrich our understanding of business problems and questions” (Molina-Azorin, 2016). Thus, the study employed a mixed-method approach, comprised of a literature review (in which both practice-based and scientific articles are included), focus groups and a job vacancy analysis. 21 European partner organizations, both academic and non-academic from 8 different countries, were involved in this study. The purpose of the study was to provide a comprehensive and accurate understanding of the current and future demand for ICT software development skills. Triangulation was used to validate the findings of the literature review with the results of the questionnaire, job vacancy analysis, and focus groups. The literature review provides guidelines for the survey and the focus groups as it specified the foundation for the roles in ICT software development that are important. The process and methods fol owed to identify gaps in the market need and provide recommendations is shown in figure 1. For the current demand, the primary data is found in labour market reports, as these reports focus on the types and numbers of skil ed workers that are needed. Complementary to the labour market reports, during the expert group meetings the need for skil ed workers and the way organizations fil these needs is discussed. Furthermore, the current need of ICT software development skills is also analysed with data collected by a questionnaire. As a validation we col ected and analysed job vacancies to determine if the competences mentioned in the text match the preliminary findings from our other data. 564 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY To also look forward to future ICT software development skils, the focus groups gathered data on the experts' views on the future skil s needed. Figure 1: Research Process and methods An overview of the data col ected is shown in table 1. Below we briefly discuss each of the data col ection methods. Table 1: Quantitative characteristics data col ection Data collection in numbers Academic literature study 18 selected top papers for detailed analysis Labour market desk research 63 national labour market reports 14 national labour market databases 905 job vacancies, 20 countries Demand questionnaire 301 respondents, 21 countries Expert groups in 9 different 10 national expert groups, 118 experts countries 1 European expert group, 14 experts Literature review Scientific papers A systematic review approach has been adopted following Webster and Watson (2002). In an iterative approach four researchers together with their teams from four different countries and universities have comprised the fol owing search terms: P. Morsch, W. van Haeften, P. Ravesteyn, G. Ongena: Future Competences for the European Software Sector: A Mixed-Method Approach 565 ‘Software engineering ‘, ‘software skills’, ‘Digital transformation', ‘Industry 4.0’, 'knowledge’, ‘software education’, ‘future', ‘programming’, ‘computational thinking’, ‘logic’. These terms ware used to search in the fol owing databases: Scopus, ACM, IEEE, Arxiv and the AIS Library. Only papers published from 2018 onward and peer-reviewed, are included in the analyses. The initial 84 papers identified through the primary database searches, are analysed based on the abstract and title. This provided 44 Articles focused on software development competences and also provided a future orientation of needed skills (Scopus: 7, Acm: 10, IEEE: 6, Arxic: 11, AIS Library: 10). All papers were discussed by al researchers to determine to which extent the papers were focused on the topic of software development skil s. Based on this full paper analysis, 18 papers (see table 2) were included in our study (Scopus: 3, Acm: 2, IEEE: 3, Arxic: 4, AIS Library: 6). Complementary to the systematic review, a snowball-method has been adopted to review older papers, 1 has been included into the research because it contained some insights relevant for the research aims. 566 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 2 Included papers and short characteristic or most revealing insight Authors Insights Albino & Souza Description of the gaps (skil s, commitment) of firms going through a (2019) digital transformation by a literature review. Beckett & Evaluation of chal enges determining current and future competences Daberkow (2019) in software sector by literature review. Cico et al. (2021) Integrate technologies and practices from software industry in curricula. Duan et al. (2020) Framework for adopting digital work in organizations. Engelbrecht et al. Necessity of context and comprehensive competences (+is (2018) challenging). Föll et al. (2018) Job vacancies and literature used to identify skills demand. Competence Framework are useful to develop curricula. Garousi et al. (2018) Prioritise the gap between software engineering and market needs. Günay et al. (2020) Description of need for and how to develop domain-specific critical thinking in education. Heintz & Manilla Students re-engineer the problem-solving process of computers. (2018) Hoover et al. (2019) Integration of AI and machine learning in creative domains like arts to establish deep learning and enlarge the programming awareness. Licorish & Dependence of role types/engagement of project members and MacDonel (2016) necessary skills. Quezada-Sarmiento Emphasis in education for developing creativity and innovativeness. et al. (2018) Reinhart & Experiences of group students taught software engineering skil s. Genovese (2019) Ryan (2020) Develop curricula beyond Industry needs (e.g. ethics, social skil s). Shaba et al. (2019) Effects on organizations after implementing Industry 4.0 technology. Thorat & Identifies loopholes in software education and gives possible solutions. Kshirsagar (2021) Schmidt (2020) Students collaborate in engineering projects representative for real-life. Zabavnik et al. UML skil s for the Automotive Industry. Avoiding mistakes is key. (2019) P. Morsch, W. van Haeften, P. Ravesteyn, G. Ongena: Future Competences for the European Software Sector: A Mixed-Method Approach 567 Labor reports In addition, researchers from 14 different countries collected and analysed 63 national labour market reports and 14 national labour market databases. On European level statistics by Eurostat were also used for analyses. We focused on col ecting data regarding ICT labour markets, and more specifically software development roles on a national or European level. Records found include white papers, formal government policy documents and professional bodies reports (such as commissioned research ‘state of the art’ reviews regarding software trends and competences). The selected databases are provided and maintained by a national statistics office, by other government agencies and a private institute depending on national context. Job vacancy analysis A job vacancy analysis was conducted to validate the current demand for software roles and competences in the EU. The methods and strategies to gather the information was left to the twelve national research teams. During the period of two months this analysis was executed the research teams exchanged the methods and the strategies for gathering their data in project meetings, so adjustments and insights were shared. 905 job vacancies from 20 countries were included. The job vacancy data were analysed to identify the most in-demand software roles and the level of proficiency required for these jobs. This data is used as input for both the results and to validate the findings of the questionnaire and focus groups. Questionnaire A questionnaire was designed to col ect data on the current and a perception of the future demand for ICT software development roles based on the project requirements and input from partners. The first phase was to invite partners to provide questions they considered relevant in the questionnaire. The first version of the questionnaire was drafted based on the requirements of the project and this input. This version was improved based on feedback from the partners. The next version was implemented in the EU Survey and tested by 17 people. The adjustments made based on this test resulted in the final version which was distributed via the partners of the consortium. 301 respondents have filled in this 568 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY questionnaire, working in various industries in various roles (mainly HR and management). The questions were designed to collect data on the demand for software development roles, the proficiency required and possible upcoming roles. The questionnaire consisted of 4 parts, 1) 4 general questions on the country and type of organization in which the respondent works, 2) 7 questions about what are considered the most relevant ICT professional role profiles, 3) 5 questions on software skil s that are not yet met in the organization, and 4) 11 questions on what the views of the organization are on training needs in relation to the identified roles and skills. Focus groups In 10 national focus groups, 118 experts participated. These meetings were held both online and offline in 2 to 3 hours sessions with 6 to 20 experts. In addition, one international focus group was held with 14 experts from different European countries. The focus groups were led by moderators who ensured the experts would feel free to discuss knowledge exchange and building new insights. The experts were selected based on their insights into future ICT developments. There was a mix of experts from HR, ICT, people working in the software sector and the research field. Types of candidates who were involved were: • Heads of HR departments of ICT companies (responsible for the staffing strategy); • Managers of employment agencies or recruitment agencies specialising in ICT; • CIOs of large companies (responsible for ICT strategy and sourcing). The participants were asked to reflect on their practices regarding the future demands for software roles, competences, needs and ICT-training. P. Morsch, W. van Haeften, P. Ravesteyn, G. Ongena: Future Competences for the European Software Sector: A Mixed-Method Approach 569 3 Findings The results of the executed research components were analyzed, evaluated and discussed in international working groups, composed by method and overal across al research components. The most important findings based on al components of this mixed-method research are discussed and presented in this section. During this data evaluation phase, it appeared that some of the valuable insights were not just about the roles or competences of the future of the software sector but were more about the learning process. For this reason, in this section the results are divided into three sections. The first two sections discuss the results to answer the two sub-questions (roles and competences). A third section is added which describes the insights and results on a more abstract level, the educational needs. What roles are needed in the future? 1. Developers needed most. Overal , we found that the role of developer is stil the most important role for the software sector. More than half of al respondents are currently looking for extra developers. Both the current as predicted future demand for software developers is high. However, there is a shift toward a new kind of developers, like full stack developers and low code developers. Another important progress mentioned is that the importance of developers increases as they are more integrated in the regular organization (business) instead of working from a separate entity (IT department). Soft skil s and knowledge of the business are therefore increasingly important for developers to be able to function. 2. Growing importance of the DevOps engineer We found a growing attention for the DevOps expert role which can be explained by the need for better integration between software production and the rest of an organization. Most DevOps job vacancies are from large companies. The DevOps role integrates development, deployment and maintenance and is the developer in agile environments using e.g., continuous integration and continuous delivery. 570 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3. Great need for solution designers The outcomes of the analyses shows there is a great need for solution designers in various industries. Solution designers play a critical role in developing effective solutions for complex problems. The largest number of vacancies are found in respectively in the programming sector 24,8% followed by the consultancy branche (18,3%). Furthermore, we found that many companies are looking for solution designers with senior-level experience, making it chal enging for entry-level designers to find suitable job opportunities. What competences are needed for future software development roles? 1. Programming principles It is not clear which new programming languages emerge or which of the current languages will last in the long run and are needed in the future. Therefore, it is difficult to determine which programming languages to include in a software training program to meet changing market requirements. A way to solve this problem is to make sure that software professionals have a solid foundation in understanding the underlying programming principles. This will help to adapt quicker to new languages and make software professionals more flexible and future proof. 2. Important profession related competences Because of its importance, security needs to be an integral part of the whole cycle of designing, developing, deployment and maintenance. Although currently not considered highly important, it is expected that in the (near) future sustainability management and sustainable software development wil become important. Besides these, also an understanding of the business and its needs (Business-IT-alignment) wil continue to grow in importance. These types of profession related competences are rapidly becoming more important for software professionals and should therefore have a prominent place in educational programmes. P. Morsch, W. van Haeften, P. Ravesteyn, G. Ongena: Future Competences for the European Software Sector: A Mixed-Method Approach 571 3. Interpersonal soft skills People in software roles need interpersonal soft skills since many activities require collaboration. The most relevant skil s found in our study are teamwork and general communication skills. During the focus group meetings, it was also often mentioned that the ability to communicate in English is increasingly important as it is often the standard language in software development projects. An additional benefit is that having good English language skills can increase job mobility which could benefit both the professionals as organizations searching for employees. 4. Personal soft skills are key Across al aspects of this research, the growing importance of personal soft skil s for people in software roles was clear. The most important seem to be critical thinking & analysis, problem solving and self-management. These soft skills are also truly transversal since these three skil s are also the top three skil s of which the World Economic Forum (2020) concludes that they will be increasingly in demand by 2025 in the total population of organizations. In other words: these skil s are important for people in any working environment and certainly for software professionals. Future educational needs Based on the analyses of the transcripts of the expert group meetings we found that besides roles and competences there is also a need for educational processes to change. In Europe there seems to be a focus on educational programmes for entry- level software roles. However, given the large and growing demand for software professional reskilling and upskilling is essential. As time for training is limited, short and modular programmes focusing on new technology trends are necessary. Collaboration between businesses and educational providers is needed to bridge the gap between education and practice, for example through company academies. Especial y large organizations can (and already do) organise their own tailormade educational programmes. Participants of the expert meetings even went as far as stating that flexible, lifelong learning of software professionals should start ideal y in primary school with a focus on programming logic and should continue throughout their careers to adapt to new 572 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY situations and technologies. Providing micro credentials can enable professionals to build up their resumes. 4 Conclusions Based on this research we conclude that the developer role is the most crucial and in-demand role. The shift towards new types of developers, such as full-stack and low-code developers, highlights the need for developers to possess soft skil s and business knowledge. The growing attention towards DevOps engineering further emphasizes the need for better integration between software production and the rest of the organization, this is also true for the solution designer role. Organizations face a chal enge in finding applicants with the necessary competences, and employees often lack the opportunities to train themselves while on the job. The educational field needs to educate and train people to possess the right competences to fil these job vacancies. Based on the survey we also found that large organizations have a more significant need for people in software roles and implement solutions like a company academy to reskill and upskill personnel. Furthermore, we found that soft skil s are becoming increasingly important for software professionals. Personal soft skills like adaptability, resilience, and problem-solving abilities are key to succeed as a software professional. Interpersonal soft skills like teamwork and communication are foundational, and English language skills are crucial. Additional y, we found that project management and security management are profession-related skills that are also becoming more important. The research indicates that time for training is limited, making upskilling a challenge, therefore initial education is crucial for reducing the competence gap in entry-level software roles. Collaboration between businesses and educational providers is needed to bridge the gap between education and practice. Besides this, training with micro credentials are seen as enabler to further close the competence gap and keep software professionals up to date. P. Morsch, W. van Haeften, P. Ravesteyn, G. Ongena: Future Competences for the European Software Sector: A Mixed-Method Approach 573 5 Reflection and directions for further research As with al research there are some limitations that need to be discussed. First, the literature review focused on a short period for publications and had a strict set of search criteria to limit the number of findings. This means that important relevant literature might have been missed. Second, during the expert meetings the questions started interesting discussions that often needed to be cut short due to time constraints which might mean we have missed additional needs. Third, the amount of data across the different methods made analysis very time consuming. Therefore, it was conducted by multiple researchers from different countries who then came together to combine their findings. However, it would have been better if paral el analyses of data were evaluated more often to get more alignment between the different researchers during their analysis process. To counter these limitations, we recommend that this research is conducted on a yearly basis. Hereby it is possible to keep up with the technological developments and the fast-emerging roles and competences needed in the (near) future. An important research question that remains unanswered in this study concerns the way in which the enormous and stil growing shortage of ICT professionals can be reduced. To close the gap between supply and demand in the future, it is important to also identify the supply side of educational programs. This enables organizations to choose the right training strategy that can train people to become exactly these professionals that the market demands. Recommendations for practice Prioritize initial education to close the competence gap for entry-level software roles, as the limited time for training makes upskilling challenging. Develop short, modular, and micro credentialing programs to help software professionals stay up to date with the newest technologies and trends. Integrate the development of soft skil s and broader education for software professionals to become T-shaped and Pi-shaped professionals who can bridge the gap between different disciplines. 574 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Encourage collaboration between businesses and educational providers to create better-tailored education and training programs that include real-life examples and environments for practice. And let businesses support the learning process by providing opportunities for software professionals to apply their skil s to real-world problems. Just be aware of the risks of this collaboration like ethical contradictions or a superficial learning process due to a practical interest of business. Acknowledgements This research was executed as part of the European Software Skil s Al iance (ESSA) project which is funded under the Erasmus+ programme (621751-EPP-1-2020-1-BE-EPPKA2-SSA-B). References Albino, R. and Souza, C., "Information and Technology’s role and digital transformation chalenges: a systematic literature review" (2019). CONF-IRM 2019 Proceedings. 5 https://aisel.aisnet.org/confirm2019/5 Allmér, H. (2018). Servicescape for digital wellness services for young elderly. Åbo Akademi University Press, Turku, Finland. Attig, C., Franke, T. (2020). Abandonment of personal quantification: a review and empirical study investigating reasons for wearable activity tracking attrition. Computers in Human Behavior, 102, 223-237. Beckett, R., & Daberkow, T. (2019). Work 4.0 and the Identification of Complex Competence Sets. MWAIS 2019 Proceedings, 0–6. https://aisel.aisnet.org/mwais2019/33 Duan, L., Cui, S., Qiao, Y., & Yuan, B. (2020). Clustering based on supervised learning of exemplar discriminative information. Ieee Transactions on Systems, Man, and Cybernetics: Systems, 50(12). https://doi.org/10.1109/TSMC.2018.2870549 Engelbrecht, L., Landes, D., & Sedelmaier, Y. (2018). A didactical concept for supporting reflection in software engineering education. IEEE Global Engineering Education Conference (EDUCON) ESSA - European Software Skills Alliance (2021). Europe’s Most Needed Software Roles and Skills: Needs Analysis Report 2021. Deliverable D.4.: “Software Services Profiles, Competences, and Skills Needs”. European Commission (2017) Digital Skills and Jobs Conference – Digital Opportunities for Europe, https://digital-strategy.ec.europa.eu/en/library/digital-skills-and-jobs-conference-7-december-2017 European Commision (2019) Analysis of shortage and surplus occupations based on national and Eurostat Labour Force Survey data 2019, https://ec.europa.eu/social/main.jsp?catId=738&langId=en&pubId=8269&furtherPubs=ye s European Commision (2020) Analysis of shortage and surplus occupations 2020, https://op.europa.eu/en/publication-detail/-/publication/22189434-395d-11eb-b27b-01aa75ed71a1/language-en Föl , P., Hauser, M., & Thiesse, F. (2018). Identifying the skil s expected of IS graduates by industry: A text mining approach. International Conference on Information Systems 2018, ICIS 2018, 1-17 Garousi, V., Giray, G., Tüzün, E., Catal, C., & Felderer, M. (2019). The Journal of Systems and Software Aligning software engineering education with industrial needs: A meta-analysis. The Journal of Systems and Software, 156, 65–83. htps://doi.org/10.1016/j.jss.2019.06.044 P. Morsch, W. van Haeften, P. Ravesteyn, G. Ongena: Future Competences for the European Software Sector: A Mixed-Method Approach 575 Günay, C., Doloc-Mihu, A., Barakat, R., Gluick, T., & Moore, C. . (2018). Improving Critical Thinking in Software Development via Interdisciplinary Projects at a Most Diverse College. SIGITE 2020 - Proceedings of the 21st Annual Conference on Information Technology Education Heintz, F., & Mannila, L. (2018). Computational thinking for al : an experience report on scaling up teaching computational thinking to al students in a major city in sweden. Acm Inroads, 9(2), 65– 71. https://doi.org/10.1145/3159450.3159586 Hoover, A. K., Spryszynski, A., & Halper, M. (2019). Deep learning in the IT curriculum. SIGI2019 - Proceedings of the 20th Annual Conference on Information Technology Education, 49–54. https://doi.org/10.1145/3349266.3351406 Licorish, S. A., & MacDonell, S. G. (2021). Exploring software developers’ work practices: Task differences, participation, engagement, and speed of task resolution. Information and Management, 54(3), 364–382. https://doi.org/10.1016/j.im.2016.09.005 Molina-Azorin, J.F. (2016). Mixed methods research: An opportunity to improve our studies and ourresearch skil s. European Journal of Management and Business Economics, 25(2), pp. 37-38. Reinhart, A., & Genovese, C. R. (2020). Expanding the scope of statistical computing: Training statisticians to be software engineers Ryan, K. (2020). We should teach our students what industry doesn’t want. Proceedings - International Conference on Software Engineering, 103–106. https://doi.org/10.1145/3377814.3381719 Schmidt, J. Y., & 51st ACM SIGCSE Technical Symposium on Computer Science Education, SIGCSE 2020 51 2020 03 11 - 2020 03 14. (2020). Reviewing cs1 materials through a collaborative software engineering exercise: an experience report. Annual Conference on Innovation and Technology in Computer Science Education, Iticse, 379-385, 379–385. https://doi.org/10.1145/3328778.3366932 Schmidt, J. Y., & 51st ACM SIGCSE Technical Symposium on Computer Science Education, SIGCSE 2020 51 2020 03 11 - 2020 03 14. (2020). Reviewing cs1 materials through a collaborative software engineering exercise: an experience report. Annual Conference on Innovation and Technology in Computer Science Education, Iticse, 379-385, 379–385. https://doi.org/10.1145/3328778.3366932 Shaba, E., Guerci, M., Gilardi, S., & Bartezzaghi, E. (2019). Industry 4.0 technologies and organizational design - Evidence from 15 Italian cases. Studi Organizzativi, 1, 9–37. https://doi.org/10.3280/so2019-001001 Thorat, S. A., & Kshirsagar, D. P. (2021). Developing logic building, problem solving, and debugging programming skil s among students. Journal of Engineering Education Transformations, 34(Special Issue), 402–406. https://doi.org/10.16920/jeet/2021/v34i0/157188 Webster, J., & Watson, R. T. (2002). Analyzing the past to prepare for the future: Writing a literature review. MIS quarterly, xiii-xxiii. World Economic Forum (2020) The Future of Jobs Report 2020, http://www3.weforum.org/docs/WEF_Future_of_Jobs_2020.pdf. Zabavnik, J., Riel, A., Marguč, M., & Rodič, M. (2019, July 3). Knowledge and skil s requirements for the software design and testing of automotive applications. SAAEI . https://hal.archives-ouvertes.fr/hal-02335566 576 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY TOWARDS A NEW APPROACH TO COMPETENCES IN DIGITAL TRANSFORMATION XOCHITL CRUZ,1 RAFAEL PALACIOS2 1 Autonomous University of Queretaro, School of Accounting and Administration, Santiago de Queretaro, Mexico xochitl.margarita@yahoo.com 2 BSP Business & Law School Berlin, Germany rafael.bustamante@businessschool-berlin.de Digital transformation necessitates knowledge in innovation management, organizational adaptation, and adaptability, crucial for business strategy. Relying solely on traditional knowledge capabilities inhibits innovation within organizations. Companies must develop specific competences. The chal enge also lies in managing the development of these competences within the dual context of specialization and breadth. This research aims to identify and analyze competences in the artificial intel igence sector, ultimately contributing by identifying a set of competences that can serve as knowledge domain capabilities for rapid and efficient adaptability to digital transformation. The research aims to determine which theoretical variables hold the necessary validity to form a specific set of competences required by companies in the context of digital transformation. The study focuses on technology-based companies in Mexico involved in Artificial Intelligence, examining the presence and occurrence of selected variables. By doing so, the hypothesis that these competences influence new forms of innovation management Keywords: and value generation in organizations will be verified. The competencies, research employs a comprehensive conceptual framework to digital transformation, identify fundamental elements related to labor competences, artificial obtaining a set of variables through theoretical contrast methods. intelligence, These variables were later operationalized using machine innovation management, learning and mining techniques, al owing for the observation and organizational discovery of new competences. adaptation DOI https://doi.org/10.18690/um.fov.4.2023.37 ISBN 978-961-286-751-5 578 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The dynamics presented by digital transformation is characterized by accelerated exponential growth that begins with an almost imperceptible pace, but grows faster as time progresses (Escott, 2020; Escott et al., 2020; Palacios et al., 2021 According to Kurzweil (2005), intel igent technologies are twice as fast as the rest of the average technologies. This implies that new technologies will require new capabilities for their use. Organizations with traditional schemes must seek their adaptation towards collaborative structures, innovative, competitive and transversal (Romero, 2016; Cabrerizo, 2016; Maydón, 2016; Corujo & Cortés, 2016; Covarrubias, 2016).In this sense, the competencies of the employees would represent -at least, in some time determined- a factor that would al ow organizations to strengthen competitiveness, increase their innovative and adaptive capacity (Micheli & Valle, 2018), particularly a dual area: broad and specialized capabilities (Palacios et al., 2021). Additional y, the crisis of the COVID-19 pandemic caused an abnormal acceleration in the development of non-routine digital tasks with a high level of qualification and collaboration (Brakman et al., 2021). Artificial Intelligence (AI) represents a fast cumulative force that, when combined with the appropriate human skills, allows organizations to adapt in changing environments of digital transformation (Malerba, 2020). The research asks the question of which theoretical variables could have the necessary validity to form a set of specific competences required by companies in the context of digital transformation. This is how this research addresses the development of an approach oriented towards new skil s and knowledge. Therefore, it is necessary to make theoretical contributions that al ow characterizing this type of competencies (Escott et al., 2020; Palacios et al., 2021). The research uses qualitative and quantitative fundamental analysis. To contrast various theories, the methodology proposed by Marquina, Álvarez, Guevara & Guevara (2013) was used, which allows identifying, systematizing, and simplifying variables; Furthermore, it has been strongly used in recent studies related to innovation (Escott, 2020; Palacios et al., 2021). Thanks to this methodology, it was possible to make an approximation to the concept of knowledge domain capacities. These capabilities according to Cantner & Vannuccini (2018) describe the organizational innovation capabilities. Final y, Machine Learning techniques are used to make predictions based on the X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 579 cases analyzed (Mohri et al., 2018), and learning algorithms are created based on the observations made (Zhou, 2021). 2 New competencies in the innovation management framework According to Schumpeter (1942), technological change is highly dynamic and is shaped by waves of innovation generated by industrial transformation. Innovations change the natural course of technology, are generated endogenously and are driven by the innovative attitude of the entrepreneur (Antonelli, 2008). Similarly, throughout history, five waves have been identified that have led to transformations in techno-economic paradigms (Pérez, 2010). Table 1 al ows us to identify how the demand for competencies has also been changing, where initial y mainly instrumental jobs were required until interpersonal and systemic competencies were reached. In previous industrial revolutions, physical work was replaced, however, currently it is sought to replace, in addition, organizational and personal cognitive processes (Pérez, 2010). Despite the fact that various authors identify technological change as responsible for the destruction of jobs, it also generates innovation and productivity in addition to offering the possibility of creating new jobs (Estevadeordal & Beliz, 2017; Iglesias et al., 2020; Schwab, 2017). By improving the skills of employees, it is possible to guarantee not only the competitiveness of the organization, but also al ows the mobility of employees towards more complex positions (Erbes et al, 2008; Roitter et al. 2013). Currently, the existing gap between the skills demanded and those offered is wide (Brynjolfsson & McAfee, 2014; Autor, 2015; Sasmita & Kumar, 2018) and it is estimated that more than 40% of workers must be re-trained (Nwaohiri & Nwosu, 2021; Chakma & Chaijinda, 2020). The job profile will be mostly oriented to skil s instead of knowledge, problem solving instead of routine tasks, adaptation instead of monotony, transversal knowledge and continuous relearning (Montes et al., 2021; Chakma & Chaijinda, 2020). In this research, it has been decided to use the term competencies in an integration of knowledge, skills, attitudes, and experience in complex and fickle environments that are directed by the individual (Kanungo & Misra, 1992), of unlimited potential, permanently developed (Sarramona, 2007) and in a broad spectrum in the performance of non-routine tasks and situations. 580 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 1: Characteristics of industrial revolutions Technological Nucleus Key Production Organizatio Type of revolution country Resources of n of competenci and knowledge production es required period First: England Cotton, Learning by Individual Instrumenta Industrial 1770/178 iron, doing and entrepreneu l Revolution 0 - machinery, technology rs and competenci 1830/184 textiles, transfer micro- es 0 chemistry enterprises Second: Age England Coal, Learning by Micro Instrumenta of steam and 1830/184 transport, doing and companies; l railways 0 - steam institutionali the competenci 1880/189 engines, zation of emergence es 0 steel, gas, knowledge of large electricity companies Third: Age of USA, Acero. Internal Rise of large Instrumenta Steel, Germany Electrical R&D companies, l electricity and 1880/189 machinery departments monopolies, competenci heavy 0 - , oligopolies es engineering 1930/194 recruitment 0 of scientists Fourth: Age USA, Petroleum, Expansion Oligopolisti Instrumenta of oil, Europe automobiles of c l automobile 1930/194 , synthetic apartments competition competenci and mass 0 – materials R&D and , es and production 1970/198 education, multination interperson 0 technology al and al transfer foreign competenci direct es investment Fifth: Age of USA, Microelectr Horizontal Business Instrumenta information Europe, onics, ICTs, integration networks, l, technology Asia digital of R&D, production interperson and 1970/198 technologie design and systems and al and Telecommuni 0 to s, robotics, production platforms systemic cations present biotechnolo competenci gy es Source: Own elaboration based on Pérez (2010); Roitter (2019), González and Wagenaar (2003) X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 581 3 Methodology In order to identify, systematize and classify the different variables proposed in the literature by various authors, the theoretical contrast methodology proposed by Marquina, Alvarez, Guevara and Guevara (2013) were used. This approach proposes 4 stages: (a) Literature search, (b) literature exploration, (c) argument development and (d) literature criticism. Table 2 breaks down the tasks that were developed in each stage. Table 2: Application of the MAGG methodology STAGE APPLICATION RESULTS LITERATURE 1. The object of study is established: 1. The necessary theoretical SEARCH Identification of the components related framework for the to knowledge domain capacities. Some delimitation of the research keywords used were: knowledge, skills, work. creativity, innovation, competence, 2. Definition of the key capability, aptitude, proficiency, talent, concepts that will be used in ability, efficiency, adaptability. the research development. 2. The literature for the theoretical framework considered various authors over time based on the main theories and their main theoretical exponents. 3. Through a quick reading of the literature, the works that must be included are selected and the rest are excluded. 4. The selected literature is read gradually, starting with a general reading until reaching a deep and specific reading. 5. The main arguments of each author are identified. For this purpose, an open coding was used, where the data was carefully examined, breaking it down into small phrases or paragraphs and assigning descriptive codes that represent their content. This coding required openness to new concepts without limitations or preconceptions. 582 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY STAGE APPLICATION RESULTS LITERATURE 6. The evolution of the topic over time 1. Main ideas of each author EXPLORATION is analyzed. identifying the specific 7. The related variables that affect the information that should not concept of competencies are identified. be omitted due to its In this stage, axial coding was used, significant contribution, aiming to establish connections and including: author, year of relationships between the codes publication, title of identified during the open coding phase publication, journal or (Step 5). Related codes were grouped publisher supporting the into broader categories or thematic publication. concepts. The relationships between 2. Selection, these categories were explored, and systematization, and conditions, contexts, actions, and classification of different consequences that are interrelated were theoretical perspectives identified. ARGUMENT 8. The literature is classified based on 1. The consistency matrix: DEVELOPMENT the approaches given by each author on allows for describing, the concept of competencies in order to categorizing, and analyzing identify their arguments in an orderly the obtained data. This manner. matrix establishes the 9. A first table is prepared where the evidence from the author, the definition and the variables theoretical review. Useful used with respect to the theoretical variables for the research approach and his views on the are considered based on competences are identified. their quality and relevance. 10. The classifiers of each variable are proposed, which al ow the variables to be unified. During the contrasting process, variables with similar meanings were grouped together in the same classification. Using the same procedure, variables with contrasting characteristics were classified separately, creating a new classification. In this stage, selective coding is used to select and develop the central or fundamental concepts that are relevant for building a grounded theory. The most significant codes are identified. This process involves reducing and consolidating the codes, as well as X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 583 STAGE APPLICATION RESULTS establishing relationships between the identified concepts. LITERATURE 11. A first table is analyzed, which 1. Contrasting theories CRITICISM allows a comparative observation of the 2. Integration of arguments of each author, identifying perspectives, approaches, the similarities and differences related and elementary concepts to the theoretical approaches of the addressed by the authors. competencies. 12. A second table is built that allows a theoretical comparison to be made, comparing the key concepts of each author from different approaches, revealing common concepts grouped through the proposed classifiers (Table 3) Source: Own elaboration based on Marquina, Álvarez, Guevara and Guevara (2013) After applying the MAGG methodology, sufficient evidence was found according to the patterns proposed by Machi & Evoy (2021), which allowed the identification of thirteen categorical variables: (1) Attitudes, (2) Strategic Alliances, (3) Capabilities, (4) Underlying characteristics or skills, (5) Knowledge, (6) Beliefs and values, (7) Entrepreneurship and innovation, (8) Standard of effectiveness, (9) Network structures, (10) Strategic architecture, (11) Experience, ( 12) Skills and (13) Productivity. Each category is made up of a group of variables that al ow it to be formed. The following table groups the thirty-four variables related to each category. Table 3: Variables by category No. Category Variable and Description Attitude: Variable that determines the traits, patterns or habitual behaviors that can be identified as patterns, tendencies and 1 Attitude dispositions to act in a certain way. The attitude can be changed or control ed. (Spencer & Spencer, 1993) Strategic Al iances: Cooperation between various entities participating in innovation that aims to improve capacities, fill 2 Strategic Al iances knowledge gaps and promote innovation from the endogenous and exogenous sphere (Lichtenthaler, 2016; Bell & Pavitt, 1995). 584 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY No. Category Variable and Description 3 Infrastructure: Equipment and physical resources necessary to be able to adequately perform a task (Tejada, 1999) Ad hoc innovation: Practices that arise as a result of taking advantage Capabilities of the internal knowledge of the workers of an organization. They 4 frequently appear from adverse economic or technological situations rather than as a result of long-term strategic decisions. (Teece et al., 1997) Aptitude: Characteristics of a certain permanence to the individual Underlying that al ow him to carry out different types of tasks and that facilitate 5 characteristics or his learning in new ones. They are often defined as innate derived from attitudes the fact that they are naturally acquired during childhood. (Sierra, 2017; Lévy-Leboyec, 2003) Competencies: It is part of an individual's personality; it is related to a standard of effectiveness and superior performance in a situation. 6 Knowledge can be understood and transferred by other subjects or organizations (Spencer & Spencer, 1993; González & Wagenaar, 2003; Rychen & Salganik, 2003; Gallardo et al., 2015). Accumulated Experience: Obtained from the performance of specific activities over time, built with reflection and from lived events. 7 Interpretation of the individual from the environment and its Knowledge perception filters (Kolb, 2014). Creativity: Innovative capacity, actions that identify, analyze and 8 solve problems in a divergent, original, effective and adaptable way (Valiente, 2017). Knowledge: Transformative learning, accumulated over time. Appropriate tacit experience in the learning by doing process. 9 Commonly known as know-how. (Cohen & Levinthal, 1990; Bell & Pavit, 1995) Social exchange: Ethical competence where, through the internalization of business objectives, transparency and awareness, the individual perceives his context and acts according to what he 10 Beliefs and considers socially important. Effort that the individual will make Values within the organization. Aspects such as the work environment, clarity of objectives and results expected of the individual are involved (Spencer & Spencer, 1993; Christis, 1988; Mertens, 1996). Distinctive competencies: Competences that are difficult for competitors to reproduce and derive from the ability to anticipate 11 market trends within an innovative and creative approach (Cohen & Entrepreneurship Levinthal, 1990; Mertens, 1996). and Innovation Inventive skil s: Combination of theoretical and practical aspects that 12 al ow generating creative and innovative solutions both external y and internally in the organization (Nadine, 1998). X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 585 No. Category Variable and Description Sensitivity: An individual's ability to timely detect the assessments of 13 the environment with respect to their proposal. It al ows to increase the number of innovations. (Marengo et al., 2021) Perseverance: Effort dedicated by an individual to improve his 14 proposal before moving on to the next function. Identifies the number of innovation possibilities (Marengo et al., 2021). Learning by doing: Learning acquired through experimentation that 15 al ows a company to innovate and align itself with the needs of the market (Benini, 2021). Incremental innovations: smal improvements made to an existing product or service in an organization. Commonly low cost and 16 represent a differentiator with respect to its competition (Freeman et al., 1982) Reverse engineering: Innovation tool general y used in emerging 17 economies. Analysis of an existing product or service in order to know its design and operation (Acosta, 2013; Benini, 2021) Effectiveness standard: Corresponds to the reference criteria 18 Effectiveness established to predict when a task is performed in an improved way standard with respect to the standard (Spencer & Spencer, 1993). Formal networks: Col aborations made up of connections with inventors, universities, government, sources of support, talent pool, 19 large corporations, incubators, accelerators or researchers (Neck et al., 2004). 20 Informal networks: Col aborations made up of family, friends and individuals in an informal way (Neck et al., 2004). Interdisciplinarity of knowledge. Col aboration carried out with individuals with different areas of study and that increases social skil s Network 21 and abilities to work as a team. That team must have clearly assigned structures its tasks and the structure that composes it (Moe et al., 2021; Thu et al., 2021). Internal networks: Participation ecosystems within the organization 22 promoted by its members as a proposal to work col aboratively on important projects (Thu et al., 2021). Proximity: Creation of territorial y concentrated organizations, with 23 qualified local labor, vertically or horizontal y integrated (Benini, 2021). Strategic units: Generates competitive advantages. Strategies that are 24 Strategic difficult to imitate derived from the combination of capabilities, architecture strategic elements, and knowledge (Brusoni, 2021). 25 Experience Informal activities: Execution of tasks that lack a written agreement and are based on casual employment. They usual y represent low 586 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY No. Category Variable and Description productivity and the individual lacks the necessary skills (Mertens, 1996). Informal activities related to competencies: Unorganized, acquired personal experiences linked to organizational products, processes, and 26 capabilities that enable effective knowledge acquisition (Coombs, 1996). Business failures: Crisis of a company derived from various aspects that have a negative impact on a company. Frequently related to financial failures, whether in an acute, chronic or terminal stage, 27 however, it occurs when the financial situation of the organization reflects a reduction in value or cash flow shortages (Durand & Lupaca, 2016). Knowledge of the founders: Determines the culture of the company. The previous experience of the founders general y gives the 28 organization a lasting effect on the capabilities, performance and learning process of the workers (Bahoo et al., 2021). 29 Skil s: Operational capacity of an individual to carry out an activity effortlessly (Nadine, 1998; Gallego-Arrufat, 2007). Complementary skil s: Skills related to customer and personal 30 service, judgment and decision making, technological design, flow of ideas and analysis of operations (Bakhshi et al., 2017) Skil s Interactive and adaptability skil s: skills that require in-depth knowledge of the organization, including the organizational culture, in 31 order to be able to adapt in the context of technological change (Holm & Lorenz, 2021). 32 Related tasks: Complementary activities performed daily by the individual (Carbonero et al., 2021). Routine tasks: Highly automatable activities, follow perfectly defined 33 processes and is made up of clear and repeated tasks (Frey & Osborne, Productivity 2013). 34 Capacities, Skil s and Competencies: Determine the priority and real importance of each variable within the organization Source: Own elaboration based on the application of Marquina, Álvarez, Guevara and Guevara (2013) 3.1 Methodological design and presentation cases of study and results The methodology was designed in three stages: a) obtaining variables related to domain knowledge capabilities, b) selection of companies with AI development1, 1 The selected companies are Mexican entrepreneurial companies in the field of AI. A total of 51 questionnaires were conducted with Mexican companies from various sectors and sizes that have integrated AI into their processes. Questionnaires were administered to companies located in 9 states of Mexico: Aguascalientes, Baja California Sur, Mexico City, Guanajuato, Jalisco, Nuevo León, Puebla, Querétaro, San Luis Potosí. According to INEGI (2022), X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 587 questionnaire preparation (For further information, please refer to the appendix), application, publication and sending of the questionnaire, c) validation of findings. Figure 1 describes the process carried out by stage. Figure 1: Methodological process of the research Source: Own In total, fifty-one questionnaires were applied to leaders of Mexican companies from various sectors that have integrated AI into their processes. The companies are segmented mainly in the tertiary economic sector according to the INEGI classification, the rest are part of the secondary sector in the category of Manufacturing companies. 4 Results and Discussions For the last stage, the results were analyzed using machine learning tools cal ed k-means and Principal Component Analysis (PCA), previously used in other applied research on skil development (Harlak et al., 2008; Nugent et al., 2010; Kuswandi et six of these states currently show a GDP higher than the national average. The surveyed individuals' profiles are closely related to managerial and executive positions with a focus on AI implementation and innovation management within their companies. 588 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY al., 2018; Shelly et al., 2020; Shamrat et al., 2020). The projection of the results can be seen in Figure 2. Strategy 1 Strategy 2 Strategy 3 Figure 2: Data projection Source: Own This figure allows us to infer the following results: (a) 3 important groups are identified: in quadrant 1, there is the productivity category that constitutes the first group, where companies such as 13, 46 and 29 would be applying strategies of this type, (b) Categories: attitude, strategic al iances and experience, where companies like 24 and 19 feel identified and finally (c) categories: underlying characteristics, skills, entrepreneurship and innovation, knowledge, strategic architecture, network structures, beliefs and values, standard of effectiveness and capabilities, where companies like 33 and 49 remain. In addition, companies that are not clustered with any strategy are observed in the third quadrant, which al ows us to infer that they do not coexist with the necessary theoretical variables within their objectives. This means that companies seek three types of strategies: the first exclusively based on promoting productivity, the second using strategic al iances and enhancing experience, and the third constitutes the union between skil s, entrepreneurship, innovation, knowledge, strategic architecture and network, beliefs, capabilities and X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 589 standards of effectiveness. However, the closeness between the variables of the second and third strategies indicate that both strategies can coexist, but not with strategy one, since it is opposed to the rest. If we analyze Figure 3 regarding the contribution of categories within the PCA, we can distinguish the existence of seven (7) highly representative variables. They are: 1) entrepreneurship and innovation, 2) knowledge, 3) experience, 4) beliefs and values, 5) effectiveness, 6) networks, and 7) Attitude. Figure 3: Contribution of the categories within the principal components. Source: Own elaboration based on information obtained from RStudio software. These variables are significant inputs not only for an approach that al ows for the theoretical configuration of them as knowledge domain capabilities in the high-tech sector but also together they can serve as inputs for both theoretical discussion and the validation of current theoretical perspectives on competencies in the context of digital transformation. regarding knowledge domain capabilities. The research also addresses the application of data mining techniques belonging to the branch of Data Science used in qualitative variables (Gupta et al., 2020). 590 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY The result obtained shows that responsibility represents a quality and challenge in the 21st century according to Fadel et al. (2016) and confirmed in the survey of businessmen; In addition, the COVID-19 pandemic required the opening of virtual workspaces that generated uncertainty (Schwab & Malleret, 2020) in the leaders regarding this scheme and had the chal enge not only of responsibility but also of communication, collaboration, trust, and coincident theoretical statements with the companies interviewed. On the contrary, those surveyed defined fear as the most important chal enge that a negative attitude would represent. In this sense, the COVID-19 pandemic provoked feelings within the population that included fear and that in many cases slowed down the process of acquiring new skil s. Faced with competition, businessmen highlighted the importance of innovation, which, according to Drucker (1985) is only carried out if it has an economic profitability effect. Finally, the surveys reveal that honesty is the most important social value within their organizations, defined by Ruiz et al. (2008) as highly significant within business ethics and university education (Osuna & Luna, 2011). In the use of text networks (Berry & Kogan, 2010; Vijayarani & Nithya, 2015; Zhang et al., 2015; Gupta et al., 2020; Kononova et al., 2021; Hickman et al., 2022) the surveys showed that teamwork, collaboration with communication, honesty with respect are groups of concepts that they value as a whole, while fear and distrust are factors that they frequently avoid. 5 Conclusions The methodology used not only al owed us to identify driving indicators of innovation processes related to competition in a dynamic environment, but also to reduce their dimensionality and identify findings that are difficult to observe in traditional analysis. Therefore, it is necessary to delve into theoretical and practical aspects that enable the development of techniques to enhance the variables analyzed within the organization, so that we can understand not only the 'what' but also the 'how'. The group of representative variables identified presents itself as determining elements that influence the development of a digital transformation strategy in companies in the artificial intelligence sector. Collectively, these variables shape the company's capabilities in knowledge domain and adaptability in a highly dynamic and changing context (technological change). X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 591 Competitive organizations in the midst of digital transformation strive to identify, retain, and develop new capabilities in their employees that enable them to coexist with new technologies and address the demands of technological change. These organizations exist in different sectors and sizes, which means that the need to enhance new skills is not exclusive to a particular industry. This research presents a finding (stil in an exploratory phase) regarding the relationship between competencies and specialization. A set of non-repetitive competencies can be identified through task specialization. The influence of technological changes outside the company drives the specialization of tasks. In this context, there is an expansion of task specialization influenced by external factors (Palacios et al., 2021; Escott, 2020) References Acosta, D. A. R. (2013). Uso de la ingeniería inversa como metodología de enseñanza en la formación para la innovación. Encuentro Internacional de Educación en Ingeniería. Antonelli, C. (2008), The Economics of Innovation, Londres, Routledge Major Works Autor, D. H. (2015). Why are there stil so many jobs? The history and future of workplace automation», en The Journal of Economic Perspectives, vol. 29, n.º 3, pp. 3-30. Bahoo, A., Fontana, R. & Malerba, F. (2021) Knowledge Contexts of Entrepreneurship, New Entrants’ Capabilities, and Performance Outcomes. 18th International Schumpeter Conference. Rome, Italia. Bakhshi, H., Downing, J. M., Osborne, M. A., & Schneider, P. (2017). The future of skil s: Employment in 2030. Pearson. Bell, M., & Pavitt, K. (1995). The development of technological capabilities. Trade, technology and international competitiveness, 22(4831), 69-101. Benini, R. (2016). Some key policy issues related to technology change, knowledge and absorption capacities in a country comparison perspective. Economic Change and Restructuring, 49(2), 95-112. Berry, M. W., & Kogan, J. (Eds.). (2010). Text mining: applications and theory. John Wiley & Sons. Brakman, S., Garretsen, H., & van Witteloostuijn, A. (2021). Robots do not get the coronavirus: The COVID-19 pandemic and the international division of labor. Journal of International Business Studies, 52(6), 1215-1224. Brusoni, S., Cassi, L., & Tuna, S. (2021). Knowledge integration between technical change and strategy making. Journal of Evolutionary Economics, 31, 1521-1552. Brynjolfsson, E., & McAfee, A. (2014). The second machine age: Work, progress, and prosperity in a time of brilliant technologies. WW Norton & Company. Cabrerizo, J. (2016) La transformación Digital en Unidad Editorial Cantner, U., & Vannuccini, S. (2018). Elements of a Schumpeterian catalytic research and innovation policy. Industrial and Corporate Change, 27(5), 833-850. Christis, J. (1988) Taylorisme en nieuwe produktieconcepties. Te Eldfer Ure, Nijmegen, n. 41. Chakma, S., & Chaijinda, N. (2020). Importance of reskilling and upskilling the workforce. วารสาร สห ศาสตร ์ ศรีปทุม ชลบุรี Interdisciplinary Sripatum Chonburi Journal (ISCJ), 6(2), 23-31. 592 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative science quarterly, 128-152. Coombs, R. (1996). Core competencies and the strategic management of R&D. R&d Management, 26(4), 345-355. Corujo, A. & Cortés, S. (2016) Todos somos startups. La auténtica transformación: Mitos y claves. Covarrubias, P. (2016). CTO vs CIO Drucker, P. (1985). Innovation and Entrepreneurship. ISBN: 0060154284, 9780060154288. Durand, C. D. D., & Lupaca, R. D. P. L. (2016). Origen, definición y modelos del fracaso empresarial: una revisión teórica. Revista de Investigación Valor Contable, 3(1), 47-56. Erbes, A., Roitter, S., & Delfini, M. (2008). Conocimiento, organización del trabajo y empleo en tramas productivas. Revista de Trabajo. Nueva Época, 4(5), 73-86. Escott, M.P. (2020). Digitalización cómo nuevo patrón tecnológico dominante: Implicaciones en la innovación universitaria en México. Universidad Autónoma de Querétaro. Tesis Doctoral. Escott, M. P., Palacios, R., & Cruz, X. M. (2020). The new complexity and new dynamics of technological change; and its effects on innovation management. IAMOT 2020 Conference Proceedings. (p.1063–1076). Cairo, Egipto: International Association for Management of Technology. Estevadeordal, A., & Beliz, G. (2017). The future of work in Latin American Integration 4.0. Integration and Trade Journal, 21(42). Fadel, C., Bialik, M., & Trilling, B. (2016). Educación en cuatro dimensiones: las competencias que los estudiantes necesitan para su realización. Freeman C., Clark J., Soete L. (1982) Unemployment and Technical Innovation: A Study of Long Waves and Economic Development, Frances Printer, London Frey, C. B., & Osborne, M. A. (2017). The future of employment: How susceptible are jobs to computerisation? Technological forecasting and social change, 114, 254-280. Gal ardo, T. E., Aguilar, J. F. L., & Aracil, A. G. (2015). Competencias Vs. Capacidades: ¿Enfoques complementarios o excluyentes? Gallego-Arrufat, M. J. (2007). Las funciones docentes presenciales y virtuales del profesorado universitario González, J. & Wagenaar, R. (2003). Tuning Educational Structures in Europe. Informe Final - Proyecto Piloto, Fase 1. Bilbao: Universidad de Deusto Gupta, A., Dengre, V., Kheruwala, H. A., & Shah, M. (2020). Comprehensive review of text-mining applications in finance. Financial Innovation, 6(1), 1-25. Harlak, H., Gemalmaz, A., Gurel, F. S., Dereboy, C., & Ertekin, K. (2008). Communication skil s training: effects on attitudes toward communication skil s and empathic tendency. Education for Health, 21(2), 62. Hickman, L., Thapa, S., Tay, L., Cao, M., & Srinivasan, P. (2022). Text preprocessing for text mining in organizational research: Review and recommendations. Organizational Research Methods, 25(1), 114-146. Holm, J. R., & Lorenz, E. (2021). The impact of artificial intelligence on skills at work in Denmark. New Technology, Work and Employment. Iglesias, E., García, A., Puig, P., & Benzaqué, I. (2020). Inteligencia artificial: la gran oportunidad del siglo XXI: documento de reflexión y propuesta de actuación. Banco Interamericano de Desarrollo. DOI: http://dx. doi. org/10.18235/0003037 [01-06-2021]. Instituto Nacional de Estadística y Geografía (INEGI) (2022). Indicador Trimestral de la Actividad Económica Estatal. Link: https://www.inegi.org.mx/temas/itaee/#Tabulados Kanungo, R. N., & Misra, S. (1992). Managerial resourcefulness: A reconceptualization of management skills. Human relations, 45(12), 1311-1332. Kolb, D. A. (2014). Experiential learning: Experience as the source of learning and development. FT press. Kononova, O., He, T., Huo, H., Trewartha, A., Olivetti, E. A., & Ceder, G. (2021). Opportunities and challenges of text mining in materials research. Iscience, 24(3), 102155. Kurzweil, R. (2005). The singularity is near: When humans transcend biology. Penguin. X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 593 Kuswandi, D., Surahman, E., Thaariq, Z. Z. A., & Muthmainnah, M. (2018). K-Means clustering of student perceptions on project-based learning model application. In 2018 4th International Conference on Education and Technology (ICET) (pp. 9-12). IEEE. Lévy-Leboyer, C., & Prieto, J. M. (1997). Gestión de las competencias. Barcelona: Gestión. Lévy-Leboyec, C. (2003). Gestión de las competencias: cómo analizarlas, cómo evaluarlas, cómo desarrollarlas. Barcelona: Gestión 2000. López, E. (2016). En torno al concepto de competencia: un análisis de fuentes. Vol. 20, núm. 1, enero-abril, 2016, pp. 311-322 Universidad de Granada, Granada, España Machi, L. A., & McEvoy, B. T. (2021). The literature review: Six steps to success. Malerba, F., & McKelvey, M. (2020). Knowledge-intensive innovative entrepreneurship integrating Schumpeter, evolutionary economics, and innovation systems. Smal Business Economics, 54(2), 503-522. Marengo, L., Mastrogiorgio, M. & Romagnoli, M. (2021). Sensitivity, Innovation Attitudes, and Perseverance as the Strategic Foundations of Exaptation Marquina, P., Alvarez, C., Guevara, D., & Guevara, R. (2013). Revisión de literatura esquema Documento de trabajo con esquema para el desarrol o del Trabajo de Investigación Final. Diss. Tesis, modalidad Revisión de Literatura, en CENTRUM católica. Lima, Perú, 2013. Maydón, L. (2016). Alto impacto, tecnología y las claves del éxito en la era de la transformación digital. Mertens, L. (1996). Competencia laboral: sistemas, surgimiento y modelos. Micheli, T., & Valle, J. (2018). La brecha digital y la importancia de las tecnologías de la información y la comunicación en las economías regionales de México. Realidad, datos y espacio revista internacional de estadística y geografía. Vol. 9, Núm. 2 mayo-agosto, 2018. Moe, K., Shotaro, B., Qi, W. & Masaru, Y. (2021). Artificial Intelligence in scientific activities: an organizational analysis of research teams Mohri, M., Rostamizadeh, A., & Talwalkar, A. (2018). Foundations of machine learning. MIT Nadine, J. (1998). Compétences et Compétitivité. París: Les Éditions dórganisation Neck, H. M., Meyer, G. D., Cohen, B., & Corbett, A. C. (2004). An entrepreneurial system view of new venture creation. Journal of small business management, 42(2), 190-208. Nugent, R., Dean, N., & Ayers, E. (2010). Skil set profile clustering: the empty K-means algorithm with automatic specification of starting cluster centers. Nwaohiri, N. M., & Nwosu, M. C. (2021). Reskil ing the library workforce for the fourth industrial revolution. In Examining the Impact of Industry 4.0 on Academic Libraries. Emerald Publishing Limited. Osuna, C., & Luna, E. (2011). Valores Éticos en la Formación Universitaria de las Áreas de Ciencias Naturales e Ingeniería y Tecnología, en el Contexto de la Sociedad del Conocimiento. Formación universitaria, 4(5), 29-36. Palacios, R.; Cruz, X., & Escott, M.P. (2021) Knowledge Domain Capabilities as a Business Mindset in the Context of Digital Transformation. International Schumpeter Society Conference Pérez, C. (2010). Technological revolutions and techno-economic paradigms. Cambridge journal of economics, 34(1), 185-202. Roitter, S., Erbes, A., & Kababe, Y. (2013). Procesos de aprendizaje en el sector servicios:¿ nuevas formas de organización del trabajo?. El sistema argentino de innovación: instituciones, empresas y redes. El desafío de la creación y apropiación de conocimiento Roitter, S. (2019). Cambio tecnológico y empleo aportes conceptuales y evidencia frente a la dinámica en curso (No. 995058592602676). International Labour Organization Romero, A. (2016) El secreto de poner a las personas en el centro de la transformación digital. Ruiz, J. K., Silva, N. G., & Vanga, M. G. (2008). Ética empresarial y el desempeño laboral en Organizaciones de Alta Tecnología (OAT). Revista Venezolana de gerencia, 13(43), 417-441. Rychen, D. S., & Salganik, L. H. (Eds.). (2003). Key competencies for a successful life and wel - functioning society. Hogrefe Publishing. Sarramona, J. (2007). Las competencias profesionales del profesorado de secundaria. Estudios sobre educación, 31-40. 594 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Sasmita, N., & Kumar, R. H. (2018). Exigency of re-skiling for organization and employee’s growth. Soc. Sci, 3, 65-67. Schumpeter, J. (1942). Creative destruction. Capitalism, socialism and democracy, 825, 82-85 Schumpeter, J. (1978). Teoría del desenvolvimiento económico. México: Fondo de Cultura Económica Schwab, K. (2017). The fourth industrial revolution. Crown Business. Schwab, K. & Malleret, T. (2020). Covid-19: El Gran Reinicio Shamrat, F. J. M., Tasnim, Z., Mahmud, I., Jahan, M. N., & Nobel, N. I. (2020). Application of K-means clustering algorithm to determine the density of demand of different kinds of jobs. International Journal of Scientific & Technology Research, 9(02), 2550-2557. Shel y, Z., Burch, R. F., Tian, W., Strawderman, L., Piroli, A., & Bichey, C. (2020). Using K-means clustering to create training groups for elite American footbal student-athletes based on game demands. International Journal of Kinesiology and Sports Science, 8(2), 47-63. Sierra, S. (2017). Experimentos científicos infantiles como medios para generar actitudes y aptitudes investigativas en los niños de 5 años del grado transición del colegio Liceo Pedagógico la Dicha del Saber de la ciudad de Bucaramanga. Spencer, L., & Spencer, S. (1993). Competency at work: models for superior performance. New York: Wiley and Sons. Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic management journal, 18(7), 509-533. Tejada, J. (1999). Acerca de las competencias profesionales. Herramientas. Thu, M. K., Beppu, S., Wang, Q., Yarime, M., & Shibayama, S. (2021). Artificial Intel igence in Scientific Activities: An Organizational Analysis of Research Teams. Valiente, C. M. (2017). La creatividad, una revisión científica. Revista científica de Arquitectura y Urbanismo, 38(2), 53-62. Vijayarani, S., Ilamathi, M. J., & Nithya, M. (2015). Preprocessing techniques for text mining-an overview. International Journal of Computer Science & Communication Networks, 5(1), 7-16. Zhang, Y., Chen, M., & Liu, L. (2015, September). A review on text mining. In 2015 6th IEEE International Conference on Software Engineering and Service Science (ICSESS) (pp. 681-685). IEEE. Zhou, Z. H. (2021). Machine learning. Springer Nature X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 595 Appendix Questions from the applied questionnaire, as well as classification of indicators - simplified variables. Variables Information 1. Attitudes Author's definition /Autor: Traits, patterns, or habitual behaviors (positive attitudes vs negative attitudes). Spencer & Spencer (1993) Questions: 1. Does the culture of your company influence the attitudes of your employees towards the tasks they perform? a) It completely influences. b) It partial y influences. c) It influences very little. d) It doesn't influence. 2. Which positive attitudes do workers develop that foster innovation? Mention up to 3 3. Which negative attitudes have been observed in the development of work tasks that hinder innovation? Mention up to 3 Concepts: a) Attitude: It is defined as a constant tendency or disposition to perceive or react in a specific way. It can be controlled and changed. (Martinez, 2004) b) Positive attitude: Actions that, based on the analysis of reality, focus attention on achieving objectives. (Monhanty, 2009; Siomkos et al., 2001; Lowe et al., 2007) c) Negative attitude: These constitute obstacles to fulfil ing and achieving goals. They are evaluative predispositions associated with a specific object or situation that generate reactions and behaviors that hinder the acquisition of knowledge. (Antonak & Livneh, 1988; Romero-Bojórquez, 2014) 2. Strategic Author's definition /Autor: Cooperation links between different sectors Alliances with the purpose of addressing knowledge gaps and promoting innovation (endogenous competencies and exogenous competencies). Lichtenthaler (2016). Participation in innovation, improvement of capabilities through involvement in current technology. Bel & Pavitt (1995) Question: 4. Do the al iances that your company develops with other organizations address the knowledge gaps and internal competency deficiencies of your employees? a) Completely address b) Partially address c) Address very little 596 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Variables Information d) Do not address 3. Capabilities Author's definition /Autor: Professional equipment and necessary resources to carry out an activity. Tejada (1999) The potential that a person has to perform a specific task. Schkolnik et al. (2005) Endures over time in the face of market changes and adapts to new environmental requirements (Dynamic Capability). Teece et al. (1997) Integrated knowledge that enables companies to develop. Teece & Linden (2017) Questions: 5. Does your company have technological infrastructure and knowledge that al ows strengthening the competencies of your employees? a) Completely available b) Partially available c) Available to a very limited extent d) Not available 6. During the COVID-19 pandemic, does the company develop ad-hoc innovation strategies to respond to continuous changes and digital transformation? a) Fully develops them b) Partially develops them c) Develops them to a very limited extent d) Does not develop them Concepts: a) Ad-hoc innovation Types of innovative practices that arise from economic situations or technological circumstances rather than as a result of long-term strategic decisions. These practices are developed by leveraging the internal knowledge of the company's employees. (Bruns, 2012; Shaugnessy, 2010) 4. Underlying Author's definition /Autor: characteristics Individual characteristics, aptitudes, personality traits. Representation of individuals based on their natural abilities. Lévy-Leboyec (2003) Question: 7. Does your company leverage the natural aptitudes of its employees for adapting to digital transformation? a) Ful y leverages them b) Partially leverages them c) Leverages them to a very limited extent d) Does not leverage them Concepts: a) Aptitude. Attributes of relative permanence in an individual that enable the development of certain tasks. Capacities that facilitate the learning of a specific X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 597 Variables Information task or position. Potential capacity to perform through training and knowledge (Lévy-Leboyec, 1997) b) Natural aptitudes. Innate skills acquired from childhood, stimulate specific processes for comprehensive development, and are acquired natural y (Sierra, 2017) 5. Knowledge Author's definition /Autor: Transformative learning, positively influenced by the prior knowledge base of potential users. This learning accumulates throughout their performance in the organization (Cohen & Levinthal, 1990) Tacit knowledge, as wel as the accumulation of experience through "learning by doing," appropriation of know-how. (Bell & Pavitt, 1995) Key competencies (knowledge bases, training, and education) (Mertens, 1996) Theoretical competencies (acquired knowledge) (Nadine, 1998) Learning that results in increased knowledge and technological absorption capacity (Morcillo et al., 2000) Capacity for understanding and assimilating knowledge related to flexibility and adaptability, prior learning, reduction of ambiguity in new knowledge, connection with existing knowledge, and application of external knowledge. Diffusion and integration into the activities of the company (Lane, 2001) Cognitive competence (knowing, knowing how). (Gal ego-Arrufat, 2007) It is a necessary resource for the entrepreneurial ecosystem to flow among agents. Stam & Van de Ven (2019) Questions: 8. Does the interaction with users or clients impact the development of your employees' competencies? a) It impacts completely. b) It impacts partial y. c) It impacts very little. d) It does not impact. 9. Has the accumulated knowledge and experience of your employees been a transformative component in the adaptation of the company? a) It has been completely. b) It has been partial y. c) It has been very little. d) It has not been. 10. What are the most important competencies that the company develops to adapt to digital transformation? Please mention at least three. 11. Does the company develop strategies for creativity and the development of new ideas among its employees? a) It develops completely. b) It develops partially. c) It develops very little. d) It does not develop. 598 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Variables Information 12. Is the accumulated knowledge of the workers used to enhance the innovation ecosystem of the business? a) It is fully utilized. b) It is partially utilized. c) It is minimally utilized. d) It is not utilized. Concepts: a) Competencies. A characteristic that is part of one's personality and is related to a standard of effectiveness and superior performance in a situation. It includes knowledge and understanding that can be transferred to other individuals or organizations (Spencer & Spencer, 1993; González & Wagenaar, 2003; Rychen & Salganik, 2003; Gallardo et al., 2015) b) Knowledge dissemination. The dissemination of knowledge can be defined as the adaptations and applications of documented knowledge in scientific publications and patents. The more social y connected it is to the creator of knowledge, the more quickly it spreads. Knowledge dissemination largely occurs through personal contact (Chen & Hicks, 2004; Breschi & Nalerba, 2005) c) Innovation ecosystem. They are formed by the proper combination of organizations and individuals that are optimal for innovation and economic growth. Innovation ecosystems develop through entrepreneurship and are intertwined with political, economic, social, and cultural aspects of each region. (Arenal et al., 2018; Malecki & Spigel, 2017; Smith, 2018; Fernández-Esquinas et al., 2017; Witte et al., 2017; Rychen & Zimmermann, 2008) 6. Beliefs and Author's definition /Autor: Values What is social y important for the individual (values and social roles). Spencer & Spencer (1993) Transparency, internalization of business goals, clarity, personal awareness (Social exchange) (Mertens, 1996) Knowing how to live in a context and how others are perceived (González & Wagenaar, 2003) Ethical competence (Gal ego-Arrufat, 2007) Questions: 13. Which values and social roles of the workers are crucial for the proper development of their job competencies? Please mention at least three. 14. Do the workers have a clear and precise understanding of the objectives pursued by the company through social exchange? a) They have a complete understanding b) They have a partial understanding c) They have a very limited understanding d) They do not have an understanding Concepts: a) Values and social roles. X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 599 Variables Information Invisible variable, it is the image that a person projects reflecting what the individual believes to be important to do or be. Spencer & Spencer (1993) b) Social exchange. Refers to the relationship that determines the quantity and quality of effort an individual will exert in an organization. It has critical factors such as the work environment, transparency, and the internalization of goals, clarity and awareness of individual expected outcomes. (Christis, 1988; Mertens, 1996) 7. Author's definition /Autor: Entrepreneurship Distinctive competencies provide potential and contribute to the consumer's and Innovation perception of benefits towards the final product. They are difficult to imitate and correspond to the ability to anticipate market trends (Prahalad & Hamel, 1990) Competitive advantages and difficult reproducibility, combined with networking outcomes, are considered within the innovation capability approach (Cohen & Levinthal, 1990) Knowledge competencies through the combination of theory, practice, and social skills to coordinate actions, seek new solutions, and generate creativity and innovation (Mertens, 1996) Integration of inventive and creative competence with the internal and external knowledge of an organization (Nadine, 1998) Improvement in companies occurs in proportion to the number of exaptations. These occur when agents create a function that is required by the environment in a subsequent period. Sensitivity is the ability of an agent to sense the evaluations of the external environment associated with the functions implemented in their artifact. This sensitivity will increase the number of innovations, and the appropriate level of perseverance wil enable the identification of possibilities (Marengo et al., 2021) "Learning by doing" and "Learning by using" promote the creation of radical innovations in the context of technological applications in the company (Grashof & Kopka, 2021) Foreign technology can be adapted and transformed according to new needs, as the process of "learning by doing" can help improve the technology itself. Imitation and reverse engineering are tools of innovation, especially in emerging economies (Benini, 2016) Questions: 15. Is the improvement of the perception of the company's products influenced by the distinctive competencies of its workers? a) It is completely influenced. b) It is influenced partially. c) It is influenced very little. d) It is not influenced. 16. Does the company integrate its inventive and creative competencies with its innovation strategy? 600 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Variables Information a) It integrates them ful y. b) It integrates them partial y. c) It integrates them very little. d) It does not integrate them. 17. What level of sensitivity do you develop regarding the market in which your company competes? a) Very high b) High c) Medium d) Low e) Very low 18. If you respond from medium to very high, does the innovation of your company increase the sensitivity you develop regarding the market in which you compete? a) It increases it completely b) It increases it partially c) It increases it very little d) It does not increase it 19. What level of perseverance do you develop regarding the demand for your company's product? a) Very high b) High c) Medium d) Low e) Very low 20. If you respond from medium to very high, does the level of perseverance you develop al ow you to identify greater possibilities for innovation in your company? a) It allows it completely b) It al ows it partially c) It allows it very little d) It does not al ow it 21. Do you apply the "learning by doing" approach as a potential for innovation in the company? a) It is applied completely b) It is applied partially c) It is applied very little d) It is not applied 22. If the "learning by doing" approach is applied, does this approach favor incremental innovations in the company? a) It favors them completely b) It favors them partial y c) It favors them very little X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 601 Variables Information d) It does not favor them 23. Does the company engage in reverse engineering processes to generate new products or services? a) It engages in them completely b) It engages in them partial y c) It engages in them very little d) It does not engage in them 24. Do you benefit from reverse engineering to improve your competencies? a) You benefit from it completely b) You benefit from it partial y c) You benefit from it very little d) You do not benefit from it Concepts: a) Sensitivity Sensitivity refers to the ability of an agent to perceive and evaluate external environmental factors associated with the functions implemented in its artifact (Marengo et al., 2021). b) Perseverance Perseverance refers to the number of attempts that agents decide to make in order to improve each detected function before moving on to the next one. It is a measure of the effort dedicated to enhancing each specific function (Marengo et al., 2021). c) Distinctive competencies Distinctive competencies are capabilities that are difficult for competitors to imitate, and when actively used and introduced into the market, they become a competitive advantage (Mertens, 1996). d) Learning by doing Learning through personal participation in the experience. Knowledge is discovered by the individual, having a meaning that impacts their behavior. It includes the freedom to establish personal goals and actively pursue them within a given framework (Ord, 2012). e) Learning by using Learning by customers or end-users through the use of a product or service, which helps a company invent a product or service aligned with their specific needs. f) Incremental innovations A set of smal improvements made to existing products or services within an organization. These low-cost improvements typical y help further differentiate a company from its competitors while building upon current offerings (Freeman et al., 1982). g) Reverse engineering It involves the analysis of a product or service available in the market in order to understand in detail its design, construction, and operation (Acosta, 2013). 602 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Variables Information 8. Effectiveness Author's definition /Autor: standard The effectiveness standard refers to a pre-defined criterion that predicts when an individual wil perform a task in an improved manner compared to the standard. (Spencer & Spencer, 1993) Question: 25. ¿Does the company develop effectiveness standards for workers' activities to evaluate the development of their competencies? a) It develops them completely. b) It develops them partial y. c) It develops them very little. d) It does not develop them. 9. Network Author's definition /Autor: structures Direct connections with inventors (Brusoni et al., 2021) It is composed of formal and informal networks. (Neck et al., 2004) Networks should be composed of large companies, researchers, universities, incubators or accelerators, support companies, and government (Sheriff & Muffatto, 2018) Collaboration involving diverse areas of study facilitates co-creation, enhances social competencies, and the ability to work in teams. It is suggested to participate in interdisciplinary networks, intraorganizational networks, and external organizations (Thu et al., 2021) The creation of geographical y concentrated firms, vertical y or horizontal y integrated with highly skilled local labor, has multiplier effects on the economy and competitiveness. This is what is referred to as the proximity approach. Benini (2016) Question: 26. ¿Do you participate in formal business or innovation networks related to your company's product? a) I participate completely. b) I participate partially. c) I participate very little. d) I do not participate. 27. If you participate partial y to completely, do these networks al ow you to improve your competencies? a) They al ow it completely. b) They allow it partially. c) They al ow it very little. d) They do not al ow it. 28. ¿Do you participate in informal business or innovation networks related to your company's product or business? a) I participate completely. b) I participate partial y. c) I participate very little. X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 603 Variables Information d) I do not participate. 29. If you participate partial y to completely, do these networks al ow you to improve your competencies? a) They al ow it completely. b) They allow it partially. c) They allow it very little. d) They do not al ow it. 30. ¿Are the internal networks of your company interdisciplinary in nature? a) They are completely. b) They are partial y. c) They are very little. d) They are not. 31. If participation is partial or complete, do these networks influence the development of your competencies? a) They influence it completely. b) They influence it partial y. c) They influence it very little. d) They do not influence it. 32. ¿Does the geographic proximity of the companies related to your company's product influence greater effectiveness and development of your competencies? a) It influences it completely. b) It influences it partially. c) It influences it very little. d) It does not influence it. Concepts: a) Formal networks. Formed by universities, government, support services, sources of capital, talent groups, and large corporations. (Neck et al., 2004) b) Informal networks. Composed of friends, family, and informal relationships. (Neck et al., 2004) c) Interdisciplinary knowledge. Composed of various specialized areas with experts from different backgrounds. It has task allocation and a clear team structure. (Moe et al., 2021) 10. Strategic Author's definition /Autor: architecture Identifies and commits to unify technical y and productively the strategic business units to generate competitive advantage. Prahalad & Hamel (1990) Complex strategies that are difficult to imitate by competitors due to the combined and integrated capabilities, strategic elements, and knowledge architecture. Brusoni et al. (2021) Question: 33. ¿Does the development of your skil s relate to the objectives of the strategic business units of the company? a) It relates completely 604 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Variables Information b) It relates partial y c) It relates very little d) It does not relate 11. Experience Author's definition /Autor: Understood as occupational trajectories. Formal and informal work as an entrepreneur, business, or hybrid. Stable and/or temporary work. Mertens (1996) Understood as core competencies. It refers to technological experiences related to products and processes and the organizational capacity to effectively deploy that expertise. Coombs (1996) Tacit knowledge is related to skills. It is personal and not organized. Morcillo et al. (2000) Conceived as situated learning, constructed with reality, knowledge, and experiential values and competencies. Rychen & Salganik (2003) Failures in business al ow for the understanding of behavior and drive knowledge. Mason & Harrison (2006) Start-ups benefit from the accumulated knowledge and prior experiences of founders, which have a lasting effect on learning capabilities and performance processes. Bahoo et al. (2021) Questions: 34. Does it develop informal activities to support the development of your company's business? a) Fully develops b) Partially develops c) Develops very little d) Does not develop 35. Do informal activities influence the development of your skil s within the company? a) Influences completely b) Influences partial y c) Influences very little d) Does not influence 36. How do business failures influence the development and strengthening of your skills? a) Influences very positively b) Influences positively c) Influences negatively d) Influences very negatively 37. Has the founders' knowledge of the company benefited the development of your innovation skil s? a) Has benefited completely b) Has benefited partial y c) Has benefited very little X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 605 Variables Information d) Has not benefited Concepts: a) Informal activities. They lack contractual agreements, there is little separation between work and capital, and they are based on casual employment. Usually, they involve a lack of skills and low productivity. 12. Skills Author's definition /Autor: What the individual knows how to do well. Spencer & Spencer (1993) Required qualifications: Academic skills, personal development, and teamwork: Mertens (1996) Practical competence: Operational actions. Nadine (1998) Functional competence (knowing how to do). Gal ego-Arrufat (2007) Ability to perform a task in a specific context. OECD (2012, 2018, 2019) Being capable, being skil ful, being able to easily perform something. González & Wagenaar (2003) They are understood as cohesive and structured skil s that enable col aboration and the generation of integrated knowledge capabilities. Brusoni et al. (2021) Within the framework of technological change, Artificial Intel igence has effects on skills, requiring complementary skills, increasing the complexity of learning skills, as well as the need for interactive and adaptive skills. Holm & Lorennz (2021) The potential of artificial intelligence in specific work activities is driven by technology. Occupations in less advanced economies demand different skil s compared to more advanced economies. Carbonero et al. (2021) Questions: 38. Do technological changes and digital transformation in the AI sector require complementary skil s beyond the ones you already possess? a) Completely requires them b) Partially requires them c) Requires them very little d) Does not require them 39. Do the complementary skil s you possess influence the effectiveness of innovation and adaptability strategies in your company? a) Completely requires them b) Partially requires them c) Requires them very little d) Does not require them 40. Do technological changes and digital transformation in the AI sector require interactive and adaptable skills? a) Completely requires them b) Partially requires them c) Requires them very little 606 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Variables Information d) Does not require them 41. Are your tasks related to the occupation or position you hold within the company? a) Completely related b) Partially related c) Related very little d) Not related Concepts: a) Complementary skills Supplementary skil s commonly associated with higher demand such as customer service and interpersonal skills, judgment and decision making, technology design, ideation fluency, science, and operations analysis. (Bakhshi et al., 2017) b) Interactive and adaptability skills Social interaction and judgment tend to set the pace of work. (Holm & Lorennz, 2021) c) Occupation Occupation is defined as those meaningful activities in which a person engages in on a daily basis and can be formalized according to the position they hold. 13. Productivity Author's definition /Autor: Within the context of technological change, automation is the key factor in the polarization of skill demand (high skill requirement or low skill requirement) and the increase in wage inequality. Automation will require premium skills from workers, shifting from routine tasks to non-routine tasks. The productivity of a task is subject to the skill of the individual. The time spent on a task determines the learning and skill acquisition time. (Bordot, 2021) Employees with skills related to artificial intelligence obtain better job positions and higher incomes. (Duch-Brown et al., 2021) Questions: 42. What percentage of your time do you spend on routine tasks during a week? a) 75% to 100% b) 50% to 75% c) 25% to 50% d) 0% to 25% 43. Are skills, competencies, or capabilities the ones that impact productivity and innovation capacity of the company? a) Skills, 75% to 100% b) Capabilities, 75% to 100% c) Competencies, 75% to 100% X. Cruz, R. Palacios: Towards a New Approach to Competences in Digital Transformation 607 Variables Information Concepts: a) Skill Refers to the acquired ability of an individual to apply their knowledge and habits in a theoretical or practical activity. (Iglesias et al., 2008) b) Capability An embedded and non-transferable resource within a company or individual, aimed at productivity. They represent cognitive, affective, and psychomotor conditions that enable learning and task development. (Makadok, 2001) c) Routine tasks Tasks that follow well-defined procedures and can be easily performed by some form of algorithm. They involve clear and repetitive activities. (Frey & Osborne, 2013) d) Non-routine tasks Tasks that are less likely to be automated. They involve different actions that vary over time and require visual, socioemotional, and motor processing capabilities that cannot be programmatical y described. 608 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY THE CROSS-CHANNEL EFFECTS OF IN-STORE CUSTOMER EXPERIENCE IN THE CASE OF OMNICHANNEL FASHION RETAILING IN FINLAND MARKUS MAKKONEN,1 LAURI FRANK,1 TIINA PAANANEN,1 MATILDA HOLKKOLA,1 TIINA KEMPPAINEN2 1 University of Jyvaskyla, Faculty of Information Technology, Jyvaskyla, Finland markus.v.makkonen@jyu.fi, lauri.frank@jyu.fi, ti na.e.paananen@jyu.fi, matilda.i.holkkola@jyu.fi 2 University of Jyvaskyla, School of Business and Economics, Jyvaskyla, Finland ti na.j.kemppainen@jyu.fi Although omnichannel retailing has emerged as a popular research topic in academic research, there are stil gaps in our understanding of this phenomenon. One such gap concerns Keywords: omnichannel customer experience and particularly the cross- in-store channel or spil over effects of how customer experience in one customer channel may affect customer behaviour not only in that specific experience, customer channel but also in the other channels of the same retailer. In this satisfaction, study, we aim to address this gap by examining how customer brick-and-mortar store experience in the offline channel affects customer behaviour in visit both the offline and the online channel, more specifical y how intention, in-store customer experience affects both brick-and-mortar store online store and online store visit intention. The study is conducted by using visit survey data from the customers of a Finnish fashion retailer on intention, their recent brick-and-mortar store visit. We find that in-store omnichannel fashion customer experience does indeed positively affect not only brick- retailing, and-mortar store but also online store visit intention. Finland DOI https://doi.org/10.18690/um.fov.4.2023.38 ISBN 978-961-286-751-5 610 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Today, more and more retailers are engaging in so-cal ed omnichannel retailing, which refers to the integration of multiple retail channels or other touchpoints between retailers, brands, and consumers in order to provide a single seamless and consistent customer experience (Rigby, 2011; Brynjolfsson et al., 2013; Verhoef et al., 2015; Akter et al., 2021). Thus, it is no surprise that omnichannel retailing has emerged as a popular research topic also in academic research (e.g., Cai & Lo, 2020; Gerea et al., 2021; Mishra et al., 2021; Asmare & Zewdie, 2022; Lopes et al., 2022; Nguyen et al., 2022; Timoumi et al., 2022). However, there are still substantial gaps in our present understanding of this phenomenon. One such gap concerns omnichannel customer experience and particularly the cross-channel or spil over effects of how customer experience in one channel may affect customer behaviour not only in that specific channel but also in the other channels of the same retailer (e.g., Gerea et al., 2021; Mishra et al., 2021; Timoumi et al., 2022). Such effects have not been examined in any prior study that we are aware of, although some prior studies have focused on related topics, such as the effects of online channel service failures on offline channel customer loyalty (Wang & Zhang, 2018) and the effects of online channel customer satisfaction on offline channel adoption (Teng et al., 2023). However, none of them has focused on customer experience from a holistic perspective, and most of them have focused only on the online-to-offline instead of offline-to-online effects. In this study, we aim to address the aforementioned research gap by examining how customer experience in the of line channel af ects customer behaviour in both the of line and the online channel, more specifical y how in-store customer experience af ects both brick-and-mortar store and online store visit intention. As with many other prior studies on omnichannel retailing (e.g., Lynch & Barnes, 2020; Truong, 2021; Riaz et al., 2022), the study is conducted in the case context of fashion retailing and in co-operation with a Finnish fashion retailer by first surveying its customers on their recent brick-and-mortar store visit and then analysing this collected data with partial least squares structural equation modelling (PLS-SEM). After this introductory section, we briefly present the research model of the study in Section 2. The methodology and results of the study are reported in Sections 3 and 4, of which the results are discussed in more detail in Section 5. Finally, we conclude M. Makkonen, L.Frank, T. Paananen, M. Holkkola, T. Kemppainen: The Cross-Channel Ef ects of In-Store Customer Experience in the Case of Omnichannel Fashion Retailing in Finland 611 the paper with a brief discussion of the limitations of the study and some potential paths for future research in Section 6. 2 Research Model We base our research model on the in-store customer experience (ISCX) scale by Bustamante and Rubio (2017), which is a holistic instrument for measuring customer experience in brick-and-mortar stores and has successful y been applied to various retail contexts in prior research (e.g., Bustamante & Rubio, 2017; Happ et al., 2021). The ISCX scale is based on the definition of customer experience by Verhoef et al. (2009) as a construct that “is holistic in nature and involves the customer’s cognitive, affective, emotional, social, and physical responses to the retailer”. As such, overall in-store customer experience is hypothesised to be constituted of four components (Bustamante & Rubio, 2017): cognitive experience (i.e., the capability of marketing stimuli to make a customer think and reflect, arouse curiosity, awaken creativity, inspire, etc.), affective experience (i.e., the capability of marketing stimuli to provoke emotions in a customer), social experience (i.e., the relationship a customer establishes with the store as a social system by interacting with the employees or other customers), and physical experience (i.e., the physiological responses of a customer in his or her interaction with the environment). However, when applying the ISCX scale to our research model, we make two modifications. First, of its four components, we omit social and physical experience and focus only on cognitive and affective experience, which have traditional y been found as the most important components of customer experience in prior research (Alan et al., 2016) and were also found as the most important components of overal in-store customer experience in the study by Bustamante and Rubio (2017). Second, we decompose affective experience, which in the ISCX scale focuses only on positive affective experience, into two components: positive affective experience and negative affective experience. This is in line with the prevailing view of positive and negative affect as two distinct affective dimensions rather than as bipolar endpoints of a single affective dimension (e.g., Watson & Tellgren, 1985; Watson et al., 1988). Of these three components, in line with the study by Bustamante and Rubio (2017), we hypothesise cognitive experience and positive affective experience to have a positive effect on overal in-store customer experience and negative affective experience to have a negative effect on overal in-store customer experience. 612 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Finally, as already mentioned in the introduction, we add brick-and-mortar (B&M) store and online store visit intention as outcomes of overal in-store customer experience in the research model. The decision to focus on these two constructs is based on the fact that store (re)visit intention has traditional y been considered one of the main manifestations of store loyalty (Bloemer & de Ruyter, 1998), thus making it a very relevant business metric for the retailers who are running the stores. In addition, because customer satisfaction has been found to at least partial y mediate the effects of customer experience in general (e.g., Brakus et al., 2009; Klaus & Maklan, 2013) and in-store customer experience in particular (Bustamante & Rubio, 2017) on customer or store loyalty, we also add it as a mediator in the research model, hypothesising that overall in-store customer experience wil positively affect brick-and-mortar store and online store visit intention both directly and indirectly via customer satisfaction. The whole resulting research model is illustrated in Figure 1. Figure 1: Research model 3 Methodology The study was conducted in co-operation with a Finnish fashion company, which wil be referred to as company X in the remainder of the paper. The company has its own clothing brand, a factory, and multiple brick-and-mortar stores around Finland. However, in this study, we focused on only one of those brick-and-mortar stores, which was the factory outlet. In addition, the company has an online store. The brick-and-mortar stores and the online store are closely integrated (e.g., they are branded consistently and there is the option to pick up the purchases made in the online store from the brick-and-mortar stores), thus making the company a suitable case company for this study on omnichannel retailing. M. Makkonen, L.Frank, T. Paananen, M. Holkkola, T. Kemppainen: The Cross-Channel Ef ects of In-Store Customer Experience in the Case of Omnichannel Fashion Retailing in Finland 613 We colected the data for the study between May 2022 and September 2022 by using both a pen-and-paper survey and an online survey conducted with the LimeSurvey service. Both the surveys were anonymous, had identical questionnaires, and were targeted at the customers of company X who were currently visiting or had recently visited its brick-and-mortar store. The surveys were promoted in two ways. First, we placed promotional boards in the brick-and-mortar store, in which customers were invited to respond to the survey either by using the pen-and-paper questionnaires at the store or by accessing the online questionnaire via a QR code or web address that was printed on the boards. Second, the salespeople at the brick-and-mortar store placed promotional leaflets into the shopping bags of people who had made a purchase at the store, in which customers were invited to respond to the survey by accessing the online questionnaire via a QR code or web address that was printed on the leaflets. In both the promotional boards and the promotional leaflets, customers were also informed about the opportunity to take part in a prize drawing of one 50 € gift card after completing the survey. This was considered a suitable incentive in terms of not causing any bias to the results but stil promoting the response rate. In the survey, customers were first inquired about their general background information and then more specifical y about their visit to the brick-and-mortar store, which also contained the items for measuring the seven constructs in the research model. All the constructs were measured reflectively by multiple items: cognitive experience, positive affective experience, negative affective experience, overall in-store customer experience, and customer satisfaction with three items each and brick-and-mortar store and online store visit intention with two items each. The wordings of these 19 items are reported in Appendix A. The items for measuring cognitive experience, positive affective experience, and overal in-store customer experience were adapted from the study by Bustamante and Rubio (2017), whereas the items for measuring negative affective experience were adapted from the study by Richins (1997) and the items for measuring brick-and-mortar store and online store visit intention were adapted from the behavioural intention measures by Fishbein and Ajzen (2010). The measurement scale for all these items was the traditional five-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, and 5 = strongly agree). In turn, the items measuring customer satisfaction were adapted from the American Customer Satisfaction Index (ACSI) by Fornell et al. (1996) and the Extended Performance Satisfaction Index 614 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY (EPSI) by Selivanova et al. (2002), which are both based on the Swedish Customer Satisfaction Barometer (SCSB) by Fornel (1992). The measurement scale for these items was a modified five-point Likert scale (1 = extremely dissatisfied, 2 = dissatisfied, 3 = neither satisfied nor dissatisfied, 4 = satisfied, and 5 = extremely satisfied). Finally, we also used five control variables in the study: gender, age, information on whether the respondent had made a purchase during the visit, and information about the average visiting frequency to the brick-and-mortar store and the online store of company X. Of these, the first three variables were used to control al the endogenous constructs in the research model (i.e., overall in-store customer experience, customer satisfaction, brick-and-mortar store visit intention, and online store visit intention), whereas brick-and-mortar store visiting frequency was used to control only brick-and-mortar store visit intention and online store visiting frequency was used to control only online store visit intention. In order to avoid forced responses, responding to all the items in the survey was voluntary, and not responding to a particular item resulted in a missing value. Due to the limited sample size, the col ected data was analysed with variance-based structural equation modelling (VB-SEM), more specifically partial least squares structural equation modelling (PLS-SEM) by using the SmartPLS 4.0.8.8 software by Ringle et al. (2022). When conducting PLS-SEM, we fol owed the recent guidelines by Hair et al. (2022). For example, we used mode A as the indicator weighting mode, path weighting as the weighting scheme, +1 as the initial weights, and < 10-7 as the stop criterion in model estimation, whereas the statistical significance of the model estimates was tested by using bootstrapping with 10,000 subsamples. As the threshold for statistical significance, we used p < 0.05, but we also took into account the results that were statistically almost significant at p < 0.10 due to the limited sample size. Because of their smal percentage, the potential missing values were handled simply by using mean replacement. 4 Results In total, we received 101 valid responses to the survey, of which 40 (39.6%) originated from the pen-and-paper survey and 61 (60.4%) originated from the online survey. The descriptive statistics of the whole sample in terms of the gender, age, and socioeconomic status of the respondents are reported in Table 1. In addition, Table 1 reports whether the respondents had made a purchase during their visit as M. Makkonen, L.Frank, T. Paananen, M. Holkkola, T. Kemppainen: The Cross-Channel Ef ects of In-Store Customer Experience in the Case of Omnichannel Fashion Retailing in Finland 615 wel as their average brick-and-mortar store and online store visiting frequency. As can be seen, most of the respondents (74.3%) were women, which was expected when considering our case context of fashion retailing. The age of the respondents ranged from 15 to 72 years, with a mean of 39.9 years and a standard deviation of 14.6 years. In terms of socioeconomic status, most of the respondents were either employees (72.3%) or students (22.8%). Most of the respondents (76.2%) had also made a purchase during their visit and almost all of them (96.0%) had visited the brick-and-mortar store prior to their present visit. In contrast, the online store was visited slightly less frequently, and there was also a substantial share of respondents (15.8%) who had never visited it. Table 1: Descriptive statistics of the sample (N = 101) N % N % Gender Made a purchase during the visit Man 26 25.7 Yes 77 76.2 Woman 75 74.3 No 20 19.8 Age No response 4 4.0 15–29 years 26 25.7 B&M store visiting frequency 30–39 years 29 28.7 At least monthly 20 19.8 40–49 years 20 19.8 At least yearly 48 47.5 50–59 years 11 10.9 Less frequently than yearly 29 28.7 60 years or over 15 14.9 Has never visited 2 2.0 Socioeconomic status No response 2 2.0 Student 23 22.8 Online store visiting frequency Employee 73 72.3 At least monthly 21 20.8 Self-employed 5 5.0 At least yearly 32 31.7 Unemployed or unable to work 3 3.0 Less frequently than yearly 29 28.7 Pensioner 7 6.9 Has never visited 16 15.8 Stay-at-home parent 1 1.0 No response 3 3.0 In the following three subsections, we first assess the estimated model in terms of the reliability and validity of its constructs and indicators. Finally, we report the model estimates. 616 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 4.1 Construct Reliability and Validity Construct reliability was assessed from the perspective of internal consistency by using the composite reliability (CR) of the constructs (Fornell & Larcker, 1981), which is commonly expected to be at least 0.7 (Nunally & Bernstein, 1994). The CR of each construct is reported in the first column of Table 2, showing that al the constructs met this criterion. In turn, construct validity was assessed from the perspectives of convergent and discriminant validity by using the two criteria by Fornell and Larcker (1981). They are both based on the average variance extracted (AVE) of the constructs, which is the average proportion of variance that a construct explains in its indicators. The first criterion concerning convergent validity expects each construct to have an AVE of at least 0.5. This means that, on average, each construct should explain at least half of the variance in its indicators. The AVE of each construct is reported in the second column of Table 2, showing that al the constructs met this criterion. The second criterion concerning discriminant validity expects each construct to have a square root of AVE that is at least equal to its absolute correlations with the other constructs in the model. This means that, on average, each construct should share at least an equal proportion of variance with its indicators to what it shares with the other constructs. The square root of AVE of each construct (on-diagonal) and the correlations between all the constructs in the model (off-diagonal) are reported in the remaining columns of Table 2, showing that this criterion was also met by al the constructs. Additional support for discriminant validity was also provided by the heterotrait-monotrait (HTMT) ratios of the aforementioned correlations, which al met the criterion of being less than 0.90 (Henseler et al., 2015). Table 2: Construct-level statistics Square roots of AVE and the correlations between the CR AVE constructs CE PAE NAE ISCX CS BMVI OSVI CE 0.870 0.690 0.831 PAE 0.871 0.693 0.681 0.832 NAE 0.884 0.718 -0.120 -0.362 0.847 ISCX 0.921 0.795 0.347 0.516 -0.397 0.892 CS 0.929 0.815 0.386 0.611 -0.517 0.616 0.903 BMVI 0.935 0.878 0.274 0.322 -0.234 0.695 0.400 0.937 OSVI 0.928 0.865 0.245 0.298 -0.085 0.376 0.385 0.459 0.930 M. Makkonen, L.Frank, T. Paananen, M. Holkkola, T. Kemppainen: The Cross-Channel Ef ects of In-Store Customer Experience in the Case of Omnichannel Fashion Retailing in Finland 617 4.2 Indicator Reliability and Validity Indicator reliability and validity were assessed by using the standardised loadings of the indicators, which are reported in Appendix A together with the means and standard deviations (SD) of the indicator scores as wel as the percentages of missing values. In the typical case of each indicator loading on only one construct, the standardised loading of each indicator is commonly expected to be statistical y significant and at least 0.707 (Fornell & Larcker, 1981). This is equivalent to the standardised residual of each indicator being at least 0.5, meaning that at least half of the variance in each indicator is explained by the construct on which it loads. As can be seen, al the indicators met this criterion. 4.3 Model Estimates The results of model estimation in terms of the standardised effect sizes and their statistical significance as well as the proportions of explained variance (R2) are reported in Figure 2 (and in Appendix B for the effects of the control variables). As can be seen, positive affective experience and negative affective experience were both found to have a statistically significant effect on overall in-store customer experience, with the effect of positive affective experience being positive and the effect of negative affective experience being negative. In contrast, the effect of cognitive experience on overal in-store customer experience was found to be statistically not significant. In turn, overall in-store customer experience was found to have a statistically significant effect on customer satisfaction, a statistically significant direct effect on brick-and-mortar store visit intention, and a statistically almost significant direct effect on online store visit intention. These effects were al positive, and especially the effect of overall in-store customer experience on brick-and-mortar store visit intention was found to be very strong. Final y, customer satisfaction was found to have a statistical y not significant effect on brick-and-mortar store visit intention and a statistically almost significant and positive effect on online store visit intention. 618 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Figure 2: Model estimates (*** = p < 0.001, ** = p < 0.01, * = p < 0.05, (*) = p < 0.10) In addition to the aforementioned direct effects, we also examined the indirect effects of overal in-store customer experience on brick-and-mortar store visit intention and online store visit intention via customer satisfaction as wel as the total effects of overal in-store customer experience on brick-and-mortar store visit intention and online store visit intention that take into account both the direct and the indirect effects. Of these, the indirect effects of overal in-store customer experience on both brick-and-mortar store visit intention (-0.035) and online store visit intention (0.120) were found to be statistically not significant, whereas the total effect of overal in-store customer experience on both brick-and-mortar store visit intention (0.682***) and online store visit intention (0.328***) were found to be statistically significant and positive. The effects of the control variables were found to be mostly statistical y not significant. The exceptions were the statistical y significant and positive effect of having made a purchase during the visit on customer satisfaction, the statistically almost significant and positive effect of past brick-and-mortar store visiting frequency on future brick-and-mortar store visit intention, and the statistical y significant and positive effect of past online store visiting frequency on future online store visit intention. In total, the model was able to explain 33.0% of the variance in overall in-store customer experience, 40.9% of the variance in customer satisfaction, 51.9% of the variance in brick-and-mortar store visit intention, and 45.7% of the variance in online store visit intention. The variance inflation factor (VIF) values of all the constructs and control variables were found to be less than three, thus indicating no multicollinearity issues (Hair et al., 2018). M. Makkonen, L.Frank, T. Paananen, M. Holkkola, T. Kemppainen: The Cross-Channel Ef ects of In-Store Customer Experience in the Case of Omnichannel Fashion Retailing in Finland 619 5 Discussion and Conclusion In this study, we examined the effect of in-store customer experience on brick-and-mortar store and online store visit intention while also considering the role of customer satisfaction as a mediator. All in all, we made three main findings. First and foremost, we found that in-store customer experience does indeed affect positively not only brick-and-mortar store but also online store visit intention, especial y when considering the total effect of in-store customer experience on online store visit intention that takes into account both the direct effect and the indirect effect via customer satisfaction. Thus, at least in our case context of fashion retailing, there seems to be a positive cross-channel or spil over effect for the outcomes of customer experience from offline to online channels. Second, our findings provide some support for the role of customer satisfaction as a mediator for the effect of in-store customer experience on online store visit intention because although the indirect effect of in-store customer experience on online store visit intention via customer satisfaction was found to be statistical y not significant in our sample, this effect was stil strong enough to result in the total effect of in-store customer experience on online store visit intention becoming statistically significant despite the statistical y not significant direct effect of in-store customer experience on online store visit intention. In contrast, our findings provide no support for the role of customer satisfaction as a mediator for the effect of in-store customer experience on brick-and-mortar store visit intention due to the statistically not significant and very weak indirect effect of in-store customer experience on brick-and-mortar store visit intention via customer satisfaction. In other words, in-store customer experience alone seems to determine brick-and-mortar store visit intention regardless of the resulting customer satisfaction. This finding conflicts with the study by Bustamante and Rubio (2017), who found store satisfaction to mediate the effect of in-store customer experience on store loyalty. The conflict may be explained by their different operationalisation of store loyalty, which did not focus only on store (re)visit intention but also on aspects like word-of-mouth behaviour. Third, we found overall in-store customer experience to be affected more strongly by positive than negative affective experience but not at al by cognitive experience as it is hypothesised in the study by Bustamante and Rubio (2017). Of these, the former finding is in line, for example, with the prior study by Makkonen et al. (2019), who found outcomes like customer satisfaction, repurchase intention, and recommendation intention to be affected more strongly by positive than negative 620 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY emotions in the context of online shopping. In turn, the latter finding may be explained by our more hedonic and less utilitarian case context of fashion retailing (e.g., Kemppainen et al., 2021), in which affective rather than cognitive experience is likely to be more dominant in determining overal customer experience. From a theoretical perspective, the findings of the study promote our understanding of the cross-channel or spil over effects of customer experience and its outcomes in omnichannel retailing, which is a topic that has been largely overlooked in prior research. For example, as already discussed in the introduction, although some prior studies have focused on related topics, none of them has focused on customer experience from a holistic perspective by comprehensively considering its various (e.g., cognitive and affective) components, and most of them have focused only on the online-to-offline instead of offline-to-online effects. In turn, from a practical perspective, the findings of the study highlight the importance of holistic channel and customer experience management in omnichannel retailing. Such holistic management can be considered important because the aforementioned cross- channel or spil over effects pose both substantial opportunities and substantial threats for omnichannel retailers. On one hand, the effects pose opportunities because good customer experience in one channel is likely to result in positive outcomes not only in that particular channel but in other channels as wel . Thus, it always pays off for omnichannel retailers to invest in the improvement of customer experience in al their channels because even if some channels may be seen as strategically less important, the resulting positive outcomes are likely to spill over and promote the success of also the strategical y more important channels. On the other hand, the effects pose threats because bad customer experience in one channel is likely to result in negative outcomes not only in that particular channel but in other channels as well. Thus, omnichannel retailers should never neglect customer experience in any of their channels because even if some channels may once again be seen as strategical y less important, the resulting negative outcomes are likely to spill over and spoil the success of also the strategical y more important channels. In summary, in omnichannel retailing, the customer experience of al channels always counts. M. Makkonen, L.Frank, T. Paananen, M. Holkkola, T. Kemppainen: The Cross-Channel Ef ects of In-Store Customer Experience in the Case of Omnichannel Fashion Retailing in Finland 621 6 Limitations and Future Research We see this study to have four main limitations. First, the study was conducted in the case context of fashion retailing, which may limit the generalisability of its findings to other retail contexts. Second, the sample size of the study was relatively small, although still sufficiently large for identifying several statistically significant effects between the constructs in our research model. The sample size was mainly limited by our methodological choice of focusing on the customers of only one case company and on their very recent visits, of which the latter is reflected by the fact that 85 out of the 101 respondents (84.2%) completed the survey on the same day of their visit. Thus, we believe that what we may have lost in data quantity, we have more than gained in data quality in terms of minimising recal bias and the effects of confounding factors that may have resulted from having multiple case companies in the study. Third, when measuring in-store customer experience, we focused only on cognitive and affective experience while omitting social and physical experience, which may explain our rather modest R2 of overall in-store customer experience (33.0%), although it cannot be seen to compromise our findings concerning the effects of overal in-store customer experience on its outcomes. Fourth, of the potential outcomes of in-store customer experience, we focused only on customer satisfaction as well as on visit intentions as manifestations of store loyalty. Obviously, also many other constructs could have been chosen as outcomes in our research model. We see that future research should address the aforementioned limitations by replicating the study in other case contexts than fashion retailing by using larger samples and even more holistic measures of in-store customer experience while also considering other potential outcomes of in-store customer experience, such as customer value, customer commitment, customer engagement, or wil ingness-to- pay. In addition, it is essential to examine the cross-channel or spil over effects not only from brick-and-mortar stores to online stores but also vice versa and between other commonly used channels in omnichannel retailing. 622 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY References Akter, S., Hossain, T. M. T., & Strong, C. (2021). What omnichannel realy means? Journal of Strategic Marketing, 29(7), 567–573. Alan, A. K., Kabadayi, E. T., & Yilmaz, C. (2016). Cognitive and affective constituents of the consumption experience in retail service settings: Effects on store loyalty. Service Business, 10(4), 715–735. Asmare, A., & Zewdie, S. (2022). Omnichannel retailing strategy: A systematic review. International Review of Retail, Distribution and Consumer Research, 32(1), 59–79. Bloemer, J., & de Ruyter, K. (1998). On the relationship between store image, store satisfaction and store loyalty. European Journal of Marketing, 32(5/6), 499–513. Brakus, J. J., Schmitt, B. H., & Zarantonel o, L. (2009). Brand experience: What is it? How is it measured? Does it affect loyalty? Journal of Marketing, 73(3), 52–68. Brynjolfsson, E., Hu, Y. J., & Rahman, M. S. (2013). Competing in the age of omnichannel retailing. MIT Sloan Management Review, 54(4), 23–29. Bustamante, J. C., & Rubio, N. (2017). Measuring customer experience in physical retail environments. Journal of Service Management, 28(5), 884–913. Cai, Y.-J., & Lo, C. K. Y. (2020). Omni-channel management in the new retailing era: A systematic review and future research agenda. International Journal of Production Economics, 229, 107729. Fishbein, M., & Ajzen, I. (2010). Predicting and Changing Behavior: The Reasoned Action Approach. New York, NY: Psychology Press. Fornell, C. (1992). A national customer satisfaction barometer: The Swedish experience. Journal of Marketing, 56(1), 6–21. Fornel , C., Johnson, M. D., Anderson, E. W., Cha, J., & Bryant, B. E. (1996). The American Customer Satisfaction Index: Nature, purpose, and findings. Journal of Marketing, 60(4), 7–18. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. Gerea, C., Gonzalez-Lopez, F., & Herskovic, V. (2021). Omnichannel customer experience and management: An integrative review and research agenda. Sustainability, 13(5), 2824. Hair, J. F., Jr., Black, W. C., Babin, B. J., & Anderson, R. E. (2018). Multivariate Data Analysis (8th ed.). Andover, United Kingdom: Cengage. Hair, J. F., Jr., Hult, G. T. M., Ringle, C. M., & Sarstedt, M. (2022). A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM) (3rd ed.). Thousand Oaks, CA: SAGE. Happ, E., Schol -Grissemann, U., Peters, M., & Schnitzer, M. (2021). Insights into customer experience in sports retail stores. International Journal of Sports Marketing and Sponsorship, 22(2), 312– 329. Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A new criterion for assessing discriminant validity in variance-based structural equation modeling. Journal of the Academy of Marketing Science, 43(1), 115–135. Kemppainen, T., Frank, L., Makkonen, M., & Hyvönen, O.-I. (2021). Barriers to responsible consumption in e-commerce: Evidence from fashion shoppers. In A. Pucihar, M. Kljajić Borštnar, R. Bons, H. Cripps, A. Sheombar & D. Vidmar (Eds.), Proceedings of the 34th Bled eConference (pp. 323–335). Maribor, Slovenia: University of Maribor Press. Klaus, P., & Maklan, S. (2013). Towards a better measure of customer experience. International Journal of Market Research, 55(2), 227–246. Lopes, J. M., Sousa, A., Calçada, E., & Oliveira, J. (2022). A citation and co-citation bibliometric analysis of omnichannel marketing research. Management Review Quarterly, 72(4), 1017–1050. Lynch, S., & Barnes, L. (2020). Omnichannel fashion retailing: Examining the customer decision-making journey. Journal of Fashion Marketing and Management, 24(3), 471–493. Makkonen, M., Riekkinen, J., Frank, L., & Jussila, J. (2019). The effects of positive and negative emotions during online shopping episodes on consumer satisfaction, repurchase intention, and recommendation intention. In A. Pucihar, M. Kljajić Borštnar, R. Bons, J. Seitz, H. Cripps & M. Makkonen, L.Frank, T. Paananen, M. Holkkola, T. Kemppainen: The Cross-Channel Ef ects of In-Store Customer Experience in the Case of Omnichannel Fashion Retailing in Finland 623 D. Vidmar (Eds.), Proceedings of the 32nd Bled eConference (pp. 931–953). Maribor, Slovenia: University of Maribor Press. Mishra, R., Singh, R. K., & Koles, B. (2021). Consumer decision‐making in omnichannel retailing: Literature review and future research agenda. International Journal of Consumer Studies, 45(2), 147–174. Nguyen, A., McClel and, R., Thuan, N. H., & Hoang, T. G. (2022). Omnichannel marketing: Structured review, synthesis, and future directions. International Review of Retail, Distribution and Consumer Research, 32(3), 221–265. Nunnal y, J. C., & Bernstein, I. H. (1994). Psychometric Theory (3rd ed.). New York, NY: McGraw-Hill. Riaz, H., Baig, U., Meidute-Kavaliauskiene, I., & Ahmed, H. (2022). Factors effecting omnichannel customer experience: Evidence from fashion retail. Information, 13(1), 12. Richins, M. L. (1997). Measuring emotions in the consumption experience. Journal of Consumer Research, 24(2), 127–146. Rigby, D. (2011). The future of shopping. Harvard Business Review, 89(12), 65–76. Ringle, C. M., Wende, S., & Becker, J.-M. (2022). SmartPLS 4. Oststeinbek, Germany: SmartPLS GmbH. Available at http://www.smartpls.com Selivanova, I., Hallissey, A., Letsios, A., & Eklöf, J. (2002). The EPSI rating initiative. European Quality, 9, 10–25. Teng, H., Xia, Q., Shou, J., & Zhao, J. (2023). Positive or negative spillover? The influence of online channel satisfaction on offline channel adoption. Journal of Business Research, 154, 113332. Timoumi, A., Gangwar, M., & Mantrala, M. K. (2022). Cross-channel effects of omnichannel retail marketing strategies: A review of extant data-driven research. Journal of Retailing, 98(1), 133– 151. Truong, T. H. H. (2021). The drivers of omni-channel shopping intention: A case study for fashion retailing sector in Danang, Vietnam. Journal of Asian Business and Economic Studies, 28(2), 143–159. Verhoef, P. C., Kannan, P. K., & Inman, J. J. (2015). From multi-channel retailing to omni-channel retailing: Introduction to the special issue on multi-channel retailing. Journal of Retailing, 91(2), 174–181. Verhoef, P. C., Lemon, K. N., Parasuraman, A., Roggeveen, A., Tsiros, M., & Schlesinger, L. A. (2009). Customer experience creation: Determinants, dynamics and management strategies. Journal of Retailing, 85(1), 31–41. Wang, X., & Zhang, Q. (2018). Does online service failure matter to offline customer loyalty in the integrated multi-channel context? The moderating effect of brand strength. Journal of Service Theory and Practice, 28(6), 774–806. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063–1070. Watson, D., & Tel egen, A. (1985). Toward a consensual structure of mood. Psychological Bul etin, 98(2), 219–235. 624 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Appendix A: Indicator-Level Statistics Item Wording Mean SD Missing Loading How satisfied were you with your visit to the B&M store of X… CS1 … overall? 4.673 0.568 0.0% 0.917*** CS2 … in relation to your expectations? 4.556 0.745 2.0% 0.916*** CS3 … in relation to your idea of an ideal store visit? 4.550 0.642 1.0% 0.875*** Visiting the B&M store of X… CE1 … taught me interesting things. 3.176 0.973 9.9% 0.793*** CE2 … awoke my creativity. 3.469 1.015 5.0% 0.841*** CE3 … brought interesting ideas to my mind. 3.680 0.984 1.0% 0.857*** PAE1 … put me in a good mood. 4.327 0.801 0.0% 0.866*** PAE2 … made me feel happy. 3.760 0.971 5.0% 0.868*** PAE3 … made me feel optimistic. 3.615 0.863 5.0% 0.759*** NAE1 … put me in a bad mood. 1.465 0.962 2.0% 0.824*** NAE2 … made me feel frustrated. 1.747 1.137 2.0% 0.800*** NAE3 … made me feel discontented. 1.740 1.070 1.0% 0.913*** ISCX1 I enjoyed visiting the B&M store of X. 4.604 0.722 0.0% 0.888*** ISCX2 I enjoyed doing business at the B&M store of X. 4.707 0.593 2.0% 0.880*** ISCX3 I enjoyed spending time in the B&M store of X. 4.490 0.835 1.0% 0.907*** BMVI1 I intend to visit the B&M store of X in the future. 4.639 0.739 4.0% 0.947*** BMVI2 I plan to visit the B&M store of X in the future 4.469 0.888 3.0% 0.927*** OSVI1 I intend to visit the online store of X in the future 4.242 0.970 9.9% 0.936*** OSVI2 I plan to visit the online store of X in the future 4.132 0.968 9.9% 0.925*** *** = p < 0.001 M. Makkonen, L.Frank, T. Paananen, M. Holkkola, T. Kemppainen: The Cross-Channel Ef ects of In-Store Customer Experience in the Case of Omnichannel Fashion Retailing in Finland 625 Appendix B: Effects of the Control Variables ISCX CS BMVI OSVI Gender (0 = male and 1 = female) -0.086 0.080 0.231 0.007 Age (in years) -0.023 0.002 -0.004 -0.039 Made a purchase during the visit (0 = no and 1 = yes) 0.270 0.435* -0.040 -0.229 B&M store visiting frequency (1 = has never visited, 2 = less frequently than yearly, 3 = at least yearly, and 4 = at least – – 0.155(*) – monthly) Online store visiting frequency (1 = has never visited, 2 = less frequently than yearly, 3 = at least yearly, and 4 = at least – – – 0.525*** monthly) *** = p < 0.001, ** = p < 0.01, * = p < 0.05, (*) = p < 0.10 626 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY THE ROLE OF ANTICIPATED GUILT AND ITS NEUTRALISATION IN EXPLAINING RESPONSIBLE ONLINE SHOPPING MARKUS MAKKONEN, LAURI FRANK, TIINA PAANANEN University of Jyvaskyla, Faculty of Information Technology, Jyvaskyla, Finland markus.v.makkonen@jyu.fi, lauri.frank@jyu.fi, ti na.e.paananen@jyu.fi Although responsible consumption has been the target of growing interest in academic research, the antecedents of responsible online shopping remain poorly understood. In this study, we address this gap in prior research by focusing on the role of anticipated guilt in explaining responsible online Keywords: shopping. By using data from 479 Finnish consumers, we aim to responsible answer two research questions: (1) how strong an antecedent of online shopping, responsible online shopping intention is anticipated guilt in anticipated relation to other potential antecedents and (2) how efficiently can guilt, neutralisation consumers regulate their resulting feelings of guilt by using techniques, different kinds of neutralisation techniques? We find anticipated theory of guilt to be a strong antecedent of responsible online shopping planned behaviour, intention and the denial of responsibility, the denial of injury, and online the appeal to higher loyalties to be the most efficient survey, neutralisation techniques for consumers to regulate their feelings structural equation of guilt that result from not engaging in responsible online modelling, shopping. Finland DOI https://doi.org/10.18690/um.fov.4.2023.39 ISBN 978-961-286-751-5 628 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Sustainable development, which is commonly defined as development that meets the needs of the present without compromising the ability of future generations to meet their needs (United Nations, 1987), has become one of the main goals of most modern societies. One central component of sustainable development is responsible consumption (Jain et al., 2022), which refers to consumption that has a less negative or more positive impact on the environment, society, self, and others (Ulusoy, 2016). Although responsible consumption has been the target of growing interest also in academic research (cf. Webb et al., 2008; Gupta & Agrawal, 2018; Jain et al., 2022), few prior studies have focused on it in the context of online shopping, which can be seen as surprising when considering its widespread adoption among modern consumers. More specifical y, no prior studies that we are aware of have holistical y examined the antecedents of responsible online shopping in terms of what actual y causes consumers to make or not to make responsible consumption choices when shopping online. In this study, we aim to address this gap in prior research by focusing on the role of anticipated guilt in explaining responsible online shopping. Anticipated guilt refers to the feelings of guilt that arise from contemplating a potential deviation from one’s standards (Rawlings, 1970), and it has been used in marketing research to explain various aspects of consumer behaviour (Antonetti & Baines, 2015), including sustainable and responsible consumer behaviour (e.g., Onwezen et al. 2013, 2014a, 2014b; Antonetti & Maklan, 2014a, 2014b; Theotokis & Manganari, 2015; Lindenmeier et al., 2017). Thus, it can be assumed to play an important role also in explaining responsible online shopping. More specifical y, we focus on answering two research questions: (1) how strong an antecedent of responsible online shopping intention is anticipated guilt in relation to other potential antecedents and (2) how efficiently can consumers regulate their resulting feelings of guilt by using different kinds of neutralisation techniques? As our data, we use the responses from 479 Finnish consumers, which were col ected with an online survey and are analysed with structural equation modelling (SEM). After this introductory section, we briefly present the research model of the study in Section 2. The methodology and results of the study are reported in Sections 3 and 4, of which the results are discussed in more detail in Section 5. Final y, we conclude M. Makkonen, L. Frank, T. Paananen: The Role of Anticipated Guilt and its Neutralisation in Explaining Responsible Online Shopping 629 the paper with a brief discussion of the limitations of the study and some potential paths for future research in Section 6. 2 Research Model In order to examine the strength of anticipated guilt as an antecedent of responsible online shopping intention in relation to other potential antecedents, we apply the theory of planned behaviour (TPB) by Ajzen (1985, 1991), which is an extension of the theory of reasoned action (TRA) by Fishbein and Ajzen (1975, 1980) and one of the most commonly applied theories for explaining human behaviour in a variety of contexts (Fishbein & Ajzen, 2010). Some examples of these are the acceptance and use of various kinds of information technologies (IT) and information systems (IS) (e.g., Makkonen et al., 2010, 2012a, 2012b; Kari & Makkonen, 2014; Weigel et al., 2014), online shopping (e.g., Pavlou & Fygenson, 2006), sustainable and responsible consumption (e.g., Onwezen et al., 2013, 2014b; Han & Stoel, 2017), and even sustainable online shopping (Yang et al., 2018). In TPB, an individual’s intention to engage in a particular behaviour is hypothesised to be explained by three antecedents (Fishbein & Ajzen, 2010): the attitude toward the behaviour (i.e., an individual’s positive or negative evaluations of engaging in the behaviour), the subjective norm toward the behaviour (i.e., an individual’s perception of social pressure to engage or not engage in the behaviour), and the perceived behavioural control over the behaviour (i.e., an individual’s sense of capability, control, and self-efficacy to engage in the behaviour). In our research model, in line with the studies by Onwezen et al. (2014a, 2014b), we add anticipated guilt as the fourth antecedent of behavioural intention in order to compare its explanatory power with the three original antecedents of TPB. We hypothesise each of these four antecedents to have a positive effect on responsible online shopping intention, meaning that the more positive the attitude, the stronger the subjective norm and perceived behavioural control, and the more guilt an individual feels from not engaging in responsible online shopping, the stronger his or her responsible online shopping intention is hypothesised to be. In turn, in order to examine how efficiently consumers can regulate the feelings of guilt that result from not engaging in responsible online shopping, we apply the neutralisation theory by Sykes and Matza (1957), which was originally developed for the context of juvenile delinquency to explain how individuals are able to justify and 630 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY deflect the feelings of guilt that result from deviant behaviour by using different kinds of neutralisation techniques. More recently, it has been applied to also various other contexts, such as employee IS security policy violations (Siponen & Vance, 2010), software piracy (Siponen et al., 2012), shadow IT use (Silic et al., 2017), employee unauthorised computer access (Lin et al., 2018), and digital media piracy (Riekkinen, 2018) in IS, as well as inappropriate consumer behaviour (Strutton et al., 1994), fair trade (Chatzidakis et al., 2007), sustainable consumption (Antonetti & Maklan, 2014b; Gruber & Schlegelmilch, 2014), and immoral and unethical consumption (McGregor, 2008) in marketing. Originally, Sykes and Matza (1957) proposed five neutralisation techniques: the denial of responsibility (i.e., claiming not to be responsible for the deviant behaviour), the denial of injury (i.e., claiming that the deviant behaviour caused no injury), the denial of victim (i.e., claiming that the deviant behaviour was rightful when considering the circumstances), the condemnation of the condemners (i.e., claiming that those who condemn the deviant behaviour engage themselves in similar behaviour), and the appeal to higher loyalties (i.e., claiming that the deviant behaviour was due to actualising a higher-order ideal or value). Later, also other neutralisation techniques have been proposed, of which the most prominent are the metaphor of the ledger (i.e., claiming that the prior good behaviour counterbalances the present bad behaviour) by Klockars (1974) and the defence of necessity (i.e., claiming that the deviant behaviour was necessary) by Minor (1981). These al except for the denial of victim have been found to be used by consumers in the context of sustainable consumption in the study by Gruber and Schlegelmilch (2014), which is why we assume them to be used by consumers also in the closely related context of responsible consumption. Thus, in our research model, we add these six neutralisation techniques, which are the same ones that have been used also in the studies by Siponen and Vance (2010), Silic et al. (2017), and Lin et al. (2018), as antecedents of anticipated guilt, hypothesising them to have a negative effect on it. The resulting research model is illustrated in Figure 1. M. Makkonen, L. Frank, T. Paananen: The Role of Anticipated Guilt and its Neutralisation in Explaining Responsible Online Shopping 631 Figure 1: Research model 3 Methodology We collected the data for testing the research model from Finnish consumers between February 2023 and March 2023 with an online survey that was conducted by using the LimeSurvey service. The respondents for the survey were recruited by promoting the survey on social media and via the various communication channels of Finnish universities and student associations. As an incentive for responding, al the respondents who completed the survey were able to take part in a prize drawing of ten gift boxes, which were worth about 25 € each. Before measuring the constructs of the research model, responsible online shopping was first defined for the respondents as making consumption choices that take various ecological and ethical values (e.g., sustainable development and fair trade) into account while shopping online. Of the constructs, responsible online shopping intention, attitude, subjective norm, perceived behavioural control, and anticipated guilt were measured reflectively by three items each, whereas the six neutralisation techniques were measured reflectively by only two items each in order to avoid respondent fatigue. The wordings of these items are reported in Appendix A. The items for measuring responsible online shopping intention, attitude, subjective norm, and perceived behavioural control were adapted from the examples by Fishbein and Ajzen (2010), whereas the items for measuring anticipated guilt were adapted from the guilt inventory by Kugler and Jones (1992) as exemplified by 632 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Onwezen et al. (2013, 2014a, 2014b). In turn, the items for measuring the six neutralisation techniques were developed for this study based on the format by Siponen and Vance (2010) and the contextualisations by Gruber and Schlegelmilch (2014), who examined the meaning of these neutralisation techniques to consumers in the context of sustainable consumption. A seven-point semantic differential scale ranging from -3 to +3 was used for measuring attitude, whereas the traditional five-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, and 5 = strongly agree) was used for measuring al the other constructs. In addition, we used three control variables to control the effects of gender, age, and social desirability bias on responsible online shopping intention and anticipated guilt. Of these, gender and age were each measured with a single item, whereas social desirability bias was measured with the ten items and the measurement procedure proposed by Kuokkanen (2017), in which the item scores are transformed into a single continuous social desirability bias variable ranging from one (minimum social desirability bias) to five (maximum social desirability bias). In order to avoid forced responses, responding to al the items in the survey was voluntary, and not responding to a particular item resulted in a missing value. The data was analysed with covariance-based structural equation modelling (CB- SEM) by using the Mplus 8.8 software (Muthén & Muthén, 2023) and following the guidelines by Gefen et al. (2011) for SEM in administrative and social science research. As the model estimator, we used the robust maximum likelihood (MLR) estimator, which is able to handle also non-normal data. The potential missing values were handled by using the ful information maximum likelihood (FIML) estimator, which uses all the available data in model estimation. As the threshold for statistical significance, we used p < 0.05. 4 Results In total, we received 479 valid responses to the online survey. The descriptive statistics of this sample in terms of the gender, age, yearly personal taxable income, socioeconomic status, and average online shopping frequency of the respondents are reported in Table 1. As can be seen, most of the respondents were women, students, and relatively young, which was not surprising when considering our recruitment of the respondents. More specifical y, the age of the respondents ranged M. Makkonen, L. Frank, T. Paananen: The Role of Anticipated Guilt and its Neutralisation in Explaining Responsible Online Shopping 633 from 19 to 75 years, with a mean of 28.4 years and a standard deviation of 9.2 years. However, most of the respondents (68.7%) were relatively active online shoppers who shopped online at least monthly on average. Table 1: Descriptive statistics of the sample (N = 479) N % N % Gender Socioeconomic status Man 88 18.4 Student 341 71.2 Woman 365 76.2 Employee or self- employed 132 27.6 Other 26 5.4 Unemployed or unable to work 10 2.1 Age Pensioner 5 1.0 Under 25 years 206 43.0 Other 4 0.8 25–49 years 253 52.8 Online shopping frequency 50 years or over 20 4.2 At least weekly 31 6.5 Yearly personal taxable income At least monthly 298 62.2 Under 15,000 € 286 59.7 At least yearly 140 29.2 15,000–29,999 € 71 14.8 Less frequently than yearly 8 1.7 30,000 € or over 98 20.5 Has never shopped online 1 0.2 No response 24 5.0 No response 1 0.2 In the fol owing three subsections, we first evaluate the estimated model in terms of the reliability and validity of its constructs and indicators as well as its goodness-of-fit with the data. Final y, we report the model estimates. 4.1 Construct Reliability and Validity Construct reliability was evaluated from the perspective of internal consistency by using the composite reliability (CR) of the constructs (Fornell & Larcker, 1981), which is commonly expected to be at least 0.7 (Nunally & Bernstein, 1994). The CR of each construct is reported in the first column of Table 2, showing that al the constructs met this criterion. 634 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 2: Construct-level statistics Square roots of AVE and the correlations between the constructs CR AVE BI ATT SN PBC AG DOR DOI COC AHL MOL DON BI 0.917 0.787 0.887 ATT 0.829 0.618 0.434 0.786 SN 0.940 0.840 0.341 0.272 0.916 PBC 0.835 0.635 0.312 0.211 0.140 0.797 AG 0.838 0.634 0.443 0.205 0.118 0.030 0.796 DOR 0.932 0.874 -0.239 -0.290 -0.103 -0.117 -0.335 0.935 DOI 0.890 0.801 -0.265 -0.269 -0.148 0.032 -0.467 0.474 0.895 COC 0.874 0.777 -0.174 -0.235 -0.032 -0.032 -0.268 0.548 0.443 0.881 AHL 0.886 0.795 -0.204 -0.162 -0.199 -0.146 -0.293 0.271 0.224 0.319 0.892 MOL 0.787 0.650 -0.112 -0.119 -0.077 0.060 -0.191 0.413 0.493 0.540 0.283 0.806 DON 0.850 0.739 -0.008 0.074 0.006 -0.196 0.021 0.145 0.034 0.061 0.350 0.140 0.860 In turn, construct validity was evaluated from the perspectives of convergent and discriminant validity by using the two criteria by Fornell and Larcker (1981). They are both based on the average variance extracted (AVE) of the constructs, which is the average proportion of variance that a construct explains in its indicators. The first criterion concerning convergent validity expects each construct to have an AVE of at least 0.5. This means that, on average, each construct should explain at least half of the variance in its indicators. The AVE of each construct is reported in the second column of Table 2, showing that al the constructs met this criterion. In turn, the second criterion concerning discriminant validity expects each construct to have a square root of AVE that is at least equal to its absolute correlations with the other constructs in the model. This means that, on average, each construct should share at least an equal proportion of variance with its indicators compared to what it shares with the other constructs. The square root of AVE of each construct (on-diagonal) and the correlations between all the constructs in the model (off-diagonal) are reported in the remaining columns of Table 2, showing that this criterion was also met by al the constructs. 4.2 Indicator Reliability and Validity Indicator reliability and validity were evaluated by using the standardised loadings of the indicators, which are reported in Appendix B together with the means and standard deviations (SD) of the indicator scores as wel as the percentages of missing M. Makkonen, L. Frank, T. Paananen: The Role of Anticipated Guilt and its Neutralisation in Explaining Responsible Online Shopping 635 values. In the typical case of each indicator loading on only one construct, the standardised loading of each indicator is commonly expected to be statistical y significant and at least 0.707 (Fornell & Larcker, 1981). This is equivalent to the standardised residual of each indicator being at least 0.5, meaning that at least half of the variance in each indicator is explained by the construct on which it loads. As can be seen, this criterion was met by al the indicators except for PBC3. However, because the slightly lower loading of this indicator was not found to compromise the reliability or validity of the perceived behavioural control construct (cf. Section 4.1), we decided to retain it in the model. 4.3 Model Fit and Model Estimates The results of model estimation in terms of the standardised effect sizes and their statistical significance, the proportions of explained variance (R2), and model fit are reported in Figure 2 (and in Appendix C for the effects of the control variables). Model fit was evaluated by using the χ2 test of model fit and four model fit indices recommended by Hu and Bentler (1999): the comparative fit index (CFI), the Tucker-Lewis index (TLI), the root mean square error of approximation (RMSEA), and the standardised root mean square residual (SRMR). Of these, the χ2 test of model fit rejected the nul hypothesis of the model fitting the data, which is common in the case of large samples (Bentler & Bonett, 1980), whereas the values of the four model fit indices all met the cut-off criteria recommended by Hu and Bentler (1999): CFI ≥ 0.95, TLI ≥ 0.95, RMSEA ≤ 0.06, and SRMR ≤ 0.08. Thus, we consider the overal fit of the model acceptable. We also found no signs of multicollinearity or common method bias in the model. For example, the variance inflation factor (VIF) values calculated from the factor scores were all less than three (Hair et al., 2018), and the Harman’s single factor test (Podsakoff et al., 2003) suggested a very bad fit with the data (χ2(324) = 4,754.140, p < 0.001, CFI = 0.261, TLI = 0.199, RMSEA = 0.169, SRMR = 0.147). 636 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Figure 2: Model fit and model estimates (*** = p < 0.001, ** = p < 0.01, * = p < 0.05) Of the three original antecedents of TPB, attitude, subjective norm, and perceived behavioural control were all found to have statistically significant and positive effects on responsible online shopping intention. Similarly, the effect of anticipated guilt on responsible online shopping intention was found to be statistically significant, positive, and even slightly stronger than the effects of the three original antecedents of TPB based on the point estimates of the effect sizes reported in Figure 2. However, the overlapping 95% confidence intervals of the estimated effect sizes of attitude ([0.122, 0.372]), subjective norm ([0.070, 0.327]), perceived behavioural control ([0.098, 0.335]), and anticipated guilt ([0.222, 0.489]) suggested that the differences in the strengths of al the effects were statistically not significant. In turn, of the six neutralisation techniques, the denial of responsibility, the denial of injury, and the appeal to higher loyalties were found to have statistically significant and negative effects on anticipated guilt, whereas the effects of the condemnation of the condemners, the metaphor of the ledger, and the defence of necessity were found to be statistical y not significant. The effects of the control variables were found to be mostly statistically not significant, with the exception that age was found to have a statistical y significant and negative effect on responsible online shopping intention and being a man was found to have a statistical y significant and negative effect on anticipated guilt. In total, the model was able to explain 42.1% of the variance in responsible online shopping intention and 32.6% of the variance in anticipated guilt. Without anticipated guilt, the model would have been able to explain only 35.3% of M. Makkonen, L. Frank, T. Paananen: The Role of Anticipated Guilt and its Neutralisation in Explaining Responsible Online Shopping 637 the variance in responsible online shopping intention, so its addition resulted in a 6.8 percentage point promotion in explanatory power. 5 Discussion and Conclusion In this study, we examined the role of anticipated guilt in explaining responsible online shopping intention by focusing on answering two research questions: (1) how strong an antecedent of responsible online shopping intention is anticipated guilt in relation to other potential antecedents and (2) how efficiently can consumers regulate their resulting feelings of guilt by using different kinds of neutralisation techniques? In terms of the first research question, we found that anticipated guilt is indeed a strong antecedent of responsible online shopping intention, although not necessarily stronger than the three original antecedents of TPB. However, its addition to the research model was able to substantial y promote explanatory power in comparison to the basic TPB model, thus highlighting the role of anticipated guilt as an important additional antecedent of responsible online shopping intention. This promotion was consistent with the meta-analysis by Rivis et al. (2009), which found that adding different types of anticipated affect as an additional antecedent in the basic TPB model tends to promote the explained variance in behavioural intention by about five percentage points. In terms of the second research question, we found that neutralisation techniques are indeed an efficient way for consumers to regulate their feelings of guilt that result from not engaging in responsible online shopping. In total, the six neutralisation techniques in the research model were able to explain almost one-third of the variation in anticipated guilt, which can be seen as a substantial proportion, especially when considering that the neutralisation techniques are not assumed to act as the main antecedents of anticipated guilt but only to regulate the feelings of guilt that are caused by other antecedents, such as the dissonance between behavioural norms and one’s behaviour. Of these six neutralisation techniques, the most efficient ones were found to be the denial of responsibility, the denial of injury, and the appeal to higher loyalties. In other words, consumers can most effectively regulate their feelings of guilt that result from not engaging in responsible online shopping by justifying their behaviour with the fact that they cannot real y change anything with their own consumption choices alone, that this does not cause actual injury to anybody, and that they have to consider also other values or criteria (e.g., price) when 638 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY making their consumption choices. Of these, especialy the last justification seemed to be used very often by consumers based on the high mean scores of the corresponding indicators reported in Appendix B. When comparing these findings with those of prior studies that have examined the effectiveness of individual neutralisation techniques, they seem to support the suggestion by Silic et al. (2017) that the effectiveness of specific neutralisation techniques varies considerably between contexts. For example, Silic et al. (2017) found the metaphor of the ledger to be the only neutralisation technique that affects the use intention and actual use of shadow IT, whereas Siponen et al. (2012) found the condemnation of the condemners and the appeal to higher loyalties to be the only two neutralisation techniques that affect software piracy intention. These are both very different findings from those of our study. From a theoretical perspective, this study makes three main contributions. First, from the perspective of responsible consumption, the study promotes a more holistic understanding of the antecedents of responsible online shopping, in this case particularly anticipated guilt and the regulation of the feelings of guilt by using different kinds of neutralisation techniques. Second, from the perspective of TPB, the study responds to the cal s by Richard et al. (1996) and Rivis et al. (2009) for more research on the role of different types of anticipated affect as additional antecedents for explaining behavioural intention and actual behaviour. Third, from the perspective of the neutralisation theory, the study continues the work of Chatzidakis et al. (2007) as well as Gruber and Schlegelmilch (2014) concerning the application of neutralisation techniques to explain responsible and sustainable consumption as wel as the integration of neutralisation techniques with other prominent theories for explaining human behaviour, such as TPB. In turn, from a practical perspective, the main contribution of the study is the implication that the most efficient way for businesses and society to promote responsible online shopping is not only to manipulate the attitudes, subjective norms, and perceived behavioural control of consumers but also to expose them to stronger feelings of guilt. This may be achieved not only by using different kinds of guilt appeals in consumer communication (cf. Turner & Rains, 2021) but also by restraining the use of different kinds of neutralisation techniques among consumers. Here, the most relevant neutralisation techniques are obviously the denial of responsibility, the denial of injury, and the appeal to higher loyalties, which were M. Makkonen, L. Frank, T. Paananen: The Role of Anticipated Guilt and its Neutralisation in Explaining Responsible Online Shopping 639 found to be the most effective ones in our study. Of these, the use of the first two neutralisation techniques may potentially be restrained, for example, by providing consumers with more metrics on both the positive and the negative consequences of their consumption choices. This could help consumers to perceive that even the consumption choices of one person do have measurable consequences, thus undermining the main arguments behind the denial of responsibility and injury. These metrics could also be further coupled with goal setting, which together were found to result in an even stronger promotion of sustainable online shopping in the study by Kanay et al. (2021). In turn, the use of the appeal to higher loyalties may potential y be restrained by making responsible consumption choices more affordable or otherwise more accessible to average consumers. This way consumers would not be forced to choose between responsibility and other values or priorities, such as providing for one’s family by choosing a cheaper but less responsible product or service, thus once again undermining the main argument behind this neutralisation technique. 6 Limitations and Future Research We see this study to have three main limitations. First, although our sample was relatively heterogenous in demographic terms, it was biased toward women and younger consumers, which may limit the generalisability of our findings. Second, our research model focused only on the role of anticipated guilt as an antecedent of responsible online shopping intention and not, for example, on the potential cross-over effects between attitude, subjective norm, perceived behavioural control, and anticipated guilt, which have been proposed in some prior studies (e.g., Onwezen et al., 2013, 2014a, 2014b; Turel, 2016), or on the effects of anticipated guilt on actual responsible online shopping behaviour. Third, of the different neutralisation techniques, our research model focused only on the six neutralisation techniques that have been used also in the studies by Siponen and Vance (2010), Silic et al. (2017), and Lin et al. (2018) instead of others, such as the claim of relative acceptability and the claim of individuality by Henry and Eaton (1999), justification by comparison by Cromwel and Thurman (2003), and the claim of entitlement by Coleman (2005). Both these two latter limitations cannot be seen to compromise the findings of this study per se. However, addressing them could help to provide an even more complete understanding of the role of anticipated guilt and neutralisation techniques in explaining responsible online shopping. In addition to addressing the 640 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY aforementioned limitations, future research could focus on examining responsible online shopping in some more specific product or service contexts instead of only in general. One example of such context is fashion retailing, on which some prior research has already been conducted (e.g., Kemppainen et al., 2021, 2022). Other potential paths for future research could be to examine also the role of other types of anticipated affect, such as anticipated anxiety, shame, and worry (cf. Rivis et al., 2009), the role of positive and negative emotions more general y (e.g., Makkonen et al., 2019b), or the role individual values (e.g., Makkonen et al., 2019a) in explaining responsible online shopping. References Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In J. Kuhl & J. Beckman (Eds.), Action Control: From Cognition to Behavior (pp. 11–39). Heidelberg, Germany: Springer. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211. Ajzen, I., & Fishbein, M. (1980). Understanding Attitudes and Predicting Social Behavior, Englewood Cliffs, NJ: Prentice-Hall. Antonetti, P., & Baines, P. (2015). Guilt in marketing research: An elicitation–consumption perspective and research agenda. International Journal of Management Reviews, 17(3), 333–355. Antonetti, P., & Maklan, S. (2014a). Exploring postconsumption guilt and pride in the context of sustainability. Psychology & Marketing, 31(9), 717–735. Antonetti, P., & Maklan, S. (2014b). Feelings that make a difference: How guilt and pride convince consumers of the effectiveness of sustainable consumption choices. Journal of Business Ethics, 124(1), 117–134. Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin, 88(3), 588–606. Chatzidakis, A., Hibbert, S., & Smith, A. P. (2007). Why people don’t take their concerns about fair trade to the supermarket: The role of neutralisation. Journal of Business Ethics, 74(1), 89–100. Coleman, J. W. (2005). The Criminal Elite: Understanding White-Collar Crime (6th ed.). New York, NY: Worth Publishers. Cromwel , P., & Thurman, Q. (2003). The devil made me do it: Use of neutralizations by shoplifters. Deviant Behavior, 24(6), 535–550. Fishbein, M., & Ajzen, I. (1975). Belief, Attitude, Intention, and Behavior: An Introduction to Theory and Research. Reading, MA: Addison-Wesley. Fishbein, M., & Ajzen, I. (2010). Predicting and Changing Behavior: The Reasoned Action Approach. New York, NY: Psychology Press. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. Gefen, D., Rigdon, E. E., & Straub, D. (2011). Editor’s comments: An update and extension to SEM guidelines for administrative and social science research. MIS Quarterly, 35(2), iii–xiv. Gruber, V., & Schlegelmilch, B. B. (2014). How techniques of neutralization legitimize norm- and attitude-inconsistent consumer behavior. Journal of Business Ethics, 121(1), 29–45. Gupta, S., & Agrawal, R. (2018). Environmental y responsible consumption: Construct definition, scale development, and validation. Corporate Social Responsibility and Environmental Management, 25(4), 523–536. M. Makkonen, L. Frank, T. Paananen: The Role of Anticipated Guilt and its Neutralisation in Explaining Responsible Online Shopping 641 Hair, J. F., Jr., Black, W. C., Babin, B. J., & Anderson, R. E. (2018). Multivariate Data Analysis (8th ed.). Andover, United Kingdom: Cengage. Han, T.-I., & Stoel, L. (2017). Explaining social y responsible consumer behavior: A meta-analytic review of theory of planned behavior. Journal of International Consumer Marketing, 29(2), 91–103. Henry, S., & Eaton, R. (1999). Degrees of Deviance: Student Accounts of Their Deviant Behavior (2nd ed.). Salem, WI: Sheffield Publishing. Hu, L.-t., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55. Jain, V. K., Dahiya, A., Tyagi, V., & Sharma, P. (2022). Development and validation of scale to measure responsible consumption. Asia-Pacific Journal of Business Administration. Kanay, A., Hilton, D., Charalambides, L., Corrégé, J. B., Inaudi, E., Waroquier, L., & Cézéra, S. (2021). Making the carbon basket count: Goal setting promotes sustainable consumption in a simulated online supermarket. Journal of Economic Psychology, 83, 102348. Kari, T., & Makkonen, M. (2014). Explaining the usage intentions of exergames. In Proceedings of the 35th International Conference on Information Systems. Atlanta, GA: Association for Information Systems. Kemppainen, T., Frank, L., & Luhtanen, V. (2022). What is meaningful for responsible shoppers in online fashion retail? In P. Bednar, A. S. Islind, H. Val o Hult, A. Nolte, M. Rajanen, F. Zaghloul, A. Ravarini & A. M. Braccin (Eds.), Proceedings of the 8th International Workshop on SocioTechnical Perspective in IS Development. CEUR Workshop Proceedings 3239. Kemppainen, T., Frank, L., Makkonen, M., & Hyvönen, O.-I. (2021). Barriers to responsible consumption in e-commerce: Evidence from fashion shoppers. In A. Pucihar, M. Kljajić Borštnar, R. Bons, H. Cripps, A. Sheombar & D. Vidmar (Eds.), Proceedings of the 34th Bled eConference (pp. 323–335). Maribor, Slovenia: University of Maribor Press. Klockars, C. B. (1974). The Professional Fence. New York, NY: Free Press. Kugler, K., & Jones, W. H. (1992). On conceptualizing and assessing guilt. Journal of Personality and Social Psychology, 62(2), 318–327. Kuokkanen, H. (2017). Fictitious consumer responsibility? Quantifying social desirability bias in corporate social responsibility surveys. Palgrave Communications, 3, 16106. Lin, T.-C., Hsu, J. S.-C., Wang, Y.-C., & Wu, S. (2018). Examining the antecedents of employee unauthorized computer access. Journal of Statistics & Management Systems, 21(3), 493–517. Lindenmeier, J., Lwin, M., Andersch, H., Phau, I., & Seemann, A.-K. (2017). Anticipated consumer guilt: An investigation into its antecedents and consequences for fair-trade consumption. Journal of Macromarketing, 37(4), 444–459. Makkonen, M., Frank, L., Kari, T., & Moilanen, P. (2012a). Examining the usage intentions of exercise monitoring devices: The usage of pedometers and route trackers in Finland. In U. Lechner, D. Wigand & A. Pucihar (Eds.), Proceedings of the 25th Bled eConference (pp. 439–453). Kranj, Slovenia: Moderna organizacija. Makkonen, M., Frank, L., Kari, T., & Moilanen, P. (2012b). Explaining the usage intentions of exercise monitoring devices: The usage of heart rate monitors in Finland. In Proceedings of the 18th Americas Conference on Information Systems. Atlanta, GA: Association for Information Systems. Makkonen, M., Frank, L., & Kemppainen, T. (2019a). The effects of individual values on online shopping spending. In A. Pucihar, M. Kljajić Borštnar, R. Bons, J. Seitz, H. Cripps & D. Vidmar (Eds.), Proceedings of the 32nd Bled eConference (pp. 969–994). Maribor, Slovenia: University of Maribor Press. Makkonen, M., Halttunen, V., & Frank, L. (2010). Applying the theory of planned behaviour to explain the usage intentions of music download stores: Gender and age differences. In P. Kommers, T. Issa & P. Isaías (Eds.), Proceedings of the IADIS International Conference on Internet Technologies & Society 2010 (pp. 22–32). IADIS Press. Makkonen, M., Riekkinen, J., Frank, L., & Jussila, J. (2019b). The effects of positive and negative emotions during online shopping episodes on consumer satisfaction, repurchase intention, and recommendation intention. In A. Pucihar, M. Kljajić Borštnar, R. Bons, J. Seitz, H. Cripps & 642 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY D. Vidmar (Eds.), Proceedings of the 32nd Bled eConference (pp. 931–953). Maribor, Slovenia: University of Maribor Press. McGregor, S. L. T. (2008). Conceptualizing immoral and unethical consumption using neutralization theory. Family and Consumer Sciences Research Journal, 36(3), 261–276. Minor, W. W. (1981). Techniques of neutralization: A reconceptualization and empirical examination. Journal of Research in Crime and Delinquency, 18(2), 295–318. Muthén, L. K., & Muthén, B. O. (2023). Mplus Home Page. Available at https://www.statmodel.com Nunnal y, J. C., & Bernstein, I. H. (1994). Psychometric Theory (3rd ed.). New York, NY: McGraw-Hill. Onwezen, M. C., Antonides, G., & Bartels, J. (2013). The Norm Activation Model: An exploration of the functions of anticipated pride and guilt in pro-environmental behaviour. Journal of Economic Psychology, 39, 141–153. Onwezen, M. C., Bartels, J., & Antonides, G. (2014a). Environmental y friendly consumer choices: Cultural differences in the self-regulatory function of anticipated pride and guilt. Journal of Environmental Psychology, 40, 239–248. Onwezen, M. C., Bartels, J., & Antonides, G. (2014b). The self‑regulatory function of anticipated pride and guilt in a sustainable and healthy consumption context. European Journal of Social Psychology, 44(1), 53–68. Pavlou, P. A., & Fygenson, M. (2006). Understanding and predicting electronic commerce adoption: An extension of the theory of planned behavior. MIS Quarterly, 30(1), 115–143. Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879–903. Rawlings, E. I. (1970). Reactive guilt and anticipatory guilt in altruistic behavior. In J. Macaulay & L. Berkowitz (Eds.), Altruism and Helping Behavior (pp. 163–177). New York, NY: Academic Press. Richard, R., van der Pligt, J., & de Vries, N. (1996). Anticipated affect and behavioral choice. Basic and Applied Social Psychology, 18(2), 111–129. Riekkinen, J. (2018). Streaming Era Digital Media Piracy: An Integration of Three Theoretical Perspectives (Doctoral dissertation, University of Jyväskylä). Jyväskylä Studies in Computing 277. Rivis, A., Sheeran, P., & Armitage, C. J. (2009). Expanding the affective and normative components of the theory of planned behavior: A meta‑analysis of anticipated affect and moral norms. Journal of Applied Social Psychology, 39(12), 2985–3019. Silic, M., Barlow, J. B., & Back, A. (2017). A new perspective on neutralization and deterrence: Predicting shadow IT usage. Information & Management, 54(8), 1023–1037. Siponen, M., & Vance, A. (2010). Neutralization: New insights into the problem of employee information systems security policy violations. MIS Quarterly, 34(3), 487–502. Siponen, M., Vance, A., & Wil ison, R. (2012). New insights into the problem of software piracy: The effects of neutralization, shame, and moral beliefs. Information & Management, 49(7–8), 334– 341. Strutton, D., Vitel , S. J., & Pelton, L. E. (1994). How consumers may justify inappropriate behavior in market settings: An application on the techniques of neutralization. Journal of Business Research, 30(3), 253–260. Sykes, G. M., & Matza, D. (1957). Techniques of neutralization: A theory of delinquency. American Sociological Review, 22(6), 664–670. Theotokis, A., & Manganari, E. (2015). The impact of choice architecture on sustainable consumer behavior: The role of guilt. Journal of Business Ethics, 131(2), 423–437. Turel, O. (2016). Untangling the complex role of guilt in rational decisions to discontinue the use of a hedonic information system. European Journal of Information Systems, 25(5), 432–447. Turner, M., & Rains, S. (2021). Guilt appeals in persuasive communication: A meta-analytic review. Communication Studies, 72(4), 684–700. Ulusoy, E. (2016). Experiential responsible consumption. Journal of Business Research, 69(1), 284–297. M. Makkonen, L. Frank, T. Paananen: The Role of Anticipated Guilt and its Neutralisation in Explaining Responsible Online Shopping 643 United Nations (1987). Report of the World Commission on Environment and Development: Our Common Future. Available at https://sustainabledevelopment.un.org/content/documents/ 5987our-common-future.pdf Webb, D. J., Mohr, L. A., & Harris, K. E. (2008). A re-examination of social y responsible consumption and its measurement. Journal of Business Research, 61(2), 91–98. Weigel, F. K., Hazen, B. T., Cegielski, C. G., & Hal , D. J. (2014). Diffusion of innovations and the theory of planned behavior in information systems research: A meta-analysis. Communications of the Association for Information Systems, 34(1), 31. Yang, S., Li, L., & Zhang, J. (2018). Understanding consumers’ sustainable consumption intention at China’s Double-11 online shopping festival: An extended theory of planned behavior model. Sustainability, 10(6), 1801. 644 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Appendix A: Item Wordings Item Wording BI1 In future, I intend to make responsible choices when shopping online. BI2 In future, I plan to make responsible choices when shopping online. BI3 In future, I wil make responsible choices when shopping online. SN1 People who are important to me think that I should make responsible choices when shopping online. SN2 People who are important to me think that it would be good if I made responsible choices when shopping online. SN3 People who are important to me would want that I make responsible choices when shopping online. PBC1 If I want to, it is possible for me to make responsible choices when shopping online. PBC2 I am able to make responsible choices when shopping online if I want to. PBC3 It is up to me whether or not I make responsible choices online when shopping online. I find the idea of me making responsible choices when shopping online… ATT1 … negative vs. positive. ATT2 … harmful vs. beneficial. ATT3 … unpleasant vs. pleasant. If I do not make responsible consumption choices when shopping online, I feel… AG1 … guilty. AG2 … remorseful. AG3 … bad. I find that is OK for me not to make responsible consumption choices when shopping online because… DOR1 … one person cannot real y trigger any change with his or her choices. DOR2 … one person cannot real y change anything with his or her choices. DOI1 … it causes no actual harm to anybody. DOI2 … it caused no actual damage to anybody. COC1 … people who cal for responsibility from others sometimes do the same. COC2 … people who cal for responsibility from others do not always themselves make responsible choices. AHL1 … I have to consider also other values or criteria (e.g., price) when making my choices. AHL2 … I have to take into account also other values or criteria (e.g., price) when making my choices. MOL1 … I have already made enough responsible choices earlier in my life. MOL2 … the responsible choices that I have made earlier in my life compensate for it. DON1 … the lack of responsible alternatives sometimes makes it necessary. DON2 … responsible alternatives are not always available. M. Makkonen, L. Frank, T. Paananen: The Role of Anticipated Guilt and its Neutralisation in Explaining Responsible Online Shopping 645 Appendix B: Indicator-Level Statistics Item Mean SD Missing Loading Item Mean SD Missing Loading BI1 4.232 0.807 2.7% 0.935*** DOR1 2.160 1.176 0.8% 0.919*** BI2 4.292 0.862 1.5% 0.855*** DOR2 2.101 1.134 0.6% 0.950*** BI3 4.195 0.792 3.8% 0.869*** DOI1 1.886 0.983 0.8% 0.861*** ATT1 2.385 0.970 1.3% 0.807*** DOI2 1.943 1.007 1.0% 0.928*** ATT2 2.316 1.070 1.7% 0.760*** COC1 2.629 1.239 6.7% 0.830*** ATT3 1.928 1.275 1.7% 0.790*** COC2 2.456 1.219 4.4% 0.930*** SN1 3.558 1.067 15.4% 0.911*** AHL1 4.326 0.790 0.6% 0.911*** SN2 3.777 1.021 13.8% 0.910*** AHL2 4.224 0.878 0.4% 0.872*** SN3 3.640 1.046 17.1% 0.928*** MOL1 2.025 1.019 1.3% 0.763*** PBC1 4.102 0.953 1.7% 0.855*** MOL2 1.834 0.969 0.6% 0.847*** PBC2 4.054 0.960 2.5% 0.901*** DON1 4.025 0.985 1.3% 0.863*** PBC3 3.996 1.006 0.8% 0.601*** DON2 4.032 1.004 1.9% 0.856*** AG1 3.504 1.215 2.7% 0.858*** AG2 3.104 1.185 3.3% 0.722*** AG3 3.305 1.255 2.9% 0.802*** *** = p < 0.001 Appendix C: Effects of the Control Variables Variable BI AG Variable BI AG Gender = man -0.011 -0.180*** Age -0.080* -0.090 Gender = other -0.058 0.010 Social desirability bias 0.074 -0.015 *** = p < 0.001, ** = p < 0.01, * = p < 0.05 646 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY CONCEPTUALIZING INFORMATION SYSTEMS AS BIOLOGICAL ECOSYSTEMS - A ”NEW” VOCABULARY FOR SPEAKING OF INFORMATION SYSTEMS ANNA SELL,1 JOHN JEANSSON,2 MIRALEM HELMEFALK,2 HANS ALLMÉR,2 LEIF MARCUSSON2 1 Åbo Akademi University, Faculty of Social Science, Business and Economics, and Law, Åbo, Finland anna.sel @abo.fi 2 Linnaeus University, School of Business and Economics, Kalmar, Sweden john.jeansson@lnu.se, miralem.helmefalk@lnu.se, hans.allmer@lnu.se, leif.marcusson@lnu.se Biological ecosystems and related concepts are wel known and have been used for a long time, also outside of the ecology context. In this conceptual paper, we explore the use of biological ecosystem concepts as a new lens to understand and analyze information systems. We propose that applying a frame of reference from a different discipline enriches information systems analysis in several ways, firstly through broadening the Keywords: perspective of information systems, secondly by offering biological connections to phenomena and areas that were previously ecosystem, ecosystem, outside of the scope of the information system, and lastly information through offering a new viewpoint on actors, roles and functions system, within an information system. Further research is needed to information system ecosystem, deepen our understanding of the information system ecosystem socio-technical and apply this approach to other business activities. system DOI https://doi.org/10.18690/um.fov.4.2023.40 ISBN 978-961-286-751-5 648 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Digital transformation represents a new chapter in the human-technology relationship, disrupting traditional structures, logics and models (McAfee & Brynjolfsson, 2017). One example of such changed logics and models are internet enabled platforms where companies leverage digital technologies and collaborate in order to create and deliver an increased value proposition (Cusumano et al., 2019; Libert et al., 2016; Parker et al., 2016). Such platforms are often described as digitally enabled ecosystems inhabited by people, organizations and resources (Parker et al., 2016; Fehrer et al., 2018). At the core of platform business models are information systems. As information systems are very much a prerequisite and backbone of the ongoing digital transformation of organizations and society, their role and the way we think about them have not been excluded from being chal enged, on the contrary. This has initiated a cal from the information systems research community for a new or extended vocabulary in order to gain a richer understanding (Nischak et al., 2017; Benedict, 2018; Guggenberger et al., 2020). Alongside of the digital transformation, however not dependent on it, a discussion about ecosystems as a framework of understanding complex phenomena has emerged. This is very much related to an increased conversation on ecosystems from a biology and ecology perspective as climate issues have become more and more visible and critical. Ecosystem is about seeing wholeness and the term was original y coined by the English botanist Arthur G. Tansley (e.g., Tansley, 1935; Tansley, 1939) after suggestions from A. R. Clapham (Willis, 1997, p. 268). The concept of ecosystem has subsequently been developed and applied also in other fields outside of ecology. According to Adner (2017, p. 40) ecosystems can be explained as “the alignment structure of the multilateral set of partners that need to interact in order to focal value propositions to materialize”. The use of the ecosystem concept within information systems research has increased over the years and several authors have stressed the ambiguity of the concept seeking to provide theoretical synthesis in order to increase relevance (Nischak et al., 2017; Benedict, 2018; Guggenberger et al., 2020). The purpose of this paper is to explore the concept of information systems through the lens of biological ecosystems, in order to gain a richer picture of information systems as wel as to contribute to the understanding of how biological ecosystems A. Sel , J. Jeansson, M. Helmefalk, H. Al mér, L. Marcusson: Conceptualizing Information Systems as Biological Ecosystems - A ”New” Vocabulary for Speaking of Information Systems 649 can be used as a blueprint to understand complex phenomena. We explore how concepts from biological ecosystems can be applied to the analysis of information systems, focusing on concepts describing ecosystem actors, environment and presumptions. Our proposed contribution is conceptual. More specifical y, we aim at offering an alternative view; “to see something that has been identified in a new way”, cal ed revising by MacInnis (2011, p 138). MacInnis uses the metaphor of a person turning a kaleidoscope to reveal new perspectives to describe conceptual work of this type: we use a different frame of reference and view on a previously identified phenomenon. To do this, MacInnis suggests researchers can make use of novel metaphors – in our case biological ecosystem metaphors – and to seek new vantage points from other disciplines. Our work can also be understood through a design science lens. March and Smith (1995) state that constructs or concepts are one possible type of output or artifact from design science. The authors emphasize the importance and impact of terminology as tools for describing and thinking in a field. Hevner et al. (2004) outline seven guidelines for the design science research process. In line with these guidelines and within the scope of this article, we present a novel conceptualization, argue the relevance and contribution of our proposed vocabulary, offer a descriptive evaluation of it, and suggest steps for further research. 2 Ecosystems in business and information systems literature Research has linked various aspects of biological and ecological theories to business contexts in various settings to compare, analyze and shed new light on current practices and business theories. Most frequently it has its ground in using the biological ecosystem as a metaphor or analogy to other contexts. Biological ecosystems and evolutionary perspectives in business studies have focused on innovation, entrepreneurial, knowledge, organizational, and industrial ecosystems (e.g., Ghazinoory et al., 2021; Kuckertz, 2019; Pilinkienė & Mačiulis, 2014; Blijleven et al., 2013). A heap of research has been made synthesizing innovation ecosystems and its analogy to biological ecosystems or ecology theories (Shaw and Allen, 2018; Geng and Côté, 2002). Mars et al. (2012) states that the ecosystem metaphor “is a useful tool for understanding and predicting the conditions that shape and influence organizational systems.” (p. 279). Criticism towards using the term eco in innovation ecosystems has been voiced e.g. by Oh et al. (2016), where a flawed analogy to the biological ecosystem is debated. The dangers of using the analogy for innovation research is that the ecosystems are designed rather than evolved in this case, and include a variety of definitions and variations that may confuse. Ritala and 650 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Almpanopoulou (2017) instead defend the analogy and discuss some of the critique by Oh et al. (2016) suggesting that the term may be needed in understanding theory and practice. McMullen (2018) did a thought-experiment by deconstructing certain concepts in business and compared it to biological terms to shed new light on hybrid organizations. Some research has also focused on how business innovation ecosystems co-evolve and the role innovation has in them (Breslin et al., 2021), or looked at symbiotic relationships (Yoon et al., 2022). In the digital realm where other rules reign, problematization and discussion of the view of ecosystems has occurred (Márton, 2022). Briscoe and De Wilde (2009) imply that a digital ecosystem is to be regarded as a digital counterpart of biological ecosystems. They describe digital ecosystems as software systems that are robust, scalable, and self-organizing to meet users’ demand for digital services. Romero and Vernadat (2016) emphasize that an Executive Information System (EIS) contains a digital ecosystem where many information systems, sometimes hundreds, are included. There are several publications on ecosystems with a focus on information systems, information technology and information communication technology (ICT) (e.g. Anjum, 2023; Bash et al., 2008; Basole et al., 2015; Brummermann et al., 2011; Brummermann et al., 2012; Chamberlain & Said, 2022; Changjun & Hongbum, 2018; Diga & May, 2016; Karl et al., 2020; Schramm et al., 2012). While the ecosystem concept is commonly used, definitions vary greatly and are not established. 3 Information systems and biological ecosystems 3.1 Information systems as socio-technical systems and core concepts Information systems col ect, process, store and share information in order to support decision making and purposeful action in organizations. Information systems are frequently characterized as socio-technical systems consisting of technology and data (the technical sub-system), people and processes/tasks (the social sub-system); these separate components interact and together form a complex system (Figure 1). The socio-technical perspective enables a more nuanced understanding of both possibilities and problems surrounding organizational technology use. For example, investments into IT frequently fail. The reasons behind failure can be organizational or social rather than technical, or the reasons might be entanglements of technical and social reasons; socio-technical thinking is a useful framework for analysis in this context (Baxter and Sommerville, 2011). When A. Sel , J. Jeansson, M. Helmefalk, H. Al mér, L. Marcusson: Conceptualizing Information Systems as Biological Ecosystems - A ”New” Vocabulary for Speaking of Information Systems 651 working with complex real-world systems, using socio-technical approaches instead of traditional systems development methodologies is necessary (Wu et al., 2015). For actionable insights, socio-technical systems need to be model ed which has proven difficult due to e.g. the complexity of the systems, and the inter-disciplinary nature of socio-technical systems leading to a scattered research field. Consequently, methods for socio-technical analysis remain immature (ibid.) and central concepts underdeveloped (Millerand and Baker, 2010). Figure 1: Sociotechnical system Source: (Oosthuizen & Pretorious, 2016) For example, when discussing the People component of socio-technical systems, the concept user is an established term to describe the individual who is making use of the system. The user concept has, however, been criticized e.g. for being narrow and simplistic, and for creating an artificial separation between users and developers (Lamb & Kling, 2003; Millerand & Baker, 2010). Within the Work System Framework, instead of ‘users’, the concepts of participants and customers are used to describe the people working with the system (Alter, 2004). Participants encompass those individuals who perform at least part of the target business process, and customers are people who receive benefit from the products or services produced. In another twist, Actor-Network theory makes use of the concepts actor and actant to denote both human and non-human elements participating in a network (Hanseth et al., 2004). In other words, by using different concepts and definitions for the people involved with the information system, we can examine the same phenomena from different viewpoints and gain new understanding. Hence, in 652 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY section 4 of this paper we explore how concepts derived from biological ecosystems can be applied to information systems in order to gain novel insight. In the following section we overview a selection of biological ecosystem concepts. 3.2 Biological ecosystems and core concepts In biology, an ecosystem includes al living things with their habitat within an area. An ecosystem has properties of self-organization, scalability, and sustainability. An ecosystem is self-regulating and consists of independent actors with selfish interests acting in an overal functioning environment. In addition to its internal integrity, an ecosystem can be affected by external factors. The word ecosystem consists of the two concepts ecology and system. Ecology, a nomenclature, includes an interaction between what lives, i.e., the biotic, and the non-living, i.e., the abiotic. A system is several parts that are connected to each other and that together form an ordered whole. In summary, the biological ecosystem consists of a biotic component, containing Organism and Function, and an abiotic component, containing Presumption and Habitat. The two components have interaction both between and within, furthermore they occupy a physical space (e.g., NE, 2023; Adner, 2017; Dhil on et al., 2013; Guggenberger et al., 2020; Nischak et al., 2017; Transley, 1939; Willis, 1997). There are different ways to describe the next order of principles in the biological ecosystem, one way is shown in figure 2. To clarify the meanings and content of an ecosystem, the Baltic Sea ecosystem is presented below. Figure 2: Principles of a biological ecosystems top three levels A. Sel , J. Jeansson, M. Helmefalk, H. Al mér, L. Marcusson: Conceptualizing Information Systems as Biological Ecosystems - A ”New” Vocabulary for Speaking of Information Systems 653 The notion Biotic means living and can be divided into society, population and individual. A society contains all populations of different living species within a geographic area. In our example the society is all living species inhabiting the Baltic Sea. A population is a group of individuals belonging to the same species, living in the same area and interacting with each other. An example of a population in the Baltic Sea is a shoal of Baltic cods. An individual is the individual animal, for example an individual Baltic cod. Organisms in an ecosystem can perform the functions or roles of producer, consumer and decomposer. These functions are key enablers for the circulation of energy within the ecosystem. A producer can, e.g., by using photosynthesis, transform energy into food, both for themselves, and for other organisms. In the Baltic Sea ecosystem, planktonic algae act as primary producers (John Nurminen Foundation, n.d.). Consumers are the animals and fungi that cannot produce their own food. The Baltic cod is a consumer in the Baltic Sea ecosystem. Decomposers break down dead plants and animals, and release energy back into circulation. Main decomposers in the Baltic Sea are various aerobic and anaerobic bacteria (Furman et al., n.d.). Abiotic means nonliving and is dependent on Geology and Climate. In the Baltic Sea ecosystem, the abiotic are for example seabed, water and seacoast. Geology deals with the outer covering of the planet earth (the Lithosphere i.e., the earth's crust with the upper mantle). This casing covers soils and how they are composed and structured, to which over time is added development. Here it is possible to present stones, bedrock, sand and mud to mention a few. Climate controls the weather and is a combination of physical average atmospheric conditions over a longer period for a geographic area. This includes, e.g. humidity, temperature and wind. The climate is natural y affected by e.g., the sun's radiation and the tilt of the earth's axis. It is also affected by human activities such as emissions of greenhouse gases and deforestation. A warmer climate leads, among other things, to increased water levels, stronger storms, and changing seasons. A nature type has common flora and fauna within a geographical y defined area. Examples of Swedish nature types are deciduous forest, coniferous forest, and mountains. The Baltic cod lives in brackish water. The biological habitat is the sum 654 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY of the external circumstances that an individual needs in order to survive. The habitat for the Baltic cod is overfertilized, contains environmental toxins and suffers from overfishing of e.g., Baltic Herring. Spread is the geographical area within which a species is found; the Baltic cod exists only in the Baltic Sea. 4 Information systems as ecosystems Considering the concepts of information systems and biological ecosystems side by side, some similarities emerge. Both information systems and biological ecosystems are inherently complex and chal enging to understand. Information systems consist of multiple layers of technology, people and social structures, and biological ecosystems consist of multiple species and environmental factors. Changes in any of the separate components of the biological ecosystem or the information system impacts al parts of the system. Furthermore, the flow, exchange and transformation of energy is the focal function of an ecosystem - the flow, exchange and transformation (or refining) of data is central to the information system. Some concepts originating from biological ecosystems are already used in discussing information systems, especial y holistic, behavioral concepts such as feedback loops, equifinality and adaptability. Even so, some authors suggest that the information systems discipline has done too little to understand the system nature of information systems, and in practice too often focusing on the technology as a tool (Alter, 2004). We suggest the use of ecosystems concepts by adapting terminology describing organisms, functions, presumptions and habitats as outlined in figure 2 to the information system. As a practical il ustration we describe the concepts in the context of how the Ladok system is used by Linnaeus university in Sweden. Ladok (2023) is a Swedish national system that provides 40 Swedish universities with support in the study administrative work and is used by students, teachers and administrators. Ladok is used to store information about students’ attendance, results and other data (Ladok, 2023). The division made within ecosystems theory between biotic and abiotic subsystems bears a resemblance to the division often made within socio-technical systems and its social and technical subsystems, the biotic component corresponding to the social, and abiotic corresponding to the technical subsystem. Taking the analogy further, we explore the use of the biotic components organism and function and the abiotic components presumption and habitat to describe and understand an A. Sel , J. Jeansson, M. Helmefalk, H. Al mér, L. Marcusson: Conceptualizing Information Systems as Biological Ecosystems - A ”New” Vocabulary for Speaking of Information Systems 655 information system (Figure 3). Doing this, a new, enriched picture of the information system emerges, distinct from and complementary to previous socio- technical and system model ing approaches. We propose using the term organism in order to depict the character and nature of stakeholders/actors that use, contribute, as wel as benefit from the data provided by the information system. When viewing Ladok through this lens, we can identify individual stakeholders, such as university teachers using the system to enter grades; intra-organizational stakeholders such as different faculty departments and educational programs using the system in order to achieve organizational goals; inter-organizational stakeholders such as funding and accreditation bodies using data from the system in order to validate and ensure quality. Figure 3: Components of the biotic and abiotic subsystems The function component is defined as the activity areas or roles enabling the flow of data within the information system and consists of the activity roles: producer, consumer and decomposer. The activity role of producer depicts and concerns how data is put into the system; the activity role of consumer depicts how data is extracted from the system and presented; the activity role of decomposer depicts how existing data is refined, mined, combined and analyzed in ways that enable new understanding. The different stakeholders identified within the organism component each, concurrently or at different times, could hold the different roles 656 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY of producer, consumer or decomposer of data. For example, an individual teacher is both a producer of data when registering course grades, and a consumer of data when accessing student course registrations ahead of an upcoming course. The biotic components of an information system provide a greater understanding of the who (organism) and why (functions) of information systems. Turning to the abiotic subsystem, here we investigate the conditions under which the information system operates. The presumption component consists of geology and climate and depict physical, technical and climactic aspects affecting the flow and use of data in the information system. Geology of the information system refers to the tangible infrastructural and technical aspects necessary for the information system to function, such as network infrastructure, hardware and software. Climate, on the other hand, refers to both tangible aspects (e.g. boundaries set by government, industry and organizational regulations and guidelines), as well as intangible aspects (e.g. strategies and culture on both an organizational and national level). Figure 4: Principles of information system as an ecosystem, top three levels A. Sel , J. Jeansson, M. Helmefalk, H. Al mér, L. Marcusson: Conceptualizing Information Systems as Biological Ecosystems - A ”New” Vocabulary for Speaking of Information Systems 657 The habitat component depicts the environments which constitute the context and place of operations of an information system. Firstly, there is the micro-habitat (type of nature) which constitutes organizational processes, routines, and activities where the information system is intended to provide support and be used. Information systems are typical y used with the purpose to enforce, improve, or automate processes, related to the previously mentioned functions of producing, consuming, and decomposing data. In the Ladok example, there are the processes related to gathering and storing student credit data. A practical consideration is that processes with the same goal might be enacted differently at different faculties of the university, or at geographically separate campuses. Secondly, there is the meso-habitat (living environment), which constitutes organizational and decision-making structures and hierarchies, as wel as interactions between different information systems. At the meso-habitat level the focal point of information systems is the support of, and alignment, with organizational or company-wide functions, strategies, and goals. Thirdly there is the macro-habitat (spread), which constitutes industry and market platforms, participation in networks etc. It is very much the external context in which information systems support the expanded organization, enabling cooperation, coopetition, and a sustainable competitive position. In our example, the Linnaeus university is a member of the Baltic University Program, a collaboration between 90 universities in the Baltic Sea region requiring the exchange of data between information systems to for example fulfil reporting requirements. The abiotic components provide a greater understanding of the what and where (habitat) as wel as the how (presumptions) of an information system. They enable a rich picture of both the context and outcome of information systems, as wel as the nature and character of needed infrastructure. 658 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Figure 5: Three levels of circulation in information system ecosystem When analyzing any system, a holistic view and investigating interdependence of the components are central (Figure 5). For example, while the organism component is useful to identify internal and external stakeholders, further insight is gained when considering how different stakeholders engage with the information system through the function component (producing, consuming and decomposing data). A deeper level of understanding is then gained when analyzing connections and interactions between sub-systems or components of the system at hand. The information system ecosystem has a system boundary providing an interface with an external environment. This environment affects and is affected by the ecosystem. The ecosystem with its biotic and abiotic components constitutes a constant, ongoing circulation, iteration, and/or feedback of data in and between all levels in an information system ecosystem. Data then becomes the heart of an information systems ecosystem. In figure 5 the example of Ladok is used to il ustrate this A. Sel , J. Jeansson, M. Helmefalk, H. Al mér, L. Marcusson: Conceptualizing Information Systems as Biological Ecosystems - A ”New” Vocabulary for Speaking of Information Systems 659 complexity. The dashed line depicts outer, inner, and sub level boundaries. Furthermore, “e” stands for education and “g” for Ladok as a grading application. 5 Conclusion We believe the biological ecosystem concepts give new insights as compared to traditional information systems analysis concepts. Ecosystem analysis goes beyond processes, taking into consideration also structures and hierarchies. For example, the concept Habitat gives a novel lens to analyze the boundary conditions impacting an information system, and by its division into three levels broadens the scope of the scrutiny both within the organization and beyond it. Further, in traditional systems analysis, tools such as data flow diagrams are used to model how data moves in the system, giving a necessary technical blueprint for how to build the system. The Function concepts, on the other hand, look at how and which data is needed in the socio-technical system for e.g. decision making, at the sources of the data, and at the usage of the data. The flow of data to and from stakeholders becomes a central focal point in the analysis, as al components of the system are analyzed in relation to the flow of data. We gain new understanding of the varied relations between stakeholders and data. Roles are not understood through the work tasks performed by the individual, or position in the organizational hierarchy, but as roles in relation to data. Using the biological ecosystem as a model forces a questioning of traditional information system concepts. It also means an extended and richer system perspective which is beneficial for the view of information systems. The ecosystem concepts give researchers and practitioners a new lens to use when studying, planning, designing or troubleshooting an information system. In future research, we intend to further explore the notion of the information system ecosystem, and to carry out a more systematic evaluation of the framework, e.g. as a case study. As an additional evaluation of the proposed approach, a thorough review of and comparison to alternative information systems analysis frameworks and models is proposed. Final y, we intend to explore the use of the suggested framework in other business contexts, e.g. marketing. 660 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY References Adner, R. (2017). Ecosystem as Structure: An Actionable Construct for Strategy. Journal of Management, volume 43, number 1, pages 39-58. Alter, S. (2004). Desperately seeking systems thinking in the IS discipline. Proceedings of ICIS-25, the International Conference on Information Systems, Washington, DC, 757-769. Anjum, B. (2023). A conversation with Grace IbukunOluwa Ufeoshi: Preparing African youth for the future of work in the IT ecosystem. Ubiquity, Doi:10.1145/3579372 Bash, C.E., Patel, C.D., Shah, A.J. & Sharma, R.K. (2008). The sustainable information technology ecosystem. Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems, p. 1126-1131 Basole, R.C., Park, H. & Barnett, B.C. (2015). Coopetition and convergence in the ICT ecosystem. Telecommunications Policy, 39:527-552 Baxter, G., & Sommerville, I. (2011). Socio-technical systems: From design methods to systems engineering. Interacting with computers, 23(1), 4-17. Benedict, M., (2018), Modelling ecosystem in information systems – a typology approach . Multikonferenz Wirtschaftsinformatik, Lüneburg Germany, March 06-09, 2018, https://www.researchgate.net/publication/323218912 Blijleven, V., van Angeren, J., Jansen, S., & Brinkkemper, S. (2013, July). An evolutionary economics approach to ecosystem dynamics. In 2013 7th IEEE International Conference on Digital Ecosystems and Technologies (DEST) (pp. 19-24). IEEE Breslin, D., Kask, J., Schlaile, M., & Abatecola, G. (2021). Developing a coevolutionary account of innovation ecosystems. Industrial Marketing Management, 98, 59-68. Briscoe, G., & De Wilde, P. (2009). Digital Ecosystems: Evolving Service-Oriented Architectures. Computer Science. Retrieved from https://ai2-s2-pdfs.s3.amazonaws.com/04f0/33e8fe766d07025c3efe4f457bea7c0d69fb.pdf Brummermann, H., Keunecke, M. & Scmid, K. (2011). Variability issues in the evolution of information system ecosystems. In Proceedings of the 5th Workshop on Variability Modeling of Software-Intensive Systems (pp. 159-164) . Brummermann, H., Keunecke, M. & Scmid, K. (2012). Formalizing distributed evolution of variability in information system ecosystems. In Proceedings of the 6th Workshop on Variability Modeling of Software-Intensive Systems (pp. 11-19) . Chamberlain, L.M. & Said, H. (2022). The early IT ecosystem: Re-envisioning dual credit, college access and affordability, and teachers as professionals. Policy Futures in Education. Doi:10.1177/14782103211066675 Changjun, L. & Hongbum, K. (2018). The evolutionary trajectory of an ICT ecosystem: A network analysis based on media users’ data. Information & Management, 55:795-805 Cusumano, M.A., Gawer, A. & Yoffie, D.B. (2019). Business of platforms, strategy in the age of digital competition, innovation, and power. HarperCollins Publishers, New York. Dhil on, S.K., Chiew, S.L., Leow, S.L., Sidhu, A.S., Shuhaimi, N.I., Leong, Y.M. & Chong, V.C. (2013). A model of digital biological ecosystem. Systematics and biodiversity, 11(4):425-435, Doi:10.1080/14772000.2013.856962 Diga, K. & May, J. (2016). The ICT ecosystem: The application, usefulness, and future of an evolving concept. Information Technology for Development, vol. 22, supplement 1, p. 1-6, Doi:10.1080/02681102.2016.1168218 Furman, E., Pihlajamäki, M., Välipakka, P. & Myrberg, K. (n.d.). The Baltic Sea. Environment and Ecology. Finnish Environment Institute, https://www.syke.fi/download/noname/%7B45C3C2FD-9301-43D7-803F- 90460B436AE5%7D/99119 Geng, Y., & Côté, R. P. (2002). Scavengers and decomposers in an eco-industrial park. The International Journal of Sustainable Development & World Ecology, 9(4), 333-340. A. Sel , J. Jeansson, M. Helmefalk, H. Al mér, L. Marcusson: Conceptualizing Information Systems as Biological Ecosystems - A ”New” Vocabulary for Speaking of Information Systems 661 Ghazinoory, S., Philips, F., Afshari-Mofrad, M., & Bigdelou, N. (2021). Innovation lives in ecotones, not ecosystems. Journal of Business Research, 135, 572-580. Guggenberger, T., Möl er, F, Haarhaus, T & Gür, I. (2020). Ecosystem types in information systems. ECIS 2020, Marrakesh, Morocco, https://www.researchgate.net/publication/341188637 Hanseth, O., Aanestad, M., & Berg, M. (2004). Guest editors’ introduction: Actor‐network theory and information systems. What's so special?. Information technology & people, 17(2), 116-123. Hevner, A.R., March, S.T., Park, J. & Ram, S., (2004). Design Science in Information System Research, MIS Quartely, vol 28 no 1 pp 75–105. John Nurminen Foundation (n.d.). Marine biodiversity. https://johnnurmisensaatio.fi/en/the-baltic-sea/marine-biodiversity/ Karl, H., Kundisch, D., Meyer auf der Heide, F. & Wehrheim, H. (2020). A case for a new IT ecosystem: On-The-Fly computing. Business Information System, 62(6):467-481, Doi:10.1007/s12599-019-00627-x Kuckertz, A. (2019). Let's take the entrepreneurial ecosystem metaphor seriously! Journal of Business Venturing Insights, 11, e00124. Ladok, (2023), Ladok, https://ladok.se Lamb, R., & Kling, R. (2003). Reconceptualizing users as social actors in information systems research. MIS quarterly, 197-236. Libert, B., Beck, M., & Wind, J. (2016). The network imperative, how to survive and grow in the age of digital business models. Harvard business press, Boston. MacInnis, D. J. (2011). A framework for conceptual contributions in marketing. Journal of Marketing, 75(4), 136-154. March, S.T. & Smith, G.F. (1995). Design and natural science research on information technology. Decision Support Systems, no 15 pp 251–266 Mars, M. M., Bronstein, J. L., & Lusch, R. F. (2012). The value of a metaphor: Organizations and ecosystems. Organizational Dynamics, 41(4), 271-280. Márton, A. (2022). Steps toward a digital ecology: ecological principles for the study of digital ecosystems. Journal of Information Technology, 37(3), 250-265. McAfee, A., and Brynjolfsson E., (2017). Machine, platform, crowd, harnessing our digital future. W.W. Norton & Company Inc., New York McMul en, J. S. (2018). Organizational hybrids as biological hybrids: Insights for research on the relationship between social enterprise and the entrepreneurial ecosystem. Journal of business venturing, 33(5), 575-590. Mil erand, F., & Baker, K. S. (2010). Who are the users? Who are the developers? Webs of users and developers in the development process of a technical standard. Information Systems Journal, 20(2), 137-161. NE. (2023). Nationalencyklopedin [Swedish encyclopedia], different keywords, https://www-ne-se.proxy.lnu.se/uppslagsverk/encyklopedi Nischak, F., Hanelt, A., & Kolbe, L. M. (2017). Unraveling the interaction of information systems and ecosystems - A comprehensive classification of literature. ICIS 2017 Proceedings, 20. Oh, D. S., Phillips, F., Park, S., & Lee, E. (2016). Innovation ecosystems: A critical examination. Technovation, 54, 1-6. Oosthuizen, R., & Pretorius, L. (2016). Assessing the impact of new technology on complex sociotechnical systems. South African Journal of Industrial Engineering, 27(2), 15-29. Parker, G. G., Van Alstyne, M. W., & Choudary, S. P. (2016). Platform revolution: How networked markets are transforming the economy and how to make them work for you. WW Norton & Company, New York. Pilinkienė, V., & Mačiulis, P. (2014). Comparison of different ecosystem analogies: The main economic determinants and levels of impact. Procedia-social and behavioral sciences, 156, 365-370. Ritala, P., & Almpanopoulou, A. (2017). In defense of ‘eco’in innovation ecosystem. Technovation, 60, 39-42. Romero, D., & Vernadat, F. (2016). Enterprise information systems state of the art: Past, present and future trends. Computers in Industry, 79, 3-13. Doi:10.1016/j.compind.2016.03.001 662 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Schramm, W., Köstinger, H., Bayrhammer, K., Fiedler, M. & Grechenig, T. (2012). Developing a hospital information system ecosystem for creating new clinical col aboration methodologies, IEEE-EMBS International Conference on Biomedical and Health Informatics Shaw, D. R., & Al en, T. (2018). Studying innovation ecosystems using ecology theory. Technological Forecasting and Social Change, 136, 88-102. Tansley, A.G. (1935). The use and abuse of vegetational concepts and terms. Ecology, vol. 16, no. 3, pp. 284-307 Tansley, A.G. (1939). British ecology during the past quarter-century: the plant community and the ecosystem. Journal of ecology, vol. 27, no. 2, pp. 513-530 Willis, A.J. (1997). The ecosystem: An evolving concept viewed historical y. Functional Ecology, vol., 11, no. 2, p. 268-271 Wu, P. P. Y., Fookes, C., Pitchforth, J., & Mengersen, K. (2015). A framework for model integration and holistic modelling of socio-technical systems. Decision Support Systems, 71, 14-27. THE THEORY OF IDENTITY MANAGEMENT EXTENDED TO THE AUTHENTICATION OF IDENTITY ASSERTIONS ROGER CLARKE Xamax Consultancy Pty Ltd, Canberra, UNSW Law, Sydney; Australian National University School of Computing, Canberra, Australia Roger.Clarke@xamax.com.au At the 35th Bled eConference, a previously-published pragmatic metatheoretic model was articulated in the context of identity management. The present paper extends that theory to the authentication of the various categories of assertions that arise in identity management activities. The extended theory reflects the important distinction between the concepts of identity and entity. It deals with the fundamental categories, in which a particular data-record is claimed to relate to a particular physical entity or virtual identity; but it also encompasses claims about Keywords: the properties of real-world things, and assertions that records in IS practice, two different data-sets apply to the same (id)entity. The analysis assertion, evidence, has important implications for the practice of IS, and for IS property, researchers whose work is intended to influence IS practice. principal-agent DOI https://doi.org/10.18690/um.fov.4.2023.41 ISBN 978-961-286-751-5 664 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Organisations have encountered ongoing difficulties in the areas of identification and authentication, particularly where the entities in question are human beings. The topic of identity authentication is addressed in the technical information technology (IT) literature, although far less so in information systems (IS). On the other hand, authentication as a general concept and a family of business processes has attracted remarkably little attention. For example, of the 465 refereed papers from the Bled eConference accessible in the AIS electronic Library (AISeL), covering the last two decades, only 2 have the term 'authentication' in Title or Abstract. In AISeL as a whole, the corresponding count is only 38 of >17,000, and across the Basket of 8 IS journals, a mere 12 of >10,000 articles. Effective IS depend on a deep appreciation by IS designers of the nature of the relevant phenomena, and hence researchers need to pay far more attention to the topic. IS professionals need to be supported by a model that is pragmatic, by which is meant that it is a fit to the needs of IS practitioners, but that also reflects insights from relevant aspects of philosophy. This paper builds on prior work in two areas. The first presents a pragmatic metatheoretic model to support IS, and the second examines the general concept of authentication. Together, these provide a suitable basis for an analysis of processes to authenticate assertions about identity. The paper commences with a recapitulation of prior work that establishes the pragmatic metatheoretic model, and explains the generic concept of assertion authentication. The main body of the paper identifies and discusses specific categories of assertion in which identities and/or entities play a key role. Processes are then presented whereby the reliability of those kinds of assertions can be assessed. The effectiveness of those processes is shown to be dependent on evidence, and hence on the quality of the data on which reliance is placed. The application of these ideas to IS practice is argued to enable improvements in authentication design and operation. The need is shown for further work on their application to particular kinds of entities and identities and in particular contexts. R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 665 2 Prior Work Organised activity involves dependence by parties on statements made by other parties. The longstanding maxim 'Trust, but verify' conveys that, in principle, all such statements on which dependence is placed need to be checked, but with effort invested proportionately to the harm that would arise from unjustified trust. Organisations need IS to be designed in a manner that reflects, supports and facilitates the checking of important statements. For that to be achieved, the model on which IS practice is founded needs to have an effective fit with the manner in which business enterprises, government agencies, not-for-profits and smal businesses perceive the realities of their operational environment. The first sub-section below summarises a model that is pragmatic, in that it supports understanding about and action in the world, yet has a firm foundation in relevant aspects of meta-theory, particularly ontology (the study of existence), epistemology (the study of knowledge), and axiology (the study of values). The second sub-section introduces a further area of prior work which examines the general notion of authentication. On these two foundations, the remainder of the paper builds an analysis of the authentication of assertions that involve identities and entities. 2.1 The Pragmatic Metatheoretic Model Prior work reported at the Australasian Conference in Information Systems (ACIS) in Clarke (2021) presents a pragmatic metatheoretic model conceived in order to support both IS practice and that portion of IS research activity that is intended to be relevant to IS practice. The foundational metatheoretic aspect is ontology, which is concerned with phenomena, and hence with the properties of things and events. The assumption adopted in the model is a conventional compromise between materialist and idealist notions, postulating that there are both material realities (the Real World) and internal mind-stuff (as that term is used by William Kingdon Clifford to refer to the intel ectual or Abstract World – Coneybeare 1892): phenomena and their properties inhabit the Real World; whereas ideas are of the Abstract World. 666 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY The second aspect, epistemology, is the study of knowledge, and its sources, varieties and limits. Competing views are empiricism, which holds that knowledge is derived from sensory experience, and apriorism or rationalism, which considers that knowledge is or at least can be innate and/or derived from the human faculty of reason. A pragmatic metatheoretic approach must support IS practitioners not only in contexts that are simple, stable and uncontroversial, but also where there is no expressible, singular, uncontested 'truth'. Some relatively closed systems, such as fly-by-wire, industrial control, and robotic assembly line management, can reasonably be treated as technical systems. In the large majority of IS, on the other hand, interaction among IT artefacts and people is intrinsic, and meaningful study of them requires the adoption of both a socio-technical perspective (Abbas & Michael 2022), and interpretivist or critical theory approaches to research. The epistemological commitment underlying the model is accordingly that knowledge depends on appropriately blending sensory experience, human imagination, and reasoning, and accommodating both tacit and codified knowledge. A third important branch of philosophy, axiology, is less familiar. It is concerned with how value is imputed to things. Organised activities depend on people, artefacts, and effective interactions among them. IS also affect people, including those participating in the system (conventional y cal ed 'users') and some who are not participants in the system, but are affected by it (useful y referred to as 'usees' – Berleur & Drumm 1991 p.388, Clarke 1992, Fischer-Huebner & Lindskog 2001, Baumer 2015). Examples of usees include people to whom records in shared industry databases refer, such as those for police suspects and tenants. Human values are accordingly central to the model. In addition, IS involve various stakeholders, and value-conflicts are inherent. The approach adopted in the pragmatic metatheoretic model is that value is dependent on the observer, that there are generally multiple observers of any given phenomenon, and hence that IS must support the integration of multiple perspectives rather than assume that one necessarily dominates (Clarke & Davison 2020), and must continue to function where tension continues among two or more perspectives. R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 667 At the Bled eConference, in Clarke (2022), the pragmatic model was applied to identity management. A diagrammatic form of the full model is in Figure 1. The Properties of phenomena in the Real World can be sensed by humans and artefacts with varying reliability. Within the Abstract World, two levels are then distinguished. At the Conceptual-Model level, the Real-World phenomena are conceptualised by people, in many cases on behalf of organisations. In the remainder of this paper, capitalised terms are defined in the text, with the ful set of definitions gathered into a Glossary, provided as Supplementary Material. At the Data-Model level, the concepts are operationalised, enabling the Data-Items to represent the states of Properties of phenomena, in a manner that enables coherent col ection, processing, management and use of the Data. At the Conceptual-Model level, Real-World Things and Events are represented as (Id)Entities and Transactions, and their Properties as Attibutes. Relationships exist among the various elements. These aspects of the model provide a useful formalisation of the way in which IS practitioners view the relevant parts of the surrounding world and the representations of that world expressed in their systems. The compound and intentional y attention-attracting term (Id)Entity is used to refer to elements of the Conceptual Model that represent Real-World Things. An Entity corresponds with a Real-World Physical Thing, and an Identity with a Real-World Virtual Thing. Virtual Things lack the corporeal nature of a Physical Thing, but are treated as being real because they perform functions relevant to IS. Virtual Things include processes running inside computing devices, and human presentations or roles. For example, the role of CEO is an Identity that is usually performed by a single human Entity at any give time; but it is performed by different human Entities over time. Virtual Things may have associations with Physical Things, e.g. a computer process with a computing device, a software-agent with a human principal, and a role with one or more humans. Some Physical Things have associations with multiple Virtual Things, and vice versa, and most associations are time-bounded. 668 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Figure 1: The Pragmatic Metatheoretic Model Each element at the Conceptual Level has 'Attributes', which contain an 'Attribute-Value' for each instance. For example, the Entity shipping-containers has Attributes such as colour, owner and type (with Attribute-Values fot type such as refrigerated and half-height). A Relationship has the Attribute of cardinality, reflecting how many of each of the elements that it links can exist – typical y zero, one or many. At the operational, Data-Model level, a Data-Item is a storage-location in which a discrete 'Data-Item-Value' can be represented. For example, Entity-Attributes of cargo-containers may be expressed at the Data Model level as Data-Items and Data- R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 669 Item-Values of Colour = Orange, Owner = MSK (indicating Danish shipping-line Maersk), Type = Half-Height. A set of Data-Items that relate to the same (Id)Entity or Transaction make up a Record. Record-Keys are Data-Items or Data-Item groups that are capable of distinguishing which (Id)Entity-Instance a particular Record relates to. Record-Keys for Identities are cal ed Identifiers, and Record-Keys for Entities are Entifiers. Identification refers to the process whereby a Data-Record is associated with a particular Virtual Thing. It involves the acquisition of a Record-Key, that is to say an Identifier. The corresponding process is Entification, whereby a Data-Record is associated with a particular Physical Thing, involving the acquisition of an Entifier. Some computing devices, being a Physical Thing, have a unique number stored internal y which can be reliably used as an Entifier; whereas each of the processes running within a device and communicating with remote devices is a Virtual Thing, for which a workable Identifer is the combination of IP-Address and Port-Number. For human beings, a customer-code is an Identifier, and a biometric is an Entifier. The (Id)Entity and (Id)Entification distinctions were first proposed in 2001, and have been applied in about 25 articles within the Google Scholar catchment, which together have over 400 citations. However, despite the concept's importance, it has to date attracted far too little attention. This aspect of the pragmatic metatheoretic model represents a major contribution to practice and theory. The basic features of the model can deal with passive Things, including natural objects (such as gems and dinosaur bones), animals, and artefacts (such as objets d'art, pallets, and stock-items). Further articulation is needed when dealing with active Things, including artefacts capable of action in the Real World (such as computing devices and robots), and human beings. The notion of identities of human entities, foundational to the discussion, has been treated elsewhere at length, both general y (e.g. Brown 2020) and specifically within IS (e.g. Clarke 1994b, Halperin & Backhouse 2008, Clarke 2008). In IS that extend beyond data-processing, to information production, to inferencing, to decision-making and even to action, values loom large, conflicts among value-sets occur, and designers must take much greater care. 670 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 2.2 Authentication Theory The pragmatic metatheoretic model outlined in the previous section postulates a Real World comprising Things and Events, which have Properties. These can be sensed by humans and artefacts with varying reliability. Authentication is a process whereby that reliability can be assessed. This paper's purpose is to present an analysis of Authentication in contexts in which Identity or Entity plays a central role. No sufficiently general body of theory has been located in the IS or any other literature. As a basis for the analysis, this sub-section accordingly summarises prior work by the author on the generic notion of Authentication (Clarke 2023). Dictionary definitions of 'assertion' refer to declarations or affirmations, and to the action of making such statements. In the present context: An Assertion is an expression of knowledge about one of more elements of the pragmatic metatheoretic model. An Assertion may be made by a party, implied by context, inferred by a party, or postulated by a party. Assertions may be about: 1. Particular Phenomena in the Real World; 2. Particular elements of an Abstract World; or 3. Relationships between elements in both the Abstract and Real Worlds. It is common in the IT industry for the process of establishing the reliability of an Assertion to be referred to as 'Verification', or sometimes 'Validation'. It is preferable to avoid those terms in favour of one that is consistent with the pragmatic model's recognition of the impracticality of the notion of humanly-accessible truth, and that interprets reliability in constructively loose terms: Authentication is a process that establishes an appropriate degree of confidence in the reliability of an Assertion. In designing Authentication processes, organisations general y select a trade-off among key factors such as cost, reliability, and convenience for, and acceptability to, affected parties. This inevitably results in shortfalls in the quality of the R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 671 Authentication process. The term 'appropriate' has been included in the working definition above, to reflect the fact that the degree of confidence is compromised by, or balanced against, other factors. Assertions in category (1) above, relating solely to the Real World, may be authenticated by empirical means, that is to say by observation of the Phenomena. Assertions in category (2), on the other hand, relate solely to the Abstract World. They may be authenticated against Data available in the Abstract World. Authentication processes for such Assertions may involve the application of reasoning, in order to infer additional Assertions. Classical logic, such as the propositional calculus, is of limited use, because it only supports binary / true-or-false assertions. More useful logics support qualitative data on nominal and ordinal scales, and preferably on ordinal scales (such as the non-linear Richter scale for the intensity of earthquakes), interval scales (with equal distances between consecutive values, cf. Celsius for temperature), and most powerfully on ratio scales featuring a natural zero (cf. Kelvin for temperature – Stevens 1946). Nomatter what approach is adopted, however, Authentication of Abstract-World Assertions cannot, alone, satisfy the criterion of 'establishing an appropriate degree of confidence', because it does nothing to test the relationship between Data and the Real World. The primary focus of Authentication processes needs to be on Assertions in category (3), straddling the two Worlds. Authentication processes depend on evidence, for which dictionary definitions refer to observations adduced in support of a conclusion or statement. In the present context: Evidence is Data that assists in determining the level of confidence in the reliability of an Assertion. An individual item of Evidence is usefully referred to as an Authenticator. A common form of Authenticator is a Document, in any form and expressed in any medium. Some Authenticators carry the imprimatur of an authority, such as a registrar or notary, and are referred to as Credentials. The term Token refers to a 672 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY recording medium on which useful Data is stored, in the present context (Id)Entifiers, Authenticators and/or Credentials. The preliminary sections of this paper have drawn on prior work to define a pragmatic metatheoretic model to support IS practice and practice-relevant research, and to establish basic theory in relation to Assertions and their Authentication. The fol owing section applies those ideas to the specific context of assertions in which Entities and Identities figure prominently. 3 Assertions Relating to (Id)Entity This section applies the theory of Authentication outlined above to Assertions that involve Entities or Identities. The first sub-section identifies the kinds of Assertion that are relevant. The later sub-sections discuss the process whereby such Assertions can be authenticated, and the nature of the Evidence that can be used in that process. 3.1 Introduction The pragmatic metatheoretic model avoids the idea that an accessible truth exists, and articulates the relativistic notion of a degree of confidence in the reliability of Assertions. Reflecting this, a key term in the category descriptions below is defined as follows: Appropriately means with a level of confidence commensurate with the reliance placed on the Assertion and the severity of the consequences if the reliance is misplaced This sub-section identifies a set of Assertion categories. Almost al of the IS and related literature on Authentication is concerned with a limited mainstream mode of thought in which the focus is on what are referred to here as (Id)Entity Assertions – categories (1) and (2) below. R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 673 (1) An Identity Assertion An assertion of a relationship between a Virtual Thing and a Record takes one of the following forms: • A particular Virtual Thing is appropriately associated with one or more Identity-Records 'This client's profile information is displayed on the screen in front of me' • The Data-Item-Values in a particular Identity-Record are appropriately associated with a particular Virtual Thing 'This data on my screen relates to this software-agent' • This Virtual Thing is the Virtual Thing with which this particular Identity- Record is appropriately associated 'This corporation is the corporation we're doing business with' 'This patient is the one to whom this medical record relates' Expression of such Assertions is facilitated by the industry standard Security Assertion Markup Language (SAML), which includes an assertion-type cal ed an 'authentication statement', which asserts that 'a particular remote Virtual Thing is appropriately associated with a particular Identity' (OASIS 2005). (2) An Entity Assertion An Assertion of a relationship between a Physical Thing and a Record takes one of the following forms: • A particular Physical Thing is appropriately associated with one or more Entity-Records 'This felon's profile information is displayed on the screen in front of me' • The Data-Item-Values in a particular Entity-Record are appropriately associated with a particular Physical Thing 'The data displayed on my screen relates to this particular stock-item' • This Physical Thing is the Physical Thing with which this particular Entity- Record is appropriately associated 'This is the shipping-container we were looking for' 674 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 'This person is Wanted Person No.1' A contention of this paper is that a serious deficiency exists in the existing literature in the form of the conflation of Entity and Identity Assertions and their Authentication. An important example is the statement within the IT industry, frequently adopted in the research literature, that Authentication is based on 'what you know, what you have, and what you are'. See, for example, Elgarah & Falaleeva (2005), Witman (2006), Carpenter et al. (2008) and Hewitt (2009). The third category ('what you are') differs fundamental y from the other two. A recent UNHCR plan reported on by Madon & Schoemaker (2021) calls for "a self-managed digital wal et [that] would al ow refugees to store a variety of different forms of identification such as biometric registration, individual ID documentation, attestation card and already-existing authenticated paper documents that have been digitised and uploaded" (p.938, emphasis added). Authenticators for Identity Assertions (such as 'I have this Accredited Refugee Code') are combined into a single Token along with a biometric Authenticator for an Entity Assertion ('I am s/he'). Biometrics strikes through a person's multiple Identities to the Entity. Physical and Virtual Things differ enormously, as do the contexts, the impacts on values, and hence the design processes and the ethicality and public acceptability of those designs. It is therefore very important to distinguish between the two Assertion categories. (3) A Simple Property Assertion Assertion categories (3)-(5) are concerned with Properties. The descriptions here use the abbreviated form (Id)Entity to encompass both sub-categories. A straightforward Assertion takes the following form: • A particular Data-Item-Value in a particular (Id)Entity Record is appropriately associated with, and reliably represents, a particular Property of a particular Thing 'This person is old enough to enter the night club' 'This customer is a frequent-buyer who qualifies for the loyalty discount' 'There are 13 widgets in stock because the inventory system says there are' R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 675 This category of assertion is at an atomic level and of high granularity, and relates to a Property that is represented by a single Data-Item. More complex circumstances are the subject of the fol owing category. The UNHCR plan discussed by Madon & Schoemaker (2021) includes in the envisaged "self-managed digital wallet [for] refugees" the inclusion of "education credentials" (p.938). The Token would therefore assist with the Authentication of Property Assertions (e.g. 'I have a High School Certificate from . .'). A Property Assertion is frequently assumed to be dependent on an accompanying (Id)Entity Assertion; but, subject to some conditions, it is feasible to perform Property Assertion Authentication without (Id)Entity, and may be far preferable to do so (e.g. for cost, confidentiality or privacy reasons). To facilitate expression of Property Assertions, the industry standard SAML includes an assertion-type cal ed an 'attribute statement', which asserts that a particular remote Virtual Thing has a particular Property (OASIS 2005). (4) A Complex Property Assertion A more complex assertion depends on an inference drawn from multiple Data- Items, and takes the following form: • A particular Thing is inferred to have a particular Property, on the basis of multiple particular Data-Item-Values in one or more particular (Id)Entity Records, and on the assumption that those Data-Item-Values are appropriately associated with that Thing and reliably represent that Property 'This borrower is behind on their loan repayments' 'This welfare recipient has been overpaid because they understated their income' Reliance on Property Assertions needs to be based on careful analysis, because each Assertion's reliability depends on data quality, and because of the wide array of quality factors that need to be satisfied to ensure that the level of confidence in the Assertion is appropriate – defined above as meaning 'commensurate with the reliance placed on the assertion and the severity of the consequences if the reliance is misplaced'. 676 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Kambil & van Heck (1998) describe Authentication as one of the five 'Basic Trade Processes' in Online Auctions, used "to verify [1] the quality and [2] features of the product offered [both of which are Complex Property Assertions], [and] [3] the authenticity of the trading parties [a combination of Identity and Attribute Assertions] . ." (p.5). This was revisited in Fairchild et al. (2007), who argued that Authentication is "central to the design of multi-attribute markets [with the market operator performing] extensive pre-qualification of suppliers by intermediary to ensure the integrity of the auction process" (pp.291-292, involving both Identity Assertions and Complex Attribute Assertions). However, where this category is evident in the IS literature, there is often a lack of clarity as to what the Assertion is, or Assertions are, whose reliability is being investigated. One example of simple mis-phrasing is "He pays for this transaction online using BankID to authenticate his payment" (Eaton et al. 2014), when what is meant is 'He authenticates himself to his bank as a person authorised to operate on that account', i.e. it is an Identity not a Value Assertion. Another is the expression 'authenticity of Internet-sourced information' (e.g. Haider 2008), which conflates source (an Identity Assertion) and content (a Fact Assertion). (5) A Principal-Agent Assertion A special case of a complex Property Assertion is of considerable commercial significance, and hence is dealt with here as a separate category. An Assertion relating one Thing's ability to act on behalf of another Thing takes the following form: • A particular Thing has a particular Property, based on one or more particular Data-Item-Values in one or more particular (Id)Entity Records, which are appropriately associated with that Thing and which reliably convey that the particular Thing has the authority to act on behalf of another particular Thing 'This human/legal-person/software-agent is the approved representative of this customer/client/accusee' R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 677 This Assertion-category involves a chain of Virtual and Physical Things, and hence requires multiple Authentication processes to be performed: "The claims of a business intermediary to be acting on behalf of another intermediary need to be subjected to testing. Moreover, the claims of a person to be acting on behalf of a business entity (which may itself be acting as an intermediary for another business entity) also need to be tested. Authentication needs to be undertaken of a particular attribute or credential that reflects the agency relationship, such as a power of attorney, or some other form of delegation of power to sign contracts" (Clarke 1999, p.9). A comprehensive study of Principal-Agent Authentication is in Basul & Muylle (2001). (6) (Id)Entity Match Assertions A further cluster of Assertion Categories are entirely within the Abstract World. Authentication of these categories is not merely valuable, but actually essential to the conduct of IS. They differ from the earlier categories, however, in that Authentication of these Assertions alone, while contributing to the level of confidence, does not satisfy al of the necessary conditions, and hence complementary Assertion Authentication processes are needed, to reliably link these Assertions within the Abstract World with Things in the Real World. Match Assertions take one of the following forms: • This Identity-Record is appropriately associated with this other Identity-Record Within a particular insurance company: 'This record in the Motor Vehicle Insurance database matches to this record in the Home & Contents database' In a government context, and subject to legal authority: 'The record containing this tax-file-identifier matches to the record containing this driver's licence number' • This Entity-Record is appropriately associated with this other Entity-Record 'This description of recovered stolen goods is of the same diamond necklace as this description of stolen goods' Subject to legal authority: 'This DNA sample data from { a crime-scene, a crash-site } is from the same person as is represented by this DNA sample data from a particular family history database' 678 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY • This Identity-Record is appropriately associated with this Entity-Record 'This process is running in this computing device' In a law enforcement context: 'The record containing this client-number corresponds to this fingerprint-based criminal record' The contexts in which matching of human Entities is undertaken often have potential y very serious consequences for the person concerned. This cal s for a very high degree of confidence in the reliability of the Assertion. 3.2 The Authentication Process Assertions are depended upon as a basis for inferencing, decision and action. They need to be authenticated, to protect against risks of error and fraud. Sources of poor quality include: • Accidental mistakes; and • Intentional mistakes, including: • intentional false positives, e.g. masquerade or 'spoofing' to enable a person or process to exercise a power that should only be exercised by some other (Id)Entity; and • intentional false negatives, e.g. avoidance, undermining or subversion of (Id)Entification. Few sources have been located that identify criteria for evaluating the quality of Authentication processes. Zviran & Erlich (2006) identify as relevant factors: effectiveness, ease of implementation, ease of use, and user attitude and acceptance. Way & Yuan (2009) provide a more substantial list, comprising Accuracy, Robustness, User Acceptance, Accessibility, Feasibility, Applicability, Responsiveness, Non-reputability [sic: Non-Refutability] and Maintainability. Those authors also note different priorities for the criteria among stakeholders, which they categorise as Management, IT Support and Users. Both are useful contributions, but are mostly technical in their orientation. Most IS require a socio-technical approach, and hence additional considerations need to be factored in, such as transparency (from user and usee perspectives) and costs and risks (from the viewpoint not only of the system sponsor, but also of users and usees). R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 679 Quality is a substantially greater challenge where other parties are motivated to achieve false positives or false negatives. Safeguards are needed to limit the extent to which such parties may succeed in undermining authentication quality. Techniques such as channel encryption (in particular SSL/TLS) and one-time password schemes are applied to these purposes. Each safeguard has vulnerabilities, and is subject to threats. It is common to distinguish multiple quality-levels or 'strengths' of Authentication, such as unauthenticated, weakly authenticated, moderately authenticated and strongly authenticated. Business enterprises and most government agencies general y adopt risk-managed approaches, accepting lower levels of assurance in return for processes that are less expensive, more practical, easier to implement and use, and less intrusive. 3.3 Evidence in Support of the Authentication Process In section 2.2, the concepts of Authenticators, Credentials and Tokens were introduced. These notions are much applied in the Authentication of Assertions involving (Id)Entity. The central form of such Assertions is category (1), an Identity Assertion, of the form: • A particular Virtual Thing is appropriately associated with one or more Identity-Records Each such association is achieved by means of an Identifier. The conventional term used in government circles for Authenticators designed to support Authentication of such Assertions is 'Proof of Identity' (PoI). This is a disingenuous term, implying infallibility of the Authenticator and the Authentication process that uses it. The notion of accessible truth in such complex circumstances lacks credibility. The appropriate term is accordingly Evidence of Identity (EoI). Several different categories of Authenticator are used as EoI. The notions of 'what you know' (i.e. Data of some kind) and 'what you have' (a Credential or a Token containing one) are useful summaries. A reasonable degree of confidence in an Assertion of Identity can only be achieved, however, if: 680 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY • A Virtual Thing appropriately associated with that Identity can be relied upon to have access to suitable Evidence, to be willing to provide that Evidence, and to do so; and • The party performing the Authentication (or the technology on which that party depends) has access to a copy of relevant Data, or some other means of being satisfied that the proffered Evidence supports the Assertion and is trustworthy. Common examples of knowledge-based EoI are passwords and PINs. Further instantiations are private keys generated by (or sometimes issued to) individuals' own devices (workstations, mobile phones, tablets, smartcards, etc.), and one-time passwords, whether generated by a separate device issued to the relevant individual, or communicated to them at the appropriate time over a separate and secure transmission channel. Tokens are useful y applied to the storage of human-visible and/or machine-readable copies of (Id)Entifiers. The same Token may also be used as EoI, by containing one or more Authenticators, which may be Credentials. Forms of Tokens include sequentially-numbered tickets issued to people required to wait in a queue; a credit-card-sized plastic card carrying a chip, sometimes cal ed a smartcard; machine-readable visual images (such as bar-codes and QR-codes); and machine-readable data-storage (such as a magnetic-stripe, solid-state memory in such artefacts as a thumbdrive or 'USB key', and transmissions from an RFID-tag). In some circumstances, the provision of an Identifier may represent EoI. For example, if a Token is used, and not even the artefact or the individual is aware of the Data-Value that is their Identifier, but the authenticating party (or its technological artefact) knows that Data-Value, the Identifier itself can represent reliable EoI. Generally, however, an Identifier is not a secret, and most schemes use Data-Item(s) other than Identifiers. The next Assertion category (2), an Entity Assertion, is of the form: • A particular Physical Thing is appropriately associated with one or more Entity-Records The Record-Key, in this case an Entifier, by definition reliably distinguishes that Entity-Instance from al other Instances of the same Entity. An Authenticator for R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 681 this purpose, Evidence of Entity (EoE), differs from EoI, in that it is required to provide strong support for the proposition that the Physical Thing is a specific object, artefact or human. An example is a Token installed in a device that provides the relevant Data to, and only to, the party doing the Authentication. Subject to careful design of the EoE creation, instal ation and storage, and of the communications protocols including transmission security features, a considerable degree of confidence can be designed into such a scheme. In the case of living things, a biometric measure may be used. Alternatively, a plant or an animal, including a human being, can be subjected to implantation of a Token in the same manner as instal ation of a Token into an artefact (Michael & Michael 2009). Assertion categories (3) to (5) all involve Property Assertions. Simplifying: • A particular Thing has a particular Property, based on one or more particular Data-Item-Values in one or more particular (Id)Entity Records Two alternative approaches to the Authentication of Property Assertions are possible. Reliance may be placed on Evidence in such forms as Assertions containing sufficient detail that can be checked against one or more other sources (e.g. a claim of a qualification against a testamur, or against a database listing graduates). The alternative approach is for the party performing the Authentication to rely on Data that they already hold. The relevant (Id)Entity Record may contain a directly-relevant Data-Item, such as a flag for 'Trade Customer' or 'Old-Age Pensioner', or for some category of disablement such as 'legal y blind'. In other cases, it wil be necessary for the organisation to apply logical processes to its Data to assess the claim. This applies, for example, to a claim of being owed a refund for a failed delivery; and to having reached a particular age or period of association with an organisation. 682 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3.4 Data Quality's Role in the Authentication Process Little material has been found that identifies criteria for data quality in support of Authentication processes. It is therefore necessary to got back to basic principles and work forward from there. Whether Data is a 'reliable representation' of the relevant phenomena depends on the factors listed in Table 1, comprising Data Quality factors (which are assessable at the time of creation and subsequently) and Information Quality factors (which are assessable only at the time of use). The generic categorisation of quality factors in Table 1 needs to be applied to each circumstance. For example, an Identity Authenticator for a role may be vital to the signing of a large contract, in which case D7 (Temporal Applicability) looms as a high priority for careful checking. Another concern, causing people to be justifiably wary of government agencies sharing data, is that D3 (Appropriate Property Association) and D4 (Appropriate Property Signification) differ greatly, depending on context. The number of children a person declares to an agency depends on that agency's definition of 'a child of a person' (and definitions may even differ between programs administered by the same agency, and over time). The same holds for marital status, and for gender. More complex examples are legion, such as a person's income, which can be measured differently, and can be accumulated and/or averaged over periods as diverse as daily and annually. Quality assurance measures needs to be layered. Safeguards must be designed into authentication processes, and implemented. Controls must be designed, implemented and monitored, to ensure that the safeguards are operational and effective. When concern arises that an Authentication process may have been malperformed or may have delivered a wrong result, it is necessary to conduct an audit. This depends on reviewability, replicability, auditability, accountability, and action. For any of these measures to be feasible, Data needs to be available that document the basis on which each decision was made. R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 683 Table 1: Quality Factors Data Quality Factors Information Quality Factors Assessable when created and later Assessable only at the time of use D1 Syntactical Validity I1 Theoretical Relevance Conformance of the Data-Item-Value with the Demonstrable capability of the Data- Domain on which the Data-Item is defined Item to, in principle, make a difference to the inferencing process in which the Data-Item is to be used D2 Appropriate Phenomenon Association I2 Practical Relevance A high level of confidence that the Data-Item- A demonstrable capability of the Data- Value is associated with the particular Real- Item-Value to, in practice, make a World Thing or Event it is intended to represent difference to the inferencing process in which the Data-Item-Value is to be used D3 Appropriate Property Association I3 Currency A high level of confidence that the Data-Item- The absence of a material lag between a Value is associated with the particular Property Real-World Event and the recording of of the Real-World Thing or Event that it is the corresponding Data-Item-Values intended to represent D4 Appropriate Property Signification I4 Completeness A high level of confidence that the Data-Item- The availability of sufficient contextual Value represents the state of the particular information that the data is not liable to Property of the Real-World Thing or Event it is be misinterpreted intended to represent D5 Accuracy I5 Controls A high level of correspondence of Data-Item- The application of business processes Value with the particular Real-World Thing or that ensure that the Data Quality and Event that it is intended to represent Information Quality factors are satisfied D6 Precision I6 Auditability The level of detail at which the data is captured, The availability of Metadata that reflecting the Domain on which the Data-Item is evidences the Data Quality and defined Information Quality factors D7 Temporal Applicability The absence of ambiguity about the date and time, or period, when the Data-Item-Value represents a particular Real-World Thing or Event Adapted from Clarke (2016), Table 1 This section has articulated the basic theory of Authentication in contexts relevant to identity management activities in IS. It identifies a range of needs. Areas in which IS practice and theory evidence shortfal s include the widespread error of conflating Identity and Entity, failure to be clear about what Assertions need what strength of 684 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Authentication, failure to evaluate threats, and ineffectiveness of relevant safeguards, resulting in inability to rational y select a good-enough approach to Assertion Authentication. 4 Implications The analysis presented above has important implications for IS practitioners, and for practice-oriented IS research. Some of these relate to the quality of IS design, while others reflect the importance of the human-values aspects of IS activities. (1) The Effectiveness of Identity Management Attention has been drawn to the significant differences between Entities that reflect Physical Things and Identities that reflect Virtual Things. The ability of an IS to support organisational activities is undermined if the model on which it is built conflates an Entifier such as a device-id with an Identifier of one of many processes running within that device. In such circumstances, users of a cybersecurity-monitoring system, for example, would be hampered in their efforts to detect runaway, rogue and compromised processes. Similarly, confusion could arise between a non-corporeal principal (such as a corporate trustee) and the individual person or software-process acting as the agent for that principal. In the case of humans, the failure to distinguish Entities from Identities means there is an implicit assumption that a person whose Assertion of Identity is authenticated is necessarily a single, particular instance of homo sapiens. This flies in the face of many real-world practices. Employees routinely share passwords. The Authenticators used by older family members continue to be provided to younger members, so that payments can be made. Commercial terms may wel provide organisations with legal assurance that they can repudiate responsibility for transactions conducted in such ways (if they can gather evidence). On the other hand, al of the organisation's work on risk assessment and risk management, including distinguishing between abuse of commercial terms by the principal and abuse by another party through unauthorised masquerade, lie outside the model, and hence are unsupportable by the IS. R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 685 Another technical problem masked by conventional, inadequate identity management models is the risk of biometrics being compromised, such that convincing biometrics-based masquerade is able to be committed. The Properties of Physical Things that are used for biometrics (fingerprints, iris-patterns, password-capture dynamics, gait, DNA, etc.), are not capable of being tightly protected. They can be captured and replicated, then simulated sufficiently wel that Entity Authentication can be duped. Each advance in liveness-testing stimulates countermeasures. The value of biometrics-based Entification may fal and/or governments may move to limit its use to tightly-controlled, priority circumstances. This would be tantamount to imposing a licensing scheme and heavily-sanctioned legal prohibitions on retention and transmission of biometric data. PIN-pads were designed to prevent capture, retention, replay and transmission of PINs. The same design approach can be applied to biometrics, by implementing a secure stored-biometric measure in a personal card, and instal ing a secure capture and processing module in the authentication device (Clarke 2003a). This has recently become more common in the form of separation of Authentication of an individual's Entity Assertion to their personal device (e.g. mobile phone, tablet) from Authentication to a remote service of the device's Entity Assertion and/or of each process's Identity Assertion. (2) The Effectiveness of Other Business Processes The widespread availability of high-capacity information infrastructure since the 1990s has had impacts on many aspects of organisational and individual activity. One key aspect is greatly increased institutional distance. For example, until the turn of the century, consumer credit business processes embodied a blend of Authentication of both category (2) and category (3) Assertions, because they were 'high-touch' in nature. Since then, most business enterprises have abandoned such labour-intensive approaches, and are now remote from their clients, making their decisions on the basis of a Digital Persona rather than the person (Clarke 1994a, 2014). Moreover, the recent fashions of big data, data analytics and AI/ML, by increasing the volume and diversity of data being used, and decreasing the transparency of the decision process, have greatly increased the scope for erroneous inferences. Authentication of category (3) Assertions are curtailed or dropped, and 686 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY the much weaker form of Authentication of category (2) Assertions relied upon instead, because they are easily automated and inexpensive and quick. These factors have given rise to greatly increased scope for faulty business decisions. These are not limited to the example used in the previous paragraph, consumer credit, but pervade many other application areas as wel . Some of the harm directly affects the organisation sponsoring the system, e.g. through faulty evaluations of loan-worthiness. In many cases, however, the harm is suffered by users and usees. The infliction of harm may not come to the organisation's attention for some time, resulting in reputational risk. Imprecisions in the models underlying identity management schemes exacerbate these risks, and hence far greater care is needed in IS design than in the past. There are few signs, however, that appropriate changes are being made to safeguards, controls and mechanisms for redress. (3) The Economics of IS Design Ineffective identity management can have efficiency and financial impacts. Reduced reputation harms market-share and sales volumes, and hence increases the gross profit margin needed to cover the overheads. In industry sectors in which consumer protection and other regulatory mechanisms exist, the scope exists for customer complaints numbers to rocket, and with that complaints-handling costs. Restitution and redress to individual complainants can be greatly exceeded by even the legal costs of a class action, and swamped by a class action damages settlement. The previous sub-section also noted the concerns of individuals about Entification processes and biometrics, about a consolidated Digital Persona, and about the risks of data-leakage and and manipulation that both give rise to. Identification processes are of particular concern where organisations have the capacity to inter-link multiple of a person's many Identities. This results in tensions between the organisation and individual users, both employees of the sponsoring organisation and external users. Those tensions play out in avoidance, obfuscation, falsification, and lowered ethicality of behaviour and loyalty among individuals with whom the organisation interacts, and in many cases on which it depends. R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 687 Identity Authentication, and particularly Entity Authentication, are expensive, both in terms of financial costs and negative impacts on users and usees. This raises the question as to whether the Authentication of other forms of Assertion may be able to deliver sufficient assurance to an organisation, and to be performed material y less expensively. Property Authentication can be achieved in many contexts without having to perform (Id)Entification and (Id)Entity Authentication processes. In addition, the prior paper which presents a generic theory of authentication identifies several other Assertion-categories that can be effective in satisfying organisations' needs, and can be at the same time more efficient. These are Fact Assertions, Content Integrity Assertions, Liquid-Asset Value Assertions and Non-Liquid-Asset Value Assertions (Clarke 2023). Remarkably few sources have been located that give systematic consideration to the Authentication of Assertions other than of (Id)Entity. Exceptions include Clarke (2001), which argued that "parties [need to] know what is being authenticated" (p.148), and gave examples of Value and Attribute Assertions; Rauniar et al. (2002), which distinguished user authentication, smartcard authentication and card-reader authentication; and Clarke (2003b), which listed "Assertions important to eBusiness" as extending to organisational and artefact Entities and Identities, Attributes, Agency (referred to in this paper as a Principal-Agent Assertion), Location and Value. The absence of serious treatment in the literature is remarkable given the relative ease and inexpensiveness of Authentication of some of the other categories of Assertion, at least in a proportion of the circumstances in which activities are conducted. (4) Stakeholder Interests The primary focus in the first three sub-sections above has been the interests of the organisational sponsor of the particular IS. This final sub-section broadens the view to encompass stakeholders general y. Many circumstances arise in which Authentication of any category of Assertion, and Authentication of (Id)Entity Assertions in particular, involve clashes of values between the system sponsor and individual users and usees. In some of those circumstances, the tension is intrinsic and largely unavoidable. For example, in both the lending and the insurance sectors, the contracting organisation is unlikely to be 688 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY able to protect their financial interest unless key Assertions by the other party can be reliably authenticated. Similarly, government agencies that make transfer payments to the needy have an obligation to authenticate the claims made by each applicant. There are many circumstances, however, in which an appreciation of the ful range of Authentication possibilities may enable the discovery of scope for the organisation to adopt less intrusive process designs. Individuals live their lives in enormously varied circumstances. Organisations that understand particular customer segments may be able to offer alternative Assertion-Authentication channels for, say, the seriously sight- or mobility-impaired, or for victims of domestic violence who are currently taking great care not to disclose their whereabouts. More general y, many people are concerned about the conflation of their separate Identities. Some have reasons that derive from a desire for physical safety, others because of an enclosed disposition, some because of embarrassing personal histories, and of course some for reasons that may deleteriously affect the interests of an organisation they deal with. Al such categories of people are likely to be concerned about the conflation of roles that they perform on behalf of an employer, with their roles on behalf of associations, clubs, their families, and their friends. Some people are even more concerned about the conflation of their Identities with multiple consumer marketing corporations, and even more so as public-private partnerships proliferate, government agency databases are compromised, and a movement gathers steam that began as a gleam in the eye of a Google CEO and Chair: the high-tech, corporatised-government State (Schmidt & Cohen 2013). This risk is inherent in Entifiers, but also increasingly common with Identifiers, where governments and consumer marketing corporations alike invest heavily in the consolidation of personal data into a singular Digital Persona. Many individuals are very concerned about the enablement of surveil ance and manipulation, the creation of 'honey-pots' that attract third parties, the inevitable abuse of insider privileges, and leakage of sensitive personal information. Arguments can be advanced for organisations to take individuals' interests into account for ethical, consumer rights or social responsibility reasons. Leaving those aside, however, organisational self-interest may wel be best served by recognising the extent to which individual user behaviour, especial y among employees and contractors, and usee behaviour, may not be what the organisation prefers. Specifically, individuals' behaviour may include objections, non-adoption, non- R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 689 compliance, avoidance, obfuscation, falsification, quiet sabotage, and whistleblowing. These drive up an organisation's costs, and drive down its internal morale and its external reputation. Opportunities exist to apply insights from the model and analysis presented in this paper to envision and implement satisfactory forms of Authentication that are less burdensome and intrusive for individuals and less expensive and otherwise harmful for organisations. The focus of this paper has been, throughout, on IS practice and practice-relevant IS research. Mention also needs to be made, however, of the broader issues that arise from the application of the pragmatic metatheoretic model to (id)entity. The axiological aspects of the model highlight the inadequate attention paid by both IS practice and research to values, and particularly the interests of users and usees. Further work is needed to draw out the implications of this work for professional y responsible IS design. 5 Conclusions This paper has built on a previously-published pragmatic metatheoretic model, that is intended to reflect the Real World and the Abstract Worlds as they are understood, manipulated and applied in IS practice and practice-oriented IS research. It has drawn further on a previously-published extension of that model into the identity management space, and a generic theory of Authentication of Assertion categories relevant to IS. The first contribution of this paper is a body of theory that identifies categories of Assertion that involve (Id)Entity, outlines key requirements of processes for the Authentication of those Assertion categories, describes the forms that may be taken by Evidence that can support those processes, and provides a framework for the evaluation of Data Quality and Information Quality of that Evidence. A second area of contribution is the discussion of the implications of this body of theory for IS practice, and for the organisations that depend on the products of IS practitioners. Proposals are made for improvements in effectiveness and efficiency within the frame of reference used by the organisations that utilise IS. Further proposals are made that reflect the somewhat different worldviews of those 690 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY organisations' employees, customers and suppliers, and of usees, who are affected by but do not directly participate in the system. The analysis presented in this paper also identifies a number of weak spots in IS and practice, and points towards ways to address them. The most significant weakness, visited at multiple points in the paper, is the conflation of the concepts of Identity and Entity, and how that plays out in faulty business models and ineffective, inefficient and unpopular business processes. Another area in which considerable scope exists for insightful work is the ways in which organisations, as Virtual Things, are represented within IS models, and how the reliability of Assertions relating to them can be evaluated. For example, see Eriksson & Ågerfalk (2022). Beyond these implications for IS practice, new directions for practice-oriented IS research are opened up. More illustrations and applications of the model will deliver deeper insights that offer value in particular contexts. Some industry sectors deal in natural objects and passive artefacts, both physical and virtual, whose Properties may demand refinements and extensions of a model and theory that are intentional y somewhat generic. Active artefacts, again both physical and virtual, are increasingly being conceived and injected into IS. There is current controversy concerning the extent to which active artefacts can reliably be delegated the power to infer, decide and act in isolation, or need to be oriented towards decision support roles, working closely with humans. The elements of the above analysis of (Id)Entification, and of the Authentication of (Id)Entity Assertions, provide what may prove to be valuable perspectives on that problem-domain. Acknowledgements This work reflects feedback from many colleagues who have participated in seminars and panels, and formal y and informal y reviewed the many working papers and refereed papers on which the present work is built. It has also benefited from the thoughtful and informed comments of the Bled Conference reviewers. R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 691 References Abbas R. & Michael K. (2022) 'Socio-Technical Theory: A review' In S. Papagiannidis (Ed), 'TheoryHub Book', TheoryHub, 2022, at https://open.ncl.ac.uk/theories/9/socio-technical-theory/ Altinkemer K. & Wang T. (2011) 'Cost and benefit analysis of authentication systems' Decision Support Systems 51 (2011) 394-404 Avison D. & Fitzgerald G. (2006) 'Information Systems Development - Methodologies, Techniques & Tools' McGraw Hill, 4th ed., 2006 Basul A. & Muylle S. (2001) 'Achieving Authentication in Electronic Markets: A Principal-Agent Perspective' Proc. 8th Research Symposium on Emerging Electronic Markets, Maastricht, The Netherlands, September 2001, at https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=538ce0bbebb44ce0159 95e3e5122092a3fa93171 Baumer E.P.S. (2015) 'Usees' Proc. 33rd Annual ACM Conference on Human Factors in Computing Systems (CHI'15), April 2015, at http://ericbaumer.com/2015/01/07/usees/ Berleur J. & Drumm J. (Eds.) (1991) 'Information Technology Assessment' Proc. 4th IFIP-TC9 International Conference on Human Choice and Computers, Dublin, July 8-12, 1990, Elsevier Science Publishers (North-Holland), 1991 Brands S.A. (2000) 'Rethinking Public Key Infrastructures and Digital Certificates: Building in Privacy' MIT Press, 2000 Breward M., Hassanein K. & Head M. (2017) 'Understanding Consumers’ Attitudes Toward Controversial Information Technologies: A Contextualization Approach' Information Systems Research 28,4 (2017) 760-774 Brown A.D. (Ed.) (2020) 'The Oxford handbook of identities in organisations' Oxford University Press, 2020 Carpenter D.R., McLeod A.J. & Clark J.G. (2008) 'Using Biometric Authentication to Improve Fire Ground Accountability: An Assessment of Firefighter Privacy Concerns' Proc. AMCIS 2008, at 11 Chaum D. (1985) 'Security Without Identification: Transaction Systems To Make Big Brother Obsolete' Communications of the ACM 28, 10 (October 1985) 1030-1044, at https://dl.acm.org/doi/pdf/10.1145/4372.4373 Clarke R. (1992) 'Extra-Organisational Systems: A Chal enge to the Software Engineering Paradigm' Proc. IFIP World Congress, Madrid, September 1992, PrePrint at http://www.rogerclarke.com/SOS/PaperExtraOrgSys.html Clarke R. (1994a) 'The Digital Persona and its Application to Data Surveillance' The Information Society 10,2 (June 1994) 77-92, PrePrint at http://www.rogerclarke.com/DV/DigPersona.html Clarke R. (1994b) 'Human Identification in Information Systems: Management Chal enges and Public Policy Issues' Information Technology & People 7,4 (December 1994) 6-37, PrePrint at ://www.rogerclarke.com/DV/HumanID.html Clarke R. (1999) 'Identified, Anonymous and Pseudonymous Transactions: The Spectrum of Choice' Proc. Conf. User Identification & Privacy Protection Conference, Stockholm, June 1999, PrePrint at http://www.rogerclarke.com/DV/UIPP99.html Clarke R. (2001a) 'The Fundamental Inadequacies of Conventional Public Key Infrastructure' Proc. Conf. ECIS'2001, Bled, Slovenia, 27-29 June 2001, PrePrint at http://www.rogerclarke.com/II/ECIS2001.html Clarke R. (2001b) 'Authentication: A Sufficiently Rich Model to Enable e-Business' Xamax Consultancy Pty Ltd, 19 October 2001, at http://www.rogerclarke.com/EC/AuthModel.html011019.html Clarke R. (2002a) 'Personal Notes on Computers, Freedom & Privacy 2002' CFP'22, San Francisco, Xamax Consultancy Pty Ltd, 16-19 April 2002, at http://rogerclarke.com/DV/NotesCFP02.html 692 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Clarke R. (2002b) 'Why Do We Need PKI? Authentication Re-visited' Xamax Consultancy Pty Ltd, Presentation at the 1st Annual PKI Research Workshop, at NIST, Gaithersburg MD, April 24-25, 2002, at http://www.rogerclarke.com/EC/PKIRW02.html Clarke R. (2003a) 'The Scope for Privacy-Sensitive Biometric Architecture' Xamax Consultancy Pty Ltd, February 2003, at http://www.rogerclarke.com/DV/BioArch.html Clarke R. (2003b) 'Authentication Re-visited: How Public Key Infrastructure Could Yet Prosper' Proc. 16th Bled eCommerce Conf., June 2003, PrePrint at http://www.rogerclarke.com/EC/Bled03.html Clarke R. (2004) 'Identity Management: The Technologies Their Business Value Their Problems Their Prospects' Xamax Consultancxy Pty Ltd, March 2004, 66 pp., at http://www.xamax.com.au/EC/IdMngt-Public.pdf Clarke R. (2008) 'Dissidentity: The Political Dimension of Identity and Privacy' Identity in the Information Society 1,1 (December, 2008) 221-228, PrePrint at http://www.rogerclarke.com/DV/Dissidentity.html Clarke R. (2009) 'A Sufficiently Rich Model of (Id)entity, Authentication and Authorisation' Proc. IDIS 2009 - The 2nd Multidisciplinary Workshop on Identity in the Information Society, LSE, London, 5 June 2009, at http://www.rogerclarke.com/ID/IdModel-090605.html Clarke R. (2014) 'Promise Unfulfilled: The Digital Persona Concept, Two Decades Later' Information Technology & People 27, 2 (Jun 2014) 182-207, PrePrint at http://www.rogerclarke.com/ID/DP12.html Clarke R. (2016) 'Big Data, Big Risks' Information Systems Journal 26, 1 (January 2016) 77-90, PrePrint at http://www.rogerclarke.com/EC/BDBR.html Clarke R. (2021) 'A Platform for a Pragmatic Metatheoretic Model for Information Systems Practice and Research' Proc. Austral. Conf. Infor. Syst. (ACIS), December 2021, PrePrint at http://rogerclarke.com/ID/PMM.html Clarke R. (2022) 'A Reconsideration of the Foundations of Identity Management' Proc. 35th Bled eConference, June 2022, pp.1-30, PrePrint at http://rogerclarke.com/ID/IDM-Bled.html Clarke R. (2023) ''A Generic Theory of Authentication to Support IS Practice and Research' Xamax Consultancy Pty Ltd, January 2023, at http://rogerclarke.com/ID/PGTA.html Clarke R. & Davidson R.M. (2020) 'Through Whose Eyes? The Critical Concept of Researcher Perspective' J. Assoc. Infor. Syst. 21, 2 (March-April 2020) 483-501, PrePrint at http://rogerclarke.com/SOS/RP.html Coneybeare F.C. (1892) 'Professor Clifford on the Soul in Nature' The Monist 2,2 (January 1892) 209-224, at https://www.jstor.org/stable/27896942 Eaton B., Hal ingby H.K., Nesse P.-J. & Hanseth O. (2014) 'Achieving Payoffs from an Industry Cloud Ecosystem at BankID' MIS Quarterly Executive 13,4 at 6 Elgarah W. & Falaleeva N. (2005) 'Adoption of Biometric Technology: Information Privacy in TAM' Proc. AMCIS 2005, at 222 Eriksson O. & Ågerfalk P.J. (2022) 'Speaking things into existence: Ontological foundations of identity representation and management' Information Systems Journal 32,1 (2022) 33-60, at https://onlinelibrary.wiley.com/doi/full/10.1111/isj.12330 Fairchild A.M., O'Reilly P., Finnegan P. & Ribbers P.M. (2007) 'Multi-Criteria Markets: An Exploratory Study of Market Process Design' Electronic Markets 17,4, (Nov 2007) 286-297, at http://www.electronicmarkets.org/fileadmin/user_upload/doc/Issues/Volume_17/Issue_04/V17I4_M ulti-Criteria_Markets__An_Exploratory_Study_of_Market_Process_Design.pdf Fischer-Huebner S. & Lindskog H. (2001) 'Teaching Privacy-Enhancing Technologies' Proc. IFIP WG 11.8 2nd World Conf. on Information Security Education, Perth, Australia Gottwald S. (2001) 'A Treatise on Many-Valued Logics', January 2001, at https://www.researchgate.net/profile/Siegfried-Gottwald/publication/259645593_A_Treatise_on_Many- Valued_Logics/links/00b7d5324ce793473c000000/A-Treatise-on-Many-Valued-Logics.pdf Haider A. (2008) 'Believable Unbelievable Internet Based Information' Proc. ACIS 2008, at 93 R. Clarke: The Theory of Identity Management Extended to the Authentication of Identity Assertions 693 Halperin R. & Backhouse J. (2008) 'A roadmap for research on Identity in the information society' Identity in the information society journal 1,1 (2008) 71-87, at https://link.springer.com/article/10.1007/s12394-008-0004-0 Hewitt B. (2009) 'Using a Hybrid Technology Acceptance Model to Explore How Security Measures Affect the Adoption of Electronic Health Record Systems' Proc. AMCIS 2009, at 328 IETF (2022) 'RFCs' Internet Engineering Task Force (IETF), December 2022, at https://www.ietf.org/standards/rfcs/ Kambil A. & van Heck E. (1998) 'Reengineering the Dutch Flower Auctions' Information Systems Research 9,1 (March 1998) 1-19 Magnusson A. (2022) 'The Definitive Guide to Authentication' Strong DM, September 2022, at https://www.strongdm.com/authentication Michael K. & Michael M.G. (2009) 'Innovative Automatic Identification and Location-based Services: From Bar Codes to Chip Implants' Information Science Reference, 2009 Mumford E. (2006) 'The story of socio-technical design: reflections on its successes, failures and potential' Info Systems J 16 (2006) 317-342, at https://executiveinsight.typepad.com/files/the-story-of-socio-technical-design.pdf Nyst C., Makin P., Pannifer S. & Whitley E. (2016) 'Digital identity: Issue analysis: Executive summary' Consult Hyperion, 2016 OASIS (2005) 'SAML V2.0 Standard' OASIS, March 2005, at https://wiki.oasis-open.org/security/FrontPage Rauniar R, Rauniar D., Shakya S. & Urcuyo C. (2002) 'A Mutual Authentication Scheme for Low Cost Smart Card Applications' Proc. AMCIS 2002, at 265 Ryan T.W. & Marett K. (2010) 'The Influence of Experiential and Dispositional Factors in Phishing: An Empirical Investigation of the Deceived' Journal of Management Information Systems 27,1 (Summer 2010) 273–303 Schmidt E. & Cohen J. (2013) 'The New Digital Age: Reshaping the Future of People, Nations and Business' Knopf, 2013 Stevens S.S. (1946) 'On the Theory of Scales of Measurement' Science 103, 2684 (7 June 1946), at https://psychology.okstate.edu/faculty/jgrice/psyc3120/Stevens_FourScales_1946.pdf Way S.C. & Yuan Y. (2009) 'Criteria for Evaluating Authentication Systems' Proc. AMCIS 2009, at 338 Witman P. (2006) 'Anti-Phishing Strong Authentication Technology Options' Proc. AMCIS 2006, at 131 Zviran M. & Erlich Z. (2006) 'Identification and Authentication: Technology and Implementation Issues' Commun. Assoc. Infor. Syst. 17,4 (2006) 90-105 694 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Supplementary Materials Glossary, at http://rogerclarke.com/SOS/FDI.html#G 'The Authentication of Assertions Relating to (Id)Entity', a comprehensive Working Paper from which the present paper is drawn, at http://rogerclarke.com/ID/IEA.html DECISION ANALYTICS - A POSITION PAPER CHRISTER CARLSSON Abo Akademi University, Institute for Advanced Management Systems Research, Turku, Finland christer.carlsson@abo.fi The context we address is the ‘digital or new economy’ for which we propose that Decision Analytics wil be one of the key drivers. The reasons are that we need to both meet the chal enges from big data/fast data and to work out new possibilities to make experience and expert knowledge accessible and usable for local, ad hoc decision makers and for automated, intel igent systems. Digitalisation brings increasing competition, slimmer margins for productivity and profitability and more pronounced requirements for effective planning, problem solving and decision making. This requires a transfer of (sometimes tacit) knowledge from experts and experi-enced people to novice Keywords: system operators—and to automated, intelligent systems—a analytics, digitalisation, transfer we call knowledge mobilisation. We will work out knowledge reasons for why Decision Analytics wil be a key part of mobilisation, knowledge mobilisation and an essential contribution to the Bled eConference, development of instruments we need for the progress of research digitalisation. paper DOI https://doi.org/10.18690/um.fov.4.2023.42 ISBN 978-961-286-751-5 696 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction INFORMS defines Operational Research and Management Science (OR/MS) as “the scientific process of transforming data into insights to making better decisions”. OR/MS was the forerunner to analytics. In a recent report cal ed “Competing in 2020: Winners and Losers in the Digital Economy” [16] Harvard Business Review worked out the impact digitalisation will have in some key industrial sectors. Among the respondents 16% stated that their companies are digital (most products/operations depend on digital technology), 23% that they are non-digital (few if any products/operations depend on digital technology), and 61% that they are hybrid (some products/operations depend on digital technology). The report found a significant performance gap between digital leaders (“digi-tals”) and the rest (“non-digitals”): 84% of the digitals use big data and analytics, but only 34% of the non-digitals; 51% of the digitals use cognitive computing/AI, but only 7% of the non-digitals; the digitals have data science and data engineering on staff (62%), the non-digitals much fewer (20%); all professionals working for the digitals have the ability to work with and make sense of data and analytics (76%), not that common for the non-digitals (30%). The insight is that a strong analytics capability is key to digital business—companies that want to compete in the digital economy wil have to invest in people, processes and technology that offer access to data and knowledge and skills in analytics. One more detail—artificial and machine intel igence appear as key interests and concerns among the business leaders [16]; the formulation is that “future success wil depend on the successful collaboration between human and machine intelligence”. Digitalisation is bringing big data (or “fast data” for streaming big data) which more recently has been claimed to make it impossible to use analytics as huge amounts of data make the algorithms impossible or impractical to use [4, 8]—or it will take too much time as fast decision making in almost real-time is a necessity in the digital economy (“the fast eat the slow” as the slogan goes). If there is time to make bad decisions (by guessing instead of using effective (but more demanding) instruments) C. Carlsson: Decision Analytics - A Position Paper 697 there should also be time to make good decisions—we only need to know how and what instruments to use. The following tale offers some insight on analytics – it offers alternatives to approaches based on closed, secret and mysterious data sets. Kahneman [18] relates the case of Orley Ashenfelter, a Princeton economist and wine lover, who wanted to find a way to predict the future value of fine Bor-deaux wines from information available in the year they are made. The experts taste the wine and use decades of experience and insight in the wine markets to decide future values. Ashenfelter, an economist, used multiple regression analysis and statistics tools as he had no possibility to actually taste exclusive Bordeaux wines. He collected statistics on London auction prices for select mature red Bordeaux wines 1990–1991. The quality of Bordeaux wines was found to be decided by (i) the age of the vintage, (ii) the average temperature over the growing season (April–September), (i i) the amount of rain in September and August (less rain gives better wine), and (iv) the amount of rain preceding the vintage (October–March). These four factors are al measurable, published and easily verifiable —tasting wines and making judgments has of course its benefits. Aschenfelter built a regression model with the four factors on vintages 1952–1980, which turned out to explain about 80% of the variation in the average price of Bordeaux wine vintages! We should notice that Aschenfelter is a professional—he avoided the fal acy of smal samples and made sure that he had observations on large selections of wine over 10 years from six major century-old Bordeaux chateaux and limited his models to Bordeaux to reduce the number of external (but actual y irrelevant) factors. Aschenfelter, in fact, fol ows what we describe as time-tested, good analytics practice: use facts and data that can be tested and verified, methods that can be validated for repeated use to work out insight that is useful, valid, and verifiable. In Section 2 we wil bring out some experience from the forest industry to show why it makes sense to take a couple of analytical steps beyond visualisation to gain insight; in Section 3 we show how Decision Analytics methods identify the core of a decision problem and clean out secondary problem elements that confuse the issues; in Section 4 we have summarized some points on what is required of Decision Analytics as a developed and effective approach for the decision context of the ‘digital or new economy’; Section 5 collects a summary and some conclusions. 698 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 2 Lessons from the Forest Industry The demand for fine paper products is slowly declining at 2-3% per year but has been rather stable for the last 10 years. Nevertheless, a paper mil we worked with showed significant variations in orders coming from the supply chain, variations that appear to be random and unexplainable. The managers had tried to use optimisation models, which did not work – they even suspected that the optimisation was part of the problem. One of the managers mentioned that he had heard about a “bul whip effect” as a possible description of the supply chain problems, and that it involved some fairly tough mathematics. Lee et al [19, 20] carried out some early, more systematic theoretical work and focused on distorted information as a driver of the bul whip effect. They found that the bul whip effect describes increasing variance in orders as they move up a supply chain (from customer trough retailer and wholesaler to producer), even when underlying customer demand shows only a smal or even negligible variance. The reason seems to be that the retailer “improves” on the orders that customers place; there are many customers, and their demand estimates show variances that are judged not to be reliable – “they are not professionals on market dynamics”. The retailers make their own demand estimates from “im-proved” customer estimates; the wholesaler gets orders from many retailers that show variations in estimated demand and “improves” on retailer orders when placing orders with the producer (the paper mill). Lee et al [19, 20] show that estimates over time of the actual demand get distorted as orders move up the supply chain. The intuition is that this cannot be a big problem, but the paper mil managers listed several problems, that were later confirmed by the literature (cf. Lee et al. [19, 20], Carlsson-Fullér [2, 3]). The supply chain actor “improvements” build excessive inventory as the actors safeguard themselves against the variations. The safeguards wil cause some part(s) of the supply chain to run out of products as actors overestimate their safety margins when deciding on orders; local shortages result in poor customer service and shortages result in lost revenue. Lost revenue translates to substandard productivity of capital allocated to operations. C. Carlsson: Decision Analytics - A Position Paper 699 On the corporate level the storyline is a bit different from the supply chain operations because of aggregated data: demand variations cause variations in the logistics chain and the planned use of transportation capacity; ad hoc changes wil result in suboptimal transportation schemes and increase transportation costs; demand fluctuations caused by the bul whip effect may unnecessarily change optimal production schedules, which shows up (much later) as increased production costs. Lee et al [19, 20] found four reasonable operations that active supply chain actors could undertake to tackle variations. The first operation is to update demand forecasts when data shows that next period demands wil be different. Forecasts are built with time series analysis of historical demand patterns from immediate customers. Then, only retailers build on actual customer demand patterns, the other actors adjust to (perhaps unmotivated) fluctuations in the orders of preceding actors. It appears that safety stocks, which are popular smoothing instruments, also will amplify the bullwhip effect [20] as they are optimal locally and subjectively for one actor but send the wrong signals up-stream. The second operation is order batching: periodic ordering and push-ordering. The costs for frequent order processing may be high and attract customers to optimised periodic ordering schemes - which in most cases wil destroy custom-er demand patterns. Standard MRP-systems use analytics models to decide optimal order size and frequency, and to activate periodical orders. Push-ordering occurs as upstream actors (working for producers, wholesalers, retailers) launch “special offers” to induce (non-optimal) out-of-period orders. This, again, contributes to variance in customer orders and destroys actual demand patterns, which contributes to the bullwhip effect. The third operation builds up on price variations. The producers initiate and control price changes both in the long- and short term. Customers are encour-aged to buy in larger quantities by attractive offers on quantity discounts, in special price campaigns, through coupons or rebates. Then the buying patterns do not necessarily reflect consumption patterns - customers buy in quantities which do not reflect their needs. This will initiate and amplify the bul whip effect. 700 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Rationing and shortage gaming drives the fourth operation, which is initiated if/when demand exceeds supply. If the producers, even once, have met shortages with rationing of customer deliveries, customers start to exaggerate their real needs if there is fear that supply will not cover demand. This starts a bullwhip effect, which wil grow if customers are al owed to cancel orders when their real demand is satisfied (and their gaming gets no downside cost). These four causes/drivers of the bul whip effect may be hard to monitor, and even harder to control in the industry. They may interact, and/or act in concert, and the resulting combined effects are not clearly understood, neither in theory nor in practice, which offer chal enges for tackling them with decision analytics modelling. In the real business case seasoned managers recognized the drivers of the bul -whip effect from their own experience and worked out some practical (producer) solutions: (i) share information with al supply chain down-stream actors; (ii) build channel alignment of pricing, transportation, inventory planning and ownership (if legal within antitrust legislation); and (i i) reduce order processing costs and shorten lead times to improve on operational efficiency. These (practical) steps make sense but there are two chal enges - to adapt them to EU regulations controlling cartel-building and to develop decision analytics instruments to find optimal solutions. In section 3 we wil show how Decision Analytics methods identify the core of a decision problem and find solutions. 3 Analytics Modelling Guides Problem Solving We wil stay with the supply chain context and focus on the retailer-wholesaler stage as a retailer reacts to the actual demand from customers. We wil add to the context description: consider a multiple period supply process where demand is non-stationary over time and demand forecasts are updated from observed demand. Assume that the retailer gets orders representing a much higher demand in one period, interprets it as a signal of increasing demand in the future, adjusts demand forecasts for future periods, and places a larger order with the wholesaler. The demand is non-stationary, and an optimal ordering policy should also be non-stationary; non-stationary ordering increases the variance of the orders which starts C. Carlsson: Decision Analytics - A Position Paper 701 the bulwhip effect. Another factor is the lead-time between the ordering point and the point of delivery; if this is long, uncertainty increases and gets the retailer to add some “safety margin” to the order, which increases the variance and adds to the bul whip effect. For analytics model ing we wil simplify the context even further by focusing on a single product, (the models can be extended to multiple items and to batches of products), and inventory for multiple periods. The process has the following structure: at the beginning of period t, the retailer decides to order a quantity zt. This is cal ed the decision point for period t. Next, goods ordered ν periods ago arrive. Retailer demand is fulfilled, and the available inventory is used to meet customer demand. Excess demand is backlogged until the next decision point. Lee at al [20] (cf. also [2, 3]) assume that the retailer faces serial y correlated demands which follow the stochastic process, Dt = d + ρDt−1 + ut where Dt is the demand in period t, ρ is a constant satisfying −1 < ρ < 1, and ut is a random variable, normal y distributed with zero mean and variance σ2. Here σ2 is assumed to be significantly smaller than d, a “usual” level of demand presumed to exist at any t, so that the probability of a negative demand is very small. The use of d is technical to avoid negative demand, which wil destroy the stochastic process and make the analytics model useless. The order quantity, which is found with a cost minimization model, is an optimal ordering policy and sheds some new light on the bul whip effect. The effect gets started by rational decision making, which decides a precise order amount and time for delivery; there is not much hope to avoid the bul whip effect by changing the ordering policy, as it is difficult to motivate people to act in an irrational way. As an optimal, precise ordering policy drives the bul whip effect we decided to modify the policy with imprecise order amounts that can be made more specific as the time of delivery gets closer. The order amounts can be intervals, which wil be made more precise over time (communicating over some joint digital platform). 702 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY We worked out such a policy where intervals replaced the precise order amounts; for this we need to add a few more instruments to analytics model ing, fuzzy numbers. A fuzzy number A is a fuzzy set [5] of the real line ℝ with a normal, (fuzzy) convex and continuous membership function of bounded support. In [1] we proved a theorem that fuzzy subsets entail smal er variance (Var): let A, B ∈ ℱ, a family of fuzzy sets, with A ⊂ B. Then Var(A) ≤ Var(B). Then, if we develop better and better estimates of future sales in period t, Dt, we can reduce the variance of zt by replacing the rule for optimal order amount with an adjusted rule. Fuzzy subsets entail smaller variance and Var(zit) < Var(z*t), i.e., the variance of zit will get smaller as Dt(i) gets closer to the point of order delivery, and the bul whip effect can be eliminated. This was now carried out for demand signal processing; fuzzy (imprecise) numbers can be applied also to price variations and - with some more model ing efforts - to cases with rationing games and order batching. The analogy with Kahneman and Ashenfelter [18] is that we found some simple, fact-based mechanism that works with data that can be tested and verified, and that we found to drive the bul whip effect. We also noted that the four practical operations – demand forecast updates, order batching, price variations and shortage gaming – in effect mostly add complexity to the storyline without offering effective means to reduce or eliminate the bul whip effect from supply chain operations. This il ustrates Kahneman’s point that “experts try to be clever … to work with (too) complex combinations of features”. 4 An Agenda for Decision Analytics Ciancimino et al [15] identifies distorted demand information – as we found out in sections 2-3 – as a key driver; in addition to this they also single out disintegrated material flow and lack of replenishment rule alignment. This resulted in a shift of focus to bul whip avoidance [15] and implementation of supply chain collaboration practices. These include alignment of planning, forecasting and replenishment systems among partners, which is made possible with digital exchange of information. This new idea is described as a Synchronised Supply Chain (SSC) [15]; it could of course not apply to industries with strong (EU) regulations that guide and secure open competition in al parts of the supply chain. On the other hand, the practical measures we col ected from the paper mil supply chain – (i) share C. Carlsson: Decision Analytics - A Position Paper 703 information downstream, (ii) channel alignment of pricing, transportation, inventory planning and ownership (if legal at least formal y) and (ii ) technical steps to improve operational efficiency (order processing costs, lead times) – are in line with the SSC archetype. On the other hand, the imprecise order amounts that become more specific as the time of order delivery gets closer are not directly reducing competition in the supply chain and should not conflict with EU regulations. The sequence of order specifications can be done with secure digital platforms that SC partners share. Transfer of accumulated expertise on how to manage large, complex, and dynamic processes from senior to novice engineers is a classical problem in industry. The transfer needs to be done as experienced engineers retire or decide to leave the company or are let go to save on salary costs. Digitalisation brought a new requirement—the accumulated expertise should somehow be transferred to automated intel igent systems. Experience that we gained from industrial cases [7, 9, 17] shows that expert insight is not easily transferred to novice replacements (or automated systems) with much less or no experience of the processes. We cal ed this process knowledge mobilisation [6, 22] and worked out a first attempt with analytics model ing [7]. We noted that analytics has not gained much support in the last couple of decades (we have found several “common wisdom” (CW) claims as explanations, but which appear not to have support in research results). The first CW-claim states that mobilised knowledge will be too limited, that simplifications clean out most of the insights on offer from a real-life context. The second CW-claim finds that experienced engineers do not work with mathematical programming models. The third CW-claim notes that mathematical programming models gives an impression of precise insight and knowledge; models come from simplifying assumptions that clean out ambiguities and imprecision. The fourth CW-claim states that there are better ways to mobilize tacit knowledge with experienced engineers than using analytics. There are, however, valid counter arguments to the claims in actual cases [21, 22]. If simplifications are made properly, they will single out core functions in interacting processes and reduce complexity. 704 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY The paper machine case [21] offered some insight on how to work out chalenges that digitalisation has introduced. Our contention was that digitalisation should build on joint human/system reasoning that combines experience, insight, intuition, social interaction, etc. with support produced by automatic, intelligent systems, which in the paper machine case were combinations of mathematical optmisation tools and fuzzy ontology [21]. We assumed and believed that this would work and managed to develop some test cases [22]. A partial answer builds on digital coaching [10, 12, 13] that offers instruments for knowledge transfer from experienced users to novice users and from knowledgeable users to automated systems. It appears that decision analytics could/should be the optimisaion part of the digital coaching. There is a context understanding for the 2020’s and an emerging research position for Decision Analytics that we can summarize in a few statements (cf. also [14]); these descriptions wil no doubt change and be updated as our understanding of digitalisation and digital economy increases and improves, but the starting point builds on the proposals {4.1-4.4}: {4.1} In a context of digital disruption and quickly growing competition in the digital economy we have in-creasing dynamic and real-time processes, which require rapid and timely problem solving and decision making in an environment of large and growing sets of giga-data that contribute to increasing and difficult to tack-le complexity. Decision Analytics develops and of ers instruments for insightful and concise representation of problems in the digital economy. {4.2} Management theory and research, combined with long-time experience, guide problem-solving and decision making in the digital economy. Digital Analytics develops and/or finds alternative and simpler forms for problem solving and decision making that also of er more insight. {4.3} Complex and competitive environments (like the supply chains of major industries) invite rapid problem solving with locally available data and supported with ad hoc insight through digital platforms and applications. Digital Analytics will search for and/or work out methods that adapt to both (optimal) local and ad hoc problem-solving as wel as to the general and generic principles for intel igent, automated systems. C. Carlsson: Decision Analytics - A Position Paper 705 {4.4} Knowledge mobilisation offers and distributes insight and knowledge in complex and competitive environments in the digital economy. Digital Analytics forms the core and common language for knowledge mobilisation over digital platforms. 5 Summary and Some Conclusions We propose that Decision Analytics wil be one of the key drivers for what is cal ed the ‘digital or new economy’ (cf. also [11]). Digitalisation brings increasing competition, slimmer margins for productivity and profitability and more pronounced requirements for effective planning, problem solving and decision making. We worked out reasons for why Decision Analytics wil be essential for meeting the chal enges from big data/fast data and for working out new possibilities for effective problem-solving and decision making based on experience and expert knowledge. Artificial and machine intel igence appear as key interests and concerns among the business leaders; the formulation is that “future success wil depend on the successful collaboration between human and machine intelligence”. We introduced “knowledge mobilisation” as instrumental for the human-machine collaboration; Decision Analytics is proposed as a key component in this collaboration. We introduced a study we carried out with a forest products corporation. The demand for the fine paper products was basical y rather stable but a paper mil showed significant variations in orders coming from the supply chain, variations that appear to be random and unexplainable. The managers had tried to use optimisation models, which did not work – they even suspected that the optimisation was part of the problem. One of the managers mentioned that he had heard about “the bul whip effect” but added that “nobody has been able to work out the bul whip effect for fine paper supply chains” a nice chal enge for analytics researchers. The work with the paper mil focused on finding some simple, fact-based mechanism that works with data that can be tested and verified to drive the bul whip effect. The principle found was to build up orders as intervals (actual y, as fuzzy numbers) that give flexibility to adapt to changes in downstream demand and wil be made more precise as a delivery time gets closer. We also found that four practical operations – demand forecast updates, order batching, price variations and shortage 706 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY gaming – mostly add complexity to the storyline without offering effective means to reduce or eliminate the bul whip effect. Further research resulted in a shift of focus to bullwhip avoidance and implementation of supply chain collaboration practices. The new ideas are described as “synchronisation of supply chain operations”, which has attracted new re-search. The paper mil offered some insight on how to work out chal enges that digitalisation has introduced. Our contention was that digitalisation should build on joint human/system reasoning that combines experience, insight, intuition, social interaction, etc. with support produced by automatic, intelligent systems, which in the paper mil case were combinations of mathematical optimisation and fuzzy ontology. A partial answer to joint human/system reasoning builds on digital coaching that offers instruments for knowledge transfer from experienced users to novice users and from knowledgeable users to automated systems. Decision Analytics could/should be the optimisation part of digital coaching. Final y, the context understanding and an emerging research position for Deci-sion Analytics formed an agenda proposal in four statements (cf. {4.1-4.4}). References C. Carlsson R. Fullér, Problem solving with multiple interdependent criteria, in: J. Kac-przyk, H. Nurmi, M. Fedrizzi (Eds.), Consensus under Fuzziness, Kluwer Academic Publishers, Boston, (1997), pp. 231–246. C. Carlsson, R. Ful ér, A fuzzy approach to the bul whip effect, Cybernetics and Systems ’2000, Proc. of the 15th European Meeting on Cybernetics and Systems Research, Vienna, April 25 –28, 2000, Austrian Society for Cybernetic Studies, (2000), pp. 228–233. C. Carlsson, R. Ful ér, Reducing the bul whip effect by means of intelligent, soft computing methods, in: Proc. of the 34th Hawai Internat. Conf. on System Sciences (HICSS-34), Island of Maui, Hawai , USA, January 3– 6, (2001). C. Carlsson, Soft computing in analytics: handling imprecision and uncertainty in strategic decisions, Fuzzy Econ. Rev. XVII (2) (2012) 3–21. C. Carlsson, J. Mezei, M. Brunel i, Decision making with a fuzzy ontology, Soft Computing 16 (7) (2012) 1143–1152. C. Carlsson, J. Mezei, M. Brunel i, Fuzzy ontology used for knowledge mobilisation, Int. J. Intell. Syst. 28 (1) (2013) 52–71. C. Carlsson, M. Brunelli, J. Mezei, A soft computing approach to mastering paper machines, in: Proceedings of HICSS-46, HICSS. 2013.61, IEEE (2013), pp. 1394–1401. C. Carlsson, Imprecision and uncertainty in management—the possibilities of fuzzy sets and soft computing, in: NOEMA XV, Romanian Academy of Science, (2016), pp. 89–114. C. Carlsson: Decision Analytics - A Position Paper 707 C. Carlsson, M. Heikkilä, J. Mezei, Fuzzy entropy used for predictive analytics, in: C. Kahraman, U. Kaymak, A. Yazici (Eds.), Fuzzy Logic in its 50th Year. New Developments, Directions and Challenges, Studies in Fuzziness, Springer, 341, (2016), pp. 187–210. C. Carlsson, Decision Analytics Mobilised with Digital Coaching, Intelligent Systems in Accounting, Finance and Management, ISAF1421, January/March, (2018), pp 3-17 DOI: 10.1002/isaf.1421 C. Carlsson, Decision Analytics – Key to Digitalization, Information Sciences, Vol 460-461, September (2018), pp 424-438 C. Carlsson, Combining ANFIS and Digital Coaching for Good Decisions in Industrial Processes, IFSA/NAFIPS'2019 (Lafayette, Louisiana, USA, June 18–21, 2019) Proceedings, Springer Verlag (2019), pp 190-200 C. Carlsson, Digital Coaching to Make Fuzzy Real Options Methods Viable for In-vestment Decisions, FUZZ-IEEE 2019 Proceedings, New Orleans (2019), pp 406-411, 978-1-5386-1728-1/19 ©2019 IEEE D. Cecez-Kecemanovic, Doing critical information systems research – arguments for a critical research methodology, Eur. J. Inf. Syst. 20 (2011) 440–455. E. Ciancimino, S. Cannella, M. Bruccoleri and J.M. Framinan, On the Bul whip Avoid-ance Phase: The Synchronised Supply Chain, EJOR, Vol 221, (2012), pp 49-63 Competing in 2020: winners and losers in the digital economy (2017), A Harvard Business Review Analytic Services Report, April 25, (2017). J. Hirvonen, T. Tommila, A. Pakonen, C. Carlsson, M. Fedrizzi, R. Fullér, Fuzzy keyword ontology for annotating and searching event reports, Proceedings of International Conference on Knowledge Engineering and Ontology Development (KEOD 2010), Valencia, Spain, (2010). D. Kahneman, Thinking Fast and Slow, Farrar, Straus and Giroux, New York, (2011) H.L. Lee, V. Padmanabhan and S. Whang, Information distortion in a supply chain: The bul whip effect, Management Science, 43(1997) 546-558. H.L. Lee, V. Padmanabhan and S. Whang, The Bul whip Effect in Supply Chains, Sloan Management Review, Spring (1997) 93-102 J. Mezei, M. Brunelli, C. Carlsson, A fuzzy approach to using expert knowledge for tuning paper machines, JORS 68 (6) (2017) 605–616. J.A. Morente-Molinera, R. Wikström, C. Carlsson, E. Viedma-Herrera, A linguistic mobile decision support system based on fuzzy ontology to facilitate knowledge mobilization, Decis. Support Syst. 81 (2016) 66–75. 708 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY RESEARCH IN PROGRESS THE ROLE OF IT IDENTITY IN THE FORMATION AND MITIGATION OF TECHNOSTRESS NAJMA SAIDANI, LAURA RUIZ SANTIAGO NEOMA Business School, Mont Saint-Aignan, France najma.saidani@neoma-bs.fr, laura.ruiz@neoma-bs.fr The pervasive and ubiquitous nature of technology has grown exponential y in the last decades. Technostress has been a common consequence of such intensive use, causing serious damage to IT users and organizations. How technostress is formed via primary appraisal toward IT and how it can be mitigated has been overlooked by prior research. This research Keywords: addresses these gaps by mobilizing an IT identity perspective. IT IT identity, identity informs about individuals’ attitudes towards IT and IT features, would explain the formation of negative feelings associated with technostress, primary the use of IT as wel as how technostress mitigation occurs in a appraisal, personal IT use setting. mitigation DOI https://doi.org/10.18690/um.fov.4.2023.43 ISBN 978-961-286-751-5 710 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The pervasive and ubiquitous nature of digital technologies has permeated every aspect of organizational and nonorganizational settings, creating important implications and chal enges for Information Systems (IS) scholars and giving rise to the so-cal ed dark side of Information Technology (IT) use (Tarafdar et al., 2013). This research stream argues that some of the features that make IT powerful (e.g., reliability, usefulness, portability, user-friendliness, and fast processing) can also threaten individual wel -being due to IT-induced stress, technology dependency, or IT misuse (Tarafdar et al., 2010). Within this line of research, the phenomenon of technostress has received great attention (Ragu-Nathan et al., 2018). Technostress has been defined as the stress experienced by individuals due to the constant need to adapt to new and evolving IT functionalities (Califf et al., 2020). It has been traditionally conceptualized as a negative aspect resulting from the use of IT, associated with negative consequences such as low job satisfaction, productivity, and high turnover rates (Ayyagari et al., 2011). Due to this negative denotation, how to mitigate technostress has also emerged as an important but overlooked topic (Galluch et al., 2015). Prior research has focused on technostress mitigation from an organizational perspective, disregarding how technostress mitigation takes place in a personal use setting (Salo et al., 2022). This research takes an IT identity perspective to answer the fol owing research question: what is the role of IT identity in the formation and mitigation of technostress in a personal IT use environment? This research contributes to two underexplored areas. First, the role of IT identity in the formation of technostress via primary appraisal toward IT, and second, how IT identity affects technostress mitigation in a personal IT use environment. IT identity refers to the extent to which an individual perceives the use of an IT as part of who he/she is and answers the question “Who am I in relation to this technology?” (Carter & Grover, 2015). It is therefore tied to the individual level of analysis. Today’s landscape is characterized by an increasing interlacing of IT and social routines as well as higher expectations from perpetual contact with IT in more complex social structures. IT identity informs about individuals’ attitudes toward IT and would explain the formation of negative feelings associated with the use of IT. We posit that IT identity plays a key role in understanding the formation and mitigation of technostress because IT N. Saidani, L. Ruiz Santiago: The Role of IT Identity in the Formation and Mitigation of Technostress 711 identity might lead to a successful self-verification of individuals’ role identities and reduce feelings of technostress (Carter & Grover, 2015). 2 Motivation & Gap First, although central to the understanding of the phenomenon of technostress, prior IS literature still did not fully explore the formation and shaping of individuals’ primary appraisal toward IT (Tarafdar et al., 2019). An individual can evaluate IT as a chal enge and a motivating factor or as a threatening and disturbing factor. A very limited number of studies have explored how individuals appraise the introduction of technologies and their impact on the formation and mitigation of technostress (Salo et al., 2022). Second, we push the boundaries of the ‘user’ concept that has been dominating the research stream of technostress. Although very insightful, the user concept does not cover the surroundings of the interaction between the sides of the technostress process (i.e., the person and the technological environment). As Lamb and Kling (2003) argue, users should be considered social actors. In fact, individuals are not merely and uniquely ‘users’ of IT, they are rather ‘social actors’ embedded in complex social settings; as ‘ their social roles and relationships are increasingly inseparable from their interactions with IT’ (Carter & Grover, 2015, p.931). We believe that the socially thin conceptualizations of individuals as merely users constitute a barrier to our understanding of such complex phenomena. Our study adopts this theoretical positioning by developing a model that captures, on the one hand, the technical and environmental settings that create technostress, and the intertwinement of social roles and IT usage in the other hand. Finally, our study mobilizes the IT identity lens to look at the formation and mitigation of technostress. IT identity helps captures the complexity of the intertwinement of IT and social structures. In fact, IT identity represents ‘ the set of meanings individuals at ach to the self in relation to IT—as a product of individuals’ personal histories of interacting with IT, as wel as a force that shapes their thinking and guides their IT use behaviors’. Thus, IT and social contexts in which individuals are embedded, are established in relation to each other and mutually and continually develop. The previous echoes with two core elements of technostress conceptualization. First, IT identity embodies the relationship between technology and social contexts as it considers IT as a social object, while technostress is theorized as a process that involves a transaction between the individual and the environment. Second, IT identity interferes with the shaping of attitudes toward IT and the use of IT while “primary appraisal” focuses on the 712 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY individual's assessment of the extent of environmental demand and influences the relationship between technology environmental conditions and techno‐stressors. In consequence, considering IT as a social object that individuals categorize themselves in relationship to would inform about the formation and mitigation of technostress. 3 Theoretical Background 3.1 Technostress: A Fragmented and Evolving Literature The questions of how and why the use of IT generates various demands on the individual are at the heart of an emerging area of scholarly investigation in the IS domain, namely technostress (Ayyagari et al., 2011; Ragu-Nathan et al., 2008). The concept of technostress, derived from the stress concept in the psychological stress literature, refers to the dynamic process in which individuals perceive that the demands of using an IT exceed one’s resources and are hard to meet (Ayyagari et al., 2011; D’Arcy et al., 2014; Galluch et al., 2015). The process includes (1) the presence of ‘technology environment conditions’, which are appraised as (2) demands or ‘techno stressors’ that are taxing the individual and require (3) coping responses resulting in (4) outcomes for the individual on the psychological, physical and behavioral levels. Studies on technostress have substantial y focused on its creators, the techno- stressors, and its outcomes. Little is known about the appraisal aspect, which unfolds how individuals appraise the technological environment conditions and interpret them as a threat and disturbing or as a chal enge and opportunity (Califf et al., 2020). Indeed, our knowledge about technostress is built on ‘fragmented investigation’ and disparate focus on specific aspects (Tarafdar et al., 2019). For example, according to Tarafdar et al. (2019)’s literature review, the antecedents of technostress have received a considerable proportion of interest (see Ayygari et al., 2011). Technostress outcomes have also been, according to the same source, under great focus. Overall, previous research falls short to discuss how individuals appraise techno-stressors neither why they perceive digital technologies as disturbing and potential y harmful. N. Saidani, L. Ruiz Santiago: The Role of IT Identity in the Formation and Mitigation of Technostress 713 Primary Appraisal One of the most overlooked dimensions of this process is the primary appraisal. Primary appraisal establishes and influences the relationship between technology environmental conditions and techno-stressors and informs about how negative feelings towards IT are formed (Tarafdar et al., 2019). Primary appraisal is the user’s assessment of the expected consequences of an IT event (Beaudry & Pinsonneault, 2005). Rooted in cognitive theories, it argues that individuals, when disrupted, proceed to an evaluation of the nature of the situation: does this constitute an opportunity or a threat? Individuals also judge the personal relevance of the disrupting event and its potential consequences (Lazarus, 1966; Lazarus & Folkman, 1984). If demands are perceived as adequate to the resources, individuals could develop a positive attitude towards the event and feel it as an opportunity. If perceived as exceeding one’s abilities, the individual will experience feelings of stress. To answer this question, different theoretical frameworks have been mobilized. The Person-Environment (P-E) fit of stress (Cooper et al., 2001; Edwards 1991) has been adopted to argue the absence of equilibrium between the IT demands and the capacity of individuals to meet them (Ayyagari et al., 2011). Other studies used the prism of individual and organizational characteristics to explore which ones would favor perceiving technology environment conditions as threat-techno-stressors. On the individual level, examples range from obsessive-compulsive personalities or neurotic dispositions (Bolger & Zuckerman, 1995; Chang, 1998) to low self-efficacy (Schaubroeck & Merritt, 1997). On the organizational level, examples range from the surveillance culture (Zuboff, 1988, 2015) to high organizational expectations (Barber & Santuzzi, 2015). Two recent studies (Califf, 2022; Salo et al., 2022) adopt an affordances lens to theorize what happens in the black box of the appraisal process. Recurrent Techno-stressors Techno-stressors refer to “IS stress creators appraised by the individual as threatening” (Tarafdar et al., 2015, p. 5). Prior research has identified several techno-stressors such as invasion, privacy concerns, complexity, overload, uncertainty, insecurity, or dependency (Califf et al., 2020; Ragu-Nathan et al., 2008; Tarafdar et al., 2007). We consider the most recurrent techno-stressors found in a personal use context: 714 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY invasion, privacy concerns, complexity, and dependency (Salo et al., 2022). Dependency involves an overreliance on IT to perform daily activities (Shu et al., 2011). Privacy concerns involve feelings of compromised individual privacy (Ayyagari et al., 2011). Complexity refers to the difficulty in using IT (Fischer & Riedl, 2017; Tarafdar et al., 2007). Invasion refers to situations in which users can be reached and available at any time (Fischer & Riedl, 2017; Tarafdar et al., 2010). Technostress Mitigation A few studies have addressed the question of how to mitigate technostress. Mainly from an organizational perspective, these studies explored the factors that help knowledge workers and IT professionals lower the negative feelings associated with using IT. Among these factors, researchers validated organizational programs, training, job control, and rewards (Arnetz, 1996; Hung et al., 2011; Tams et al., 2020), in addition to specific organizational support such as technical support, literacy facilitation, support with work–home boundaries, co-worker support, and user involvement (Benlian, 2020; Fuglseth & Sørebø, 2014; Maier et al., 2019; Ragu-Nathan et al., 2008; Tarafdar et al., 2015; Yan et al., 2013). On a more individual level, aspects pertaining to IT self-efficacy, IT experience and competencies have been associated with lower technostress levels (Shu et al., 2011; Tams et al., 2018) and better performance (Tarafdar et al., 2015). Studies on mitigating technostress have also explored the impact of individual actions such as positive re-interpretation, distancing or escaping from IT work on reducing technostress (Galluch et al., 2015; Pirkkalainen et al., 2019). Recent work by Salo et al. (2022) explored the formation of technostress using IT affordance as a theoretical lens (Majchrzak & Markus, 2014) and how individuals mitigate technostress feelings through a self-regulation perspective (Bandura, 1991; Baumeister et al., 1994). 3.2 IT Identity IT identity has been articulated by prior research to better understand one’s behavior with respect to IT in embedded social contexts. The conceptualization of IT identity builds on the assumption that identities represent the set of meanings and expectations that individuals internalize for their own behaviors (Carter & Grover, 2015). Identities are tied to social categories on both the collective and individual levels. While the former focuses on how identity emerges from membership in social N. Saidani, L. Ruiz Santiago: The Role of IT Identity in the Formation and Mitigation of Technostress 715 groups or categories (Tajfal & Turner, 2004), the latter covers several forms of identity such as role identity (e.g., family role, work role) (McCall & Simmons, 1978), person identity (i.e., values and norms that individuals define themselves with) (Burke & Stets, 2009) and material identity. IT identity, as presented by Carter and Grover (2015), falls under the material identity form where individuals tie their identity to material objects such as places or personal possessions (such as IT) (Clayton, 2003). Given that IT identity applies to the individual, this variable is tied to the individual level of analysis. Prior research on identity indicates that IT identity should be a multidimensional construct (Clayton, 2003). IT identity is reflected in individuals’ perceptions of relatedness, dependence, and emotional energy with respect to the IT (Carter & Grover, 2015). Relatedness refers to the extent to which the boundary between the self and IT becomes blurred and individuals show feelings of connectedness with the IT. Dependence captures the extent to which individuals are reliant on IT to achieve important instrumental goals. Emotional energy captures feelings of enthusiasm and energy when interacting with an IT. 4 Preliminary Research Design and Research Model On the one hand, our sample includes undergraduate students from a Business School in France. Students are an appropriate target because they are digital natives who widely use IT for personal purposes in a voluntary way (Craig et al., 2019). This decision is also consistent with prior research on technostress (Gal uch et al., 2015) and IT identity (Carter et al., 2020). On the other hand, we will focus on a specific unit of technology given that, according to prior work, users may develop many IT identities, each one tied to a specific IT (Carter & Grover, 2015; Stets & Burke, 2005). Given that technologies with broader use are more likely to enact IT identity (Carter & Grover, 2015), we will use a social networking site (SNS) (e.g., Facebook, Instagram). In addition, Carter and Grover (2015) stated that technologies such as Facebook or Instagram with material properties such as sharing status updates and photos, instant messaging notifications, and dynamic information feeds, which are at the same time portable and networked, are particularly amenable to IT identity formation. We plan to perform a mixed-methods design. Mixed-methods design is an appropriate method because of its ability to “address confirmatory and explanatory research questions” (Venkatesh et al., 2016, p. 437). As our focus is on understanding the role 716 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY of IT identity in the formation and mitigation of technostress, we first plan to conduct a qualitative study to help us develop the research model and understand how the different variables are related to each other. We plan to conduct focus group to understand how SNS features are appraised by students, how primary appraisal affects recurrent techno-stressors, and how students’ IT identity affects primary appraisal and technostress mitigation. Once the hypotheses and research model are developed based on the first study, a quantitative approach will be followed for the empirical testing. Figure 1 shows the preliminary model we might test. We plan to develop a survey instrument adopting existing Likert scale measures of IT features, IT identity, and recurrent techno-stressors to ensure content validity. We will survey students from the same institution with a screening question to only include those students using the chosen SNS. IT features wil be specified as a second-order composite construct and wil include four dimensions: functionality, bandwidth, mobility, and malleability (Carter & Grover, 2015; Esmaeilzadeh, 2021). IT identity wil be considered a second-order composite construct with three reflective dimensions whose items will be adopted from Carter et al. (2020). Measures of the techno-stressors wil be adapted from prior work on technostress (e.g., Ragu-Nathan et al., 2008; Tarafdar et al., 2007; Tarafdar et al., 2017). Figure 1: Research model N. Saidani, L. Ruiz Santiago: The Role of IT Identity in the Formation and Mitigation of Technostress 717 5 Conclusion and Expected Contributions In order to expand the understanding of technostress, examining the formation of negative feelings towards IT and how users mitigate their effects represents a focal point. As part of that endeavor, this study looks at the formation and mitigation of technostress through the lens of IT identity. While prior IS research has acknowledged that understanding the role of IT use is critical to understanding how technostress forms over time via primary appraisal and how its mitigation takes place, our literature review reveals that little is known about the role of IT identity in such phenomena. The potential contributions of this research are the following. First, we contribute and extend prior work on technostress formation and mitigation by investigating how primary appraisal for threat techno-stressors happens when IT identity is considered while examining whether IT identity is a way to mitigate technostress. Second, IT identity has been understudied in prior IS research. Although there are studies that explain who people are in relation to IT, they do not treat IT as an integral part of one’s self (Esmaeilzadeh, 2021). This study will examine IT identity associated with a SNS and its role in the formation and mitigation of technostress. We wil be able to elucidate whether IT identity reinforces or weakens the relationship between technology environmental conditions, techno-stressors, and user behavior. Therefore, this work will contribute to theory building on IT identity by also providing empirical evidence for the multidimensional nature of the construct. References Arnetz, B. B. (1996). Techno-Stress: A perspective psychophysiological study of the impact of a control ed stress-reduction program in advanced telecommunications systems design work. Journal of Occupational and Environmental Medicine, 38(1), 53-65. Ayyagari, R., Grover, V., & Purvis, R. (2011). Technostress: Technological antecedents and implications. MIS Quarterly, 35(4), 831-858. Bandura, A. (1991). Social cognitive theory of self-regulation. Organizational behavior and human decision processes, 50(2), 248-287. Barber, L. K., & Santuzzi, A. M. (2015). Please respond ASAP: Workplace telepressure and employee recovery. Journal of Occupational Health Psychology, 20(2), 172–189. Baumeister, R. F., Heatherton, T. F., & Tice, D. M. 1994. Losing Control: How and Why People Fail at Self-Regulation, San Diego, CA: Academic Press. Beaudry, A., & Pinsonneault, A. (2005). Understanding User Responses to Information Technology: A Coping Model of User Adaptation . MIS Quarterly, 29(3), 493–524. 718 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Belk, R. W. (1988). Possessions and the Extended Self. Journal of Consumer Research, 15(2), 139-168. Benlian, A. (2020). A Daily Field Investigation of Technology-Driven Spil overs from Work to Home. MIS Quarterly, 44(3). Bolger, N., & Zuckerman, A. (1995). A framework for studying personality in the stress process. Journal of Personality and Social Psychology, 69(5), 890–902. Burke, P. J., & Stets, J. E. 2009. Identity Theory, New York: Oxford University Press. Califf, C. B. (2022). Stressing affordances: Towards an appraisal theory of technostress through a case study of hospital nurses' use of electronic medical record systems. Information and Organization, 32(4), 100431. Califf, C., Sarker, S., & Sarker, S. (2020). The bright and dark sides of technostress: A mixed-methods study involving healthcare IT. MIS Quarterly, 44(2), 809-856. Carter, M., & Grover, V. (2015). Me, Myself, and I(T): Conceptualizing Information Technology Identity and its Implications. MIS Quarterly, 39(4), 931-957. Carter, M., Petter, S., Grover, V., & Thatcher, J. B. (2020). Information Technology Identity: A Key Determinant of IT Feature and Exploratory Usage. MIS Quarterly, 44(3). Cavanaugh, M. A., Boswel , W. R., Roehling, M. V., & Boudreau, J. W. (2000). An Empirical Examination of Self-Reported Work Stress among U.S. Managers. Journal of Applied Psychology, 85(1), 65-74. Chang, E. C. (1998). Dispositional optimism and primary and secondary appraisal of a stressor: Control ing for confounding influences and relations to coping and psychological and physical adjustment. Journal of Personality and Social Psychology, 74(4), 1109–1120. Clayton, S. (2003). Environmental Identity: A Conceptual and an Operational Definition, in Identity and the Natural Environment, S. Clayton and S. Opotow (eds.), Cambridge, MA: MIT Press, pp. 45-67. Cooper, C. L., Dewe, P. J., & O’Driscol , M. P. (2001). Organizational Stress: A Review and Critique of Theory, Research, and Applications, Thousand Oaks, CA: SAGE Publications. D’Arcy, J., Herath, T., & Shoss, M. (2014). Understanding Employee Responses to Stressful Information Security Requirements: A Coping Perspective. Journal of Management Information Systems, 31(2), 285-318. Edwards, J. R. (1991). Person-Job Fit: A Conceptual Integration, Literature Review, and Methodological Critique. International Review of Industrial and Organizational Psychology, 6, 283-357. Esmaeilzadeh, P. (2021). How does IT identity affect individuals’ use behaviors associated with personal health devices (PHDs)? An empirical study. Information & Management, 58(1), 103313. Fuglseth, A. M., & Sørebø, Ø. (2014). The effects of technostress within the context of employee use of ICT. Computers in Human Behavior, 40, 161-170. Gal uch, P., Grover, V., & Thatcher, J. (2015). Interrupting the Workplace: Examining Stressors in an Information Technology Context. Journal of the Association for Information Systems, 16(1), 1-47. Hung, W. H., Chang, L. M., & Lin, C. H. (2011). Managing the Risk of Overusing Mobile Phones in the Working Environment: A Study of Ubiquitous Technostress. In Proceedings of the Pacific Asia Conference on Information Systems, Brisbane. Kahn, R., & Byosiere, P. (1992). Stress in organizations in M. D. Dunnette, and L. M. Hough (eds.) Handbook of industrial and organizational psychology (2nd ed.), vol. 3, Palo Alto: Consulting Psychologists Press, 571–650. Lamb, R., & Kling, R. (2003). Reconceptualizing Users as Social Actors in Information Systems Research. MIS Quarterly, 27(2), 197-235. Lazarus, R. S. (1966). Psychological stress and the coping process, New York, NY, USA: McGraw-Hill. Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping, New York, NY, USA: Springer publishing company. Maier, C., Laumer, S., Wirth, J., & Weitzel, T. (2019). Technostress and the hierarchical levels of personality: a two-wave study with multiple data samples. European Journal of Information Systems, 28(5), 496-522. N. Saidani, L. Ruiz Santiago: The Role of IT Identity in the Formation and Mitigation of Technostress 719 Majchrzak, A., & Markus, M. L. (2014). Technology Affordances and Constraints in Management Information Systems (MIS), in Encyclopedia of Management Theory, E. H. Kessler (ed.), Thousand Oaks, CA: SAGE Publications. McCall, G. J. (2003). The Me and the Not-Me: Positive and Negative Poles of Identity, in Advances in Identity Theory and Research, P. J. Burke, T. Owens, R. T. Serpe, and P. A. Thoits (eds.), New York: Springer, pp. 11-26. McCal , G. J., & Simmons, J. L. (1978). Identities and Interactions, New York: Free Press. Pirkkalainen, H., Salo, M., Tarafdar, M., & Makkonen, M. (2019). Deliberate or instinctive? Proactive and reactive coping for technostress. Journal of Management Information Systems, 36(4), 1179-1212. Ragu-Nathan, T., Tarafdar, M., Ragu-Nathan, B., & Tu, Q. (2008). The consequences of technostress for end users in organizations: Conceptual development and empirical validation. Information Systems Research, 19(4), 417-433. Salo, M., Pirkkalainen, H., Chua, C. E. H., & Koskelainen, T. (2022). Formation and Mitigation of Technostress in the Personal Use of IT. MIS Quarterly, 46(2), 1073-1108. Schaubroeck, J., & Merritt, D. (1997). Divergent effects of job control on coping with work stressors: The key role of self-efficacy. Academy of Management Journal, 40(3), 738-754. Shu, Q., Tu, Q., & Wang, K. (2011). The Impact of Computer Self-Efficacy and Technology Dependence on Computer- Related Technostress: A Social Cognitive Theory Perspective. International Journal of Human–Computer Interaction, 27(10), 923-939. Stets, J. E., & Burke, P. J. (2005). A Sociological Approach to Self and Identity. In Handbook of Self and Identity, M. R. Leary and J. P. Tangney (eds.), New York: The Guilford Press, pp. 128-152. Tajfal, H., & Turner, J. C. (2004). The Social Identity Theory of Intergroup Behavior. In Political Psychology: Key Readings, J. T. Jost and J. Sidanius (eds.), New York: Psychology Press, pp. 276-293. Tams, S., Ahuja, M., Thatcher, J., & Grover, V. (2020). Worker stress in the age of mobile technology: The combined effects of perceived interruption overload and worker control. Journal of Strategic Information Systems, 29(1), Article 101595. Tams, S., Thatcher, J. B., & Grover, V. (2018). Concentration, competence, confidence, and capture: An experimental study of age, interruption-based technostress, and task performance. Journal of the Association for Information Systems, 19(9), 2. Tarafdar, M., Cooper, C., & Stich, J. (2019). The technostress trifecta-techno eustress, techno distress and design: Theoretical directions and an agenda for research. Information Systems Journal, 29(1), 6-42. Tarafdar, M., Gupta, A., & Turel, O. (2013). The Dark Side of Information Technology Use. Information Systems Journal, 23(3), 269-275. Tarafdar, M., Pul ins, E. B., & Ragu‐Nathan, T. S. (2015). Technostress: negative effect on performance and possible mitigations. Information Systems Journal, 25(2), 103-132. Tarafdar, M., Tu, Q., & Ragu-Nathan, T. (2010). Impact of technostress on end-user satisfaction and performance. Journal of Management of Information Systems, 27(3), 303-334. Tarafdar, M., Tu, Q., Ragu-Nathan, B. S., & Ragu-Nathan, T. S. (2007). The impact of technostress on role stress and productivity. Journal of Management Information Systems, 24(1), 301-328. Venkatesh, V., Brown, S. A., & Sullivan, Y. (2016). Guidelines for conducting mixed-methods research: An extension and il ustration. Venkatesh. Journal of the Association for Information Systems, 17(7), 435-495. Yan, Z., Guo, X., Lee, M. K., & Vogel, D. R. (2013). A conceptual model of technology features and technostress in telemedicine communication. Information Technology & People, 26(3), 283-297. Zuboff, S. (1988). In the Age of the Smart Machine: The future of work and power, New York, NY: Basic Books. Zuboff, S. (2015). Big other: surveil ance capitalism and the prospects of an information civilization. Journal of Information Technology, 30(1), 75-89 720 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY RESEARCH IN PROGRESS CONCEPTUALIZING THE IMPACT OF DIGITAL BUSINESS MODELS ON PRIVACY CONCERNS MIRANDA KAJTAZI,1 ERDELINA KURTI2, 3 1 Lund University, School of Economics and Management, Department of Informatics, Lund, Sweden miranda.kajtazi @ics.lu.se 2 Malmö University, Faculty of Technology and Society, Department of Computer Science and Media Technology, Malmö, Sweden erdelina.kurti@mau.se 3 Linnaeus University, Faculty of Technology, Department of Informatics, Växjö, Sweden erdelina.kurti@lnu.se Digital technologies have enabled novel forms and reconfigurations of value creation, delivery, and capture. These new reconfigurations chal enge the conventional notion of value creation with digital business models. On that premise, the widening of privacy concerns, alert us that organizations of the elite digital, like Netflix, Amazon, and Spotify, design technology to feed on personal data, based on algorithmic profiling capabilities. Then, privacy itself becomes their digital business model. In this paper we conceptualize the impact of digital Keywords: business models on privacy concerns, by presenting a focused digital literature review that presents 4 waves of research on business understanding privacy from the context of digital business models, privacy models. With our initial findings, we recommend that future concerns, technological development should pay central attention to value privacy-preserving digital business models, by making it possible creation, invisibility, that data privacy is envisioned with the right safeguards, targeting right 'invisibility' of the user. safeguards DOI https://doi.org/10.18690/um.fov.4.2023.44 ISBN 978-961-286-751-5 722 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The exponential advancement and widespread utilization of digital technologies has spawned profound innovations, which have disrupted traditional businesses and reconfigured a number of industries (Nambisan et al., 2020). Digitization, which is defined as the conversion of analog data to digital (Yoo et al., 2010), of products and services, is the cornerstone of innovations, which transcend geographical and industrial boundaries enabling novel business models (Constantinides et al. 2018; Nambisan et al., 2020). A business model represents a key source of performance and competitive advantage of organizations (Teece, 2010), hence becoming an imperative for digital transformation. It refers to the “architecture of value creation, delivery and capture mechanism” (Teece, 2010, p.172), in a multi-actor network. The core of business models is on value creation and capture, not only for the organization itself, but also for other actors in the ecosystem (Amit and Zott, 2020). For organizations, the advent of information technology (IT) in the 1990s became a ground to breed a new generation of entrepreneurs that redefined the rules of doing business, primarily on the basis of competition, facilitated by IT (Gordon, 2000). On that end, the dot com bust in the 2000s, mandated a few entrepreneurs to reinvent the use of IT by crafting a new economic order (Zuboff, 2015). Pioneers like Netflix, Amazon, and Spotify, started to become the best attendants of feeding on personal data (Loebbecke and Picot, 2015), with a constant online surveillance, often without the knowledge of a person (Zuboff, 2019). The legal enforcement, however, with the ratification of the General Data Protection Regulation (GDPR) as the most powerful regulation ever created, presented a shift in the mind-set of how data protection is handled by such organizations. In this paper, we look at digital business models vis-á-vis privacy concerns, with the aim to provide an initial conceptual model on the interplay of digital business models and privacy concerns, over time. The rest of the paper is organized as follows. We first present our conceptualizations on digital business models and privacy. We then present a focused literature review fol owed by an initial conceptualization to view privacy as a direct consequence of reconfigurations of business and the growth of digital business models. We then highlight potential contributions of our initial conceptualization, followed by future work. M. Kajtazi, E. Kurti: Conceptualizing the Impact of Digital Business Models on Privacy Concerns 723 2 Digital Business Models and Privacy Concerns Digital technologies have enabled novel forms and reconfigurations of value creation, delivery, and capture. These new reconfigurations chal enge the conventional notion of value creation, postulating that value is co-created by “aggregating recombinant technology components by interacting with diverse resources and often across firm boundaries” (Hukal and Henfridsson, 2017 p. 488). As a result, the notion of digital business model has gained widespread popularity both in scholarly work, but also in practice. Digital business models refer to business models enabled by the utilization of digital technologies (Amit and Zott, 2020). Bärenfänger and Otto (2015, p. 18) define digital business models “as a business model whose underlying business logic deliberately acknowledges the characteristics of digitization and takes advantage of them; both in interaction with customers and business partners, and in its internal operations”. In consideration of the foregoing, it is no surprise that we have now reached a point when digital business models have influenced the generation of an organizational mind-set that even knows our deepest secrets (Acquisti et al., 2022; Zuboff, 2019). The value of personal information has made it possible for goods to increase prices tenfold on personalized services. Even the simplest case of M&M’s legendary milk chocolate candy pack is no stranger to that. M&M owns a platform that al ows you to personalize a chocolate pack, where you share personal information, e.g. dates and photographs, taking the opportunity of such data to turn it into a commodity (Crain, 2016). That also allows such platforms not only to influence our future consumer behavior, it also al ows for an astronomical price tag, al made possible by the new wave of digital business models configured for personalization. Then, it is not new to us that IT has become a constant in reconfiguring traditional roles of people in the digital realm, including their traditional view on privacy concerns (Zhang et al., 2022). Organizations driven by information capitalism (Zuboff, 2019), especially the elite digital, show an unstoppable appetite for data that forms 95% of the global economy (Srnicek, 2017). From a macro perspective, the digitization of an organization presented an opportunistic reality (Thrift, 2011) where concepts like “everyware” (Greenfield, 2006) came to life. From a micro perspective, however, secrecy in such organizations came at the expense of privacy (Solove, 2011). Zubbof’s “big other” became a precondition to argue that we are in 724 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY the hands of a new form of capitalism that she termed surveilance capitalism (Zuboff, 2015), where personal boundaries on our own privacy are put to test (Zhang et al., 2022). Contrary to this view, we know that digitization is key to produce a number of digital business models that deserted the spatial and temporal limits, often empowering people. Just to name a few, from the speed of information, to the significance of online payments for simple transactions, technological capabilities can steer progress in the right direction. However, the digital era is a new reality for people that has brought us more tension than consensus, putting people’s own privacy-protective behaviors to test (Quach et al., 2022; Zhang et al., 2022). We live with pressure trying to balance our and others' physical presence with the digital presence (Acquisti et al., 2022). Alongside the backdrop of this pressure, digital business models are leading us to conceptualize our personal digital self as a type of self that transcends the borders and acknowledges our physical self, often recognizing privacy as a loss in that transcension (Zuboff, 2019). 2.1 Focused Literature Review: Digital Business Models and Personal Boundaries on Privacy In a focused literature review, we identify 22 research articles (listed in Appendix) published at the European Journal of Information Systems (EJIS) in the course of 3 decades. The focus on EJIS stems from their distinctive European perspective on theory and practice for a global audience. Coupled with European Union laws and regulations on data privacy, such as GDPR, it makes for a unique candidate to study the conceptualization of digital business models vis-á-vis privacy concerns, over time. The article analyses yielded 4 waves of research on understanding privacy from the context of digital business models. Important to highlight, is the fact that all articles recognize that privacy is a human right, but the difference across the waves is noticed on the fact that early studies have the tendency to conceptualize on privacy concerns from the user perspective, compared to current studies that place a lot of responsibility on the design of IT itself. Wave 1, presented the early take on the use of digital business models, such as in the form of e-commerce and social media from a network perspective, where privacy conceptions were formed around the “user privacy concern” (e.g. Junglas et al., 2008), where personal privacy and the identifiable person were key, along with the data rights, but centered on the actual “user x”. M. Kajtazi, E. Kurti: Conceptualizing the Impact of Digital Business Models on Privacy Concerns 725 Wave 2 presented the wake of new digital business models that further fueled the presence and use of e-commerce and social media for more personalized services, including services for personal digital healthcare, which influenced a complex analysis on privacy as e.g. a right and commodity (Smith et al., 2011). This made it critical to view “user disclosure and personalization” as a real physical person and that data and information disclosure became pivotal to our understanding of privacy concerns (Posey et al., 2010; Warkentin et al., 2011). Wave 3, showed that new techniques on big data analyses with machine learning algorithms, made it more concerning that “personal data and privacy loss” is a real threat (Parks et al., 2017). Then, wave 4 and the current wave, focuses on important aspects, such as traceability and integrity (Raddatz et al., 2021; Parks et al., 2022), where a clear motivator for such studies depends on the movements of human rights perspective, where “personal privacy rights” are linked back to laws and regulations. In fact, the focus on privacy as a fundamental human right (UN Declaration of Human Rights, Article 12), is key to guide recent studies on design of technology for privacy protection and privacy-preservation mechanisms. Figure 1 il ustrates these waves with an example where user x is identified as Sarah Smith, which leads to show how the other waves address privacy concerns, over time. Figure 1: From User Privacy to Privacy as a Right In reference to these waves, our analysis identified several types of digital business models studied in these waves, e.g. platform business models of e-commerce or datadriven business models; etc. In describing Figure 1, we conceptualize them as digital business models. Such digital business models rely on digital technologies, which contribute to the generation and proliferation of data, which has recently exceeded in growth and profit (Wiener et al., 2020). 726 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY As such, digital business model- dependent organizations have harnessed the potential of digital technologies to create novel reconfigurations of value creation and capture, either through novel offerings, reconfiguration of activities, transactions, structure and/or governance mechanisms (Amit and Zott, 2020), al dependent on personal data feeding. To that end, successful organizations as Netflix, Amazon, and Spotify, have configured their business models and innovations around data. Netflix for example, has shifted its focus from a retailer of DVDs mail delivery to innovating its business model around data to improve customer experience through personalization and customization (Mier and Kohli, 2021). The common link across these organizations is that their business models are configured and innovated as data dependent digital platforms. The latter leads us to reflect back on how privacy concerns have shifted from conceptualizing about the unknown “user x” to the actual “physical person”. 3 Initial Findings and Future Research Direction Despite the fact that we have ample opportunities with the introduction and exponential growth of digital business models, and that new IT developments present ideas, tools, and models with privacy-preserving mechanisms, we identify that challenges with data privacy still remain detrimental. Novel configurations of digital business models where privacy becomes the core value creation mechanism, leads us to term them as privacy-based business models. In this relationship, privacy itself becomes the business model. We recommend future technological design to focus on privacy-preserving digital business models, which should make it possible that data privacy is envisioned with the right safeguards, targeting ‘invisibility’. Otherwise, Mann’s and Matzner’s (2019) cal that we risk producing technology that does not account for privacy and data protection rights, goes against GDPR’s cal on the ‘right’ not to be ruled by automated decisions. Our future work depends on bringing the question of ethics into view, on how technology is shaped to feed on personal data, where user awareness and digital literacy remain challenging. The vast majority of today’s digital users have limited awareness about algorithmic profiling capabilities and how detrimental its effects are on their privacy. At the same time, these users have become pivotal in supporting data-driven business models to thrive, letting such models to feed on their valuable personal data. M. Kajtazi, E. Kurti: Conceptualizing the Impact of Digital Business Models on Privacy Concerns 727 Acknowledgements This paper is partialy funded by the Crafoord Foundation with project number 20200668. References Acquisti, A., Brandimarte, L., Hancock, J. (2022). How privacy’s past may shape its future. Science, 375(6578), 270-272. Amit, R., Zott, C. (2020). Business Model Innovation strategy: Transformational Concepts and Tools for Entrepreneurial Leaders. John Wiley & Sons: Hoboken, NJ, USA. Bärenfänger, R., Otto, B. (2015). Proposing a capability perspective on digital business models. In 2015 IEEE 17th Conference on Business Informatics, 1, 17-25. Crain, M. (2016). The limits of transparency: Data brokers and commodification, New Media and Society, 20(1), 88-104. Gordon, R. J. (2000). “Does the "New Economy" Measure Up to the Great Inventions of the Past?" Journal of Economic Perspectives, 14(4), 49-74. Greenfield, A. (2006) Everyware: The Dawning Age of Ubiquitous Computing, New Riders, Boston, USA. Hukal, P., Henfridsson, O. (2017). Digital Innovation—A definition and integrated perspective. In The Routledge Companion to Management Information Systems, 1st ed.; Routledge: London, UK, 2017; pp. 360–369. Junglas, I.A., Johnson, N.A., Spitzmüller, Ch. (2008). Personality traits and concern for privacy: an empirical study in the context of location-based services. European Journal of Information Systems, 17, 387-402. Loebbecke, C., Picot, A. (2015). Reflections on societal and business model transformation arising from digitization and big data analytics: A research agenda. The Journal of Strategic Information Systems, 24(3), 149-157. Mann, M., Matzner, T. (2019). Chal enging algorithmic profiling: The limits of data protection and anti-discrimination in responding to emergent discrimination. Big Data & Society, 1-11. Mier, J. & Kohli, A.K. (2021). Netflix: reinvention across multiple time periods, reflections and directions for future research. AMS Review, 11, pp.1-12. Parks, R., Xu, H., Chu, Ch-H., Lowry, P.B. (2017). Examining the intended and unintended consequences of organisational privacy safeguards. European Journal of Information Systems, 26, 37-65. Parks, R. F., Wigard, R. T., Lowry, P.B. (2022). Balancing information privacy and operational utility in healthcare: proposing a privacy impact assessment (PIA) framework. European Journal of Information Systems, 1-18. Posey, C., Lowry, P.B., Roberts, T.L., Selwyn El is, T. (2010). Proposing the online community self-disclosure model: the case of working professionals in France and the U.K. who use online communities. European Journal of Information Systems, 19, 181-195. Quach, S., Thaichon, P., Martin, K.D., Weaven, S., Palmatier, R.W. (2022). Digital technologies: tensions in privacy and data. Journal of the Academy of the Marketing Sciences, 50, 1299-1323. Raddatz, N., Coyne, J., Menard, Ph., Crossler, R.E. (2021). Becoming a blockchain user: understanding consumers’ benefits realization to use blockchain based- applications. European Journal of Information Systems, 1-28. Smith, J.H., Dinev, T., Xu,H. (2011). Information Privacy Research: An Interdisciplinary Review. MIS Quarterly, 35(4), 989-1015. Srnicek, N. (2017). The chal enges of platform capitalism: Understanding the logic of a new business model, Juncture, 23(4), 254-257. Teece, D.J. (2010). Business models, business strategy and innovation. Long Range Planning, 43, 172– 194. 728 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Wiener, M., Saunders, C., Marabeli, M. (2020). Big-data business models: A critical literature review and multiperspective research framework. Journal of Information Technology, 35(1), 66-91. Zhang, N., Wang, Ch., Karahanna, E., Xu, Y. (2022). Peer Privacy Concerns: Conceptualization and Measurement. MIS Quarterly, 46(1), 491-530. Zuboff, Sh. (2015). Big other: surveillance capitalism and the prospects of an information civilization. Journal of Information Technology, 30, 75-89. Zuboff, Sh. (2019). The age of surveil ance capitalism: The fight for a human future at the new frontier of power, Profile Books, UK. M. Kajtazi, E. Kurti: Conceptualizing the Impact of Digital Business Models on Privacy Concerns 729 Appendix Context of Context of No Citation Key Focus of the Theoretical Study Informatio Data/Infor- n Systems mation Foundation Information privacy threats and maintaining utility Means-end chain Parks et al. in a healthcare Patient theory 1 eHealth (2022) privacy compliance Information Value-Focused context with value- thinking approach focused thinking (VFT) approach. Blockchain as data store to promote data privacy, Blockchain transaction research Transactiona integrity. Health belief Raddatz et Blockchain l data, 2 Factors that model al. (2021) databases personal influence Perceived benefits data consumers’ of blockchain- perceptions of based databases blockchain-based databases’ benefits Information privacy (“the ability of individuals to control Contact tracing Information Lin, Carter Smartphone the terms technology, citizen Privacy 3 & Liu s, contact under which information Technology (2021) tracing-apps their privacy concerns. Adoption personal information is acquired and used” p.389) Social Gamified SETA Online self- networking Dincelli & artefact using the disclosure (OSD) sites (SNS) 4 Chengalur- formats of text and Data privacy Attitudes and Social Smith visual to identify intentions towards engineering security threats. OSD behavior Gamification 730 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Context of Context of No Citation Key Focus of the Theoretical Study Informatio Data/Infor- n Systems mation Foundation Provides an SETA SETA and understanding of gamification the linkage between technology artefacts and human experiences. Benefits of tracing apps Appeals for prosocial behavior Contact-tracing Contact- Constant usage Trang et apps during the tracing apps and usability 5 Contact data al. (2020) pandemic. Mobile requirements Mass acceptance. technology Sensitive data and privacy concerns App acceptance, user-centered design The failure in the Rowe, design and E-GOV Apps for Nqwenya Tracing, adoption of Stop- Col ection of crisis management 6 ma & smartphone COVID app in data Alienation in Richet app France. Conditions critical theory (2020) of such failure. Information privacy in the context of peer relationships on commercial social Privacy research Ozdemir, media sites. Privacy-related Smith & Personal 7 A model that Social media constructs Benamati information considers Information (2017) relationships disclosure between the constructs of privacy experiences. Lowry, Important IT artefacts to IS 8 IS Research Big data Dinev & concerns in the artefacts M. Kajtazi, E. Kurti: Conceptualizing the Impact of Digital Business Models on Privacy Concerns 731 Context of Context of No Citation Key Focus of the Theoretical Study Informatio Data/Infor- n Systems mation Foundation Willison hope of improving Security and (2017) the effectiveness privacy research of security and Opportunities – privacy research. online platforms, Outlines three IoT, big data promising opportunities for IS research that is compelling to security and privacy researchers. Investigate the consequences of Health informatics privacy safeguard Privacy safeguards enactment in in healthcare medical practices, The intended including whether Parks et al. Health Information versus unintended 9 it influences their (2017) informatics privacy consequences of ability to enacting privacy meet privacy safeguards in requirements and organizations. whether workflows are impeded Analyzed the influences of the attitudes, subjective norms and perceived Information Foth behavioral Health care Data 10 security (2016) control on systems protection employees’ intentions to comply with data protection regulations. Bansal, Important website Privacy Concern, Zahedi & features: privacy Internet, Col ection of 11 Trust, and Privacy Gefen policy statements websites data Assurance (2015) + privacy 732 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Context of Context of No Citation Key Focus of the Theoretical Study Informatio Data/Infor- n Systems mation Foundation assurance cues are what online providers use to increase individuals’ trust and wil ingness to disclose private information online. Comprehensive examination of the process by which privacy assurance mechanisms influence trust and the moderating role of privacy concern in this process Facebook users’ learning-based attitude formation Self-disclosure Chen & Data in and the Social networking 12 Sharma Social media social relationship sites (2015) network between member Attitude literature attitude and self- disclosure. Determinants for success and failure of identity Success factors of Roßnagel management Identity we identity 13 et al. systems. management User data management (2014) Analyze the systems solutions preferences and wil ingness to pay of prospective users. M. Kajtazi, E. Kurti: Conceptualizing the Impact of Digital Business Models on Privacy Concerns 733 Context of Context of No Citation Key Focus of the Theoretical Study Informatio Data/Infor- n Systems mation Foundation Existing privacy compliance Methodology that procedures and systematical y privacy-by-design Oetzel & considers privacy IT Data Risk assessment 14 Speikerma issues by using a applications protection methodologies that nn (2014) step-by-step tackle security and privacy impact privacy issues assessment. PIA Examines how Information European citizens privacy decide to disclose Situationally Miltgen & and protect Antecedents and Peyrat- Information 15 their personal data NA. consequences Guillard privacy and thereby reveals The importance of (2014) cultural and trust generational Privacy-related divides. issues Privacy definition and conceptualization Privacy in the IT, e- Anthropological information age, Dinev commerce, Personal and cultural angle 16 future (2014) social data of privacy opportunities for networks Regulation research. Privacy and convenience. Privacy paradox Develops and tests a framework of information privacy and its The concept of Dinev et correlates, the Information 17 NA. privacy – literature al. (2013) latter often being privacy review confused with or built into definitions of information privacy per se. 734 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Context of Context of No Citation Key Focus of the Theoretical Study Informatio Data/Infor- n Systems mation Foundation Perceived personalization quality can outweigh the impact of privacy concerns. Personalization Li & Personalizati Service providers Customers’ privacy 18 Unger on NA. can improve the concerns (2012) applications perceived quality Privacy protection of personalization services being offered in order to offset customer privacy concerns. Investigates the antecedents of Personal Warkentin, information Social learning Healthcare data, 19 et a. privacy policy theory systems? sensitive (2011) compliance Compliance data? efficacy by individuals. An online Social exchange community self- theory disclosure model, Social penetration tested in a cross Online theory Posey et cultural communities 20 NA. Cross-cultural al. (2010) setting using data , social theory related to provided by networking individualism- French and British collectivism working profes sionals Concern for Fill the gap of privacy (CFP) “research has The co-evolving shown that the nature of privacy CFP Junglas et World wide and technology 21 can have a NA. al. (2008) web PMT and threat negative influence appraisals on the adoption of Personality traits information and threat technology; but appraisals M. Kajtazi, E. Kurti: Conceptualizing the Impact of Digital Business Models on Privacy Concerns 735 Context of Context of No Citation Key Focus of the Theoretical Study Informatio Data/Infor- n Systems mation Foundation little is known about factors likely to influence such concern.” Internet and e- commerce Examines cross- Personal diffusion in Italy cultural differences information Hofstede’s cultural Dinev et beliefs related to e- 22 NA. Information theory al. (2006) commerce use privacy Fukuyama’s theory for Italy and the concerns of trust and social United States. capital 736 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY RESEARCH IN PROGRESS INTEGRATION OF THE SUSTAINABLE DEVELOPMENT GOALS IN PROJECT-BASED IT EDUCATION JORIS GRESNIGT,1 MARLIES VAN STEENBERGEN2 1 HU University of Applied Sciences Utrecht, Institute for ICT, Utrecht, Netherland joris.gresnigt@hu.nl 2 HU University of Applied Sciences Utrecht, Research Group Digital Ethics, Utrecht, Netherland marlies.vansteenbergen@hu.nl Keywords: sustainable Higher education institutions play a significant role in reaching development goals, the Sustainable Development Goals (SDGs). There is however a SDGs, gap between the abstract nature of SDGs and the need to values, integrate these in the day-to-day educational environment. This sustainability, ethics, paper presents an ongoing study that reports preliminary findings information regarding the integration of insights from Value Sensitive Design technology, into frequently employed artifacts within a project-based IT IT, ICT, educational context, with the aim of translating abstract higher Sustainable Development Goals into teaching practice. education DOI https://doi.org/10.18690/um.fov.4.2023.45 ISBN 978-961-286-751-5 738 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The adoption of the 2030 Agenda for Sustainable Development by United Nations Member States in 2015 set forth a comprehensive framework for achieving global sustainable development through 17 interrelated Sustainable Development Goals (SDGs). These goals target social, economic, and ecological sustainability and seek to promote universal values, such as health, equality, and social justice (UN, 2015; UN, 2017). In this paper we use the term value in the sense of 'what is important to living things, with a focus on ethics and morality', expanding the Friedman and Hendry (2019) definition referring to humans, in order to prevent an anthropocentric focus as indicated by Borthwick et al. (2022). Higher education institutions have a significant role in contributing to these global efforts (Žalėnienė & Pereira, 2021). Literature on sustainable education highlights the importance of integrating sustainability across the entire curriculum to develop truly competent students on the SDGs (Robinson et al., 2022; Wu & Shen, 2016). However, operationalizing this integration presents a challenge. While research and international policy documents focus on sustainable competences (Dias, 2022; Peet et al., 2004; Wiek et al., 2015; Sinakou et al., 2019; UNESCO, 2017; UNECE, 2011; Bianchi et al., 2022), they do not address the integration of the SDGs in the methods, models and theories of a specific domain. In light of the abstract nature of the SDGs and their contextual dependency within specific domains (Leal et al., 2019), there is a need for an approach that can translate these abstract goals into practical applications within a particular educational setting. In this ongoing study we explore such a translation in the context of project-based IT education, a form of education that mirrors working practice and is increasingly employed in higher education (Chen & Yang, 2019). The research question we address is: How can the SDGs be integrated in project-based IT education? In this study we focus on the design phase of IT projects since it entails the primary decisions on embedding values in IT products, as wel as determining product functionality and non-functional characteristics (Becker et al., 2015; Friedman & Hendry, 2019; Lago et al., 2015; Shapira et al., 2017). Consistent with Umbrello et al. (2021), we utilize insights from the Value Sensitive Design (VSD) approach (Friedman et al., 2006) as a bridge between the values in the SDGs and IT projects. VSD is an approach to integrate values into technological design (Friedman et al., J. Gresnigt, M. van Steenbergen: Integration of the Sustainable Development Goals in Project-Based IT Education 739 2006). We use insights from VSD to incorporate values into the artifacts that students deliver in the design phase of their IT projects, such as a stakeholder analysis, persona descriptions or a prototype. Our aim is to adapt existing artifact formats to contain value dimensions by default, which we expect will lead to easier adoption among students, University staff and project clients than when we introduce completely new artifacts. We expect that these adapted artifacts wil consequently lead to IT products that better incorporate SDG values. Two examples of this translation from SDG values to artifact formats are presented in the preliminary results section. In the next section, we present the theoretical background on integrating sustainable development in the artifacts of project-based IT education. Section 3 discusses our research method. In section 4 we present preliminary results. We end with conclusions and further research in section 5. 2 Theoretical background The Sustainable Development Goals (SDGs) embody specific values, as identified by Keitsch (2018), Muñoz et al. (2022), and Umbrello et al. (2021). The SDGs list ecological values, such as harmony with nature (SDG 12) and clean air (SDG 3), social values, including justice (SDG 16) and equality (SDG 5), as well as economic values, such as inclusive sustainable growth (SDG 8) and productivity (SDG 2) (UN, 2017). Value Sensitive Design (VSD) is an approach to integrate values into technological design (Friedman et al., 2006). It is characterized by assessing the impact a new design may have on stakeholders' values. The values of different stakeholders may not always align and value tensions may be created. VSD aims to make values and value tensions explicit and carefully weigh them in making design choices. VSD offers a wide range of methods to do so, such as value dams and flows, value scenarios, and envisioning cards, and encourages to rework other existing methods and instruments to include a value perspective (Friedman & Hendry, 2019). VSD can be used to operationalize values mentioned in the SDGs using norms, leading to specific design criteria or requirements (Friedman & Hendry, 2019; Umbrel a et al., 2021). The exact choice of which values and corresponding norms and design criteria are used depends on the interplay between the different stakeholders. 740 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3 Research method We adopt a design science approach in our study (Hevner & Chatterjee, 2010), since our object is to design a way to incorporate SDGs in project-based IT education. Structuring our study according to Peffers et al. (2007), we use the five steps of their design science research process: problem identification and motivation, objectives of the solution ( as stated in the introduction) , design and development, demonstration and evaluation. The research is currently in the design and development phase. Insights from VSD are used to design adaptations of existing formats commonly used in project-based IT education. As a first step in the design and development phase we made the SDG values explicit and translated them to the IT field. This was done by finding academic articles on the relationship between the SDGs and IT and publications such as Tjoa & Tjoa (2016) of mostly UN organizations aligned with a certain SDG. E.g. the UN agricultural organization FAO publicizes about SDG 2 on hunger and IT. Analyzing these publications we created a list of positive and negative influences of IT on each of the SDGs which contain a myriad of values. IT can e.g. lead to different types of inequalities: caused by lack of access for poor people (SDG 1), for people with disabilities or caused by discrimination due to the use of racially or gender skewed data sets (SDG 10). This list is not meant to be exhaustive, but it aims to help the thought process of translating SDGs to the IT context. Having such an overview is however not enough if values are not part of the artifacts the students produce. Therefore for each of the artifacts commonly used in the IT design phase, we analyzed the gap between the artifact in use and insights from VSD on how to incorporate values and consequently adapted the artifacts to include a value dimension. For example, personas typical y do not contain values and tend not to include non-typical personas, therefore we added these aspects to the persona format. Student and lecturers can use the overview of (SDG) values in IT produced in the first step as an inspiration to draw values from. To make it easier for students to produce the adapted artifacts we made formats which include worked examples of the artifact and ‘how to’ steps. These formats were tested with ten IT lecturers who applied the formats to a real case in a workshop setting and adapted based on the lecturers’ feedback. We additionally provided supporting material such as videos and an easy-to-fill-out lay-out on an online collaborative platform, learning materials, assignments and assessment criteria, and embedded the SDGs in the learning J. Gresnigt, M. van Steenbergen: Integration of the Sustainable Development Goals in Project-Based IT Education 741 objectives. We are currently in the process of testing the formats with students in different years of their IT studies. The students use the formats in their projects, and we gather feedback from both lecturers and students. 4 Preliminary results Until now we produced 11 formats to be used by students in the design phase, including formats for personas, formats for customer journeys and additional prompt questions. For example we extended the commonly used stakeholder analysis. In the regular IT design process, especially in an agile setting, the focus is on end-users. This creates IT that does not by default take into account the interests and values of a broad range of stakeholders and might create harm or fail to obtain benefits. The SDGs require including a broad group of stakeholders. VSD methods on stakeholder involvement offer insights, such as inclusion of indirect stakeholders (f.i. non-users, future generations or other cultures) and their values. The extended format aims to help students get a broader picture of stakeholders and their values so these can be taken into account as input for the design. Another example is the extension of the process of defining the problem. The typical questions used by students in defining the problem are limited in scope and tend to not contain questions that are vital for sustainable development such as looking into the relevance of the problem for future generations and relationships among various problems. VSD provides, with e.g. the envisioning cards method (Friedman et al., 2011), a whole range of added questions that can open up the students’ horizon. While testing the formats with the lecturers during the workshop they indicated that using them has definitely helped in making the abstract SDGs tangible for them. They indicated that it was easier for them to use IT specific values such as accessibility or privacy than the related more abstract SDGs reduce inequality or ensure fundamental freedoms, especial y when integrated into artifact formats they were already used to. However, lecturers less familiar with the original artifacts, indicated that they felt overwhelmed by the added complexity. The first results on the quality of the artifacts students made using the formats show a remarkable difference between students being coached by lecturers already familiar with the adapted formats and those to whom these were entirely new. Only the students with more experienced lecturers in this area were able to produce artifacts 742 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY that incorporate ethical/sustainable values in a sensible way. Students from classes with a less experienced lecturer tended not to use the provided new sustainable formats or to use them in such a way that they did not provide any valuable insights into values and value tensions. The use of the formats also created discussion between students and lecturers. Some students, especial y those that were newly introduced to these sustainable demands in the later part of their study, were protesting against them, even cal ing them “woke”, with a negative connotation. 5 Conclusion and further research The preliminary results of this ongoing research indicate the feasibility of making the SDGs tangible for students and lecturers by enhancing existing methods with a value perspective. These methods can help to bridge the gap found in literature and in practice between SDGs and IT development. Even so, the inclusion of the value perspective adds complexity that may be overwhelming. Considering the still new, contested and volatile nature of the subject, there is clearly a need for training of lecturers, embedded in an aligned sustainable learning environment along with an ongoing open discussion among the stakeholders on SDGs and values in a specific context. Our next steps are to complete the design and development phase by validating the usability and effect of the formats we developed with both lecturers and students. We wil do so by continuing to use the formats in actual student projects throughout all years of the IT course and validating their use in focus groups. We will also conduct a comparative evaluation of the artifacts produced in various settings. Based on the outcomes we wil further extend the formats with an implementation approach and supporting teaching materials. We intend to demonstrate the resulting way of integrating SDGs in project-based IT education by implementing it in other parts of the IT institute which offer project-based education. Acknowledgements We thank al lecturers and students who participated in this study. We also thank the anonymous reviewers for their valuable comments on an earlier draft. J. Gresnigt, M. van Steenbergen: Integration of the Sustainable Development Goals in Project-Based IT Education 743 References Becker, C., Betz, S., Chitchyan, R., Duboc, L., Easterbrook, S., Penzenstadler, B., Seyff, N., Venters, C. (2015). Requirements: The Key to Sustainability. IEEE Software. 33. 1-1. 10.1109/MS.2015.158. Bianchi, G., Pisiotis, U., Giraldez, M. (2022). GreenComp – The European sustainability competence framework. 10.2760/13286. Borthwick, M., Tomitsch, M., Gaughwin, M. (2022). From human-centred to life-centred design: Considering environmental and ethical concerns in the design of interactive products. Journal of Responsible Technology. 10. 100032. 10.1016/j.jrt.2022.100032. Chen, C. H., & Yang, Y. C. (2019). Revisiting the effects of project-based learning on students’ academic achievement: A meta-analysis investigating moderators. Educational Research Review, 26, 71-81. Cross, N. (2011). Design Thinking: Understanding how designers think and work. Bloomsbury/Berg. ISBN 9781847886361 Dias, B.,Onevetch, R., Santos, J., Lopes, G.(2022). Competences for Sustainable Development Goals: The Chal enge in Business Administration Education. Journal of Teacher Education for Sustainability, 24(1) 73-86. Friedman, B., Hendry, D. G. (2019). Value Sensitive Design: Shaping Technology with Moral Imagination. The MIT Press Friedman, B., Kahn, P., Borning, A., Zhang, P., Gal etta, D. (2006). Value Sensitive Design and Information Systems. 10.1007/978-94-007-7844-3_4 Friedman, B., Nathan, L.P., Kane, S., Lin, J. (2011). Envisioning cards, Seattle , W.A.: University of Washington. http://www.envisioningcards.com Hevner, A., Chatterjee, S. (2010). Design science research in information systems. Design research in information systems: theory and practice, 9-22. Keitsch, M. (2018), Structuring ethical interpretations of the sustainable development goals-Concepts, implications and progress. Sustainability 2018, 10, 829. Lago, P., Aklini Kocak, S., Crnkovic, I., & Penzensradler, B. (2015). Framing Sustainability as a Property of Software Quality. Communications of the ACM, 58(10), 70-78. https://doi.org/10.1145/2714560 Leal Filho W., Shiel C., Paço A., Mifsud M., Ávila L.V., Brandli L.L., Molthan-Hill P., Pace P., Azeiteiro, U.M., Vargas V.R., Caeiro S. (2019), Sustainable Development Goals and sustainability teachingat universities: Fal ing behind or getting ahead of the pack?, Journal of Cleaner Production, doi: https://doi.org/10.1016/j.jclepro.2019.05.309. Muñoz-Arteaga, J., García, H.L., Col azos, C.A., Granol ers, A. (2022). Impact of Human-Computer Interaction in the Global Sustainable Development Goals. In: Verdegay, J.L., Brito, J., Cruz, C. (eds) Computational Intelligence Methodologies Applied to Sustainable Development Goals. Studies in Computational Intelligence, vol 1036. Springer, Cham. https://doi.org/10.1007/978-3-030-97344-5_3 Peet, D.,J. Mulder, K., Bijma, A. (2004), Integrating SD into engineering courses at the Delft University of Technology, International Journal of Sustainability in Higher Education. 5. 278-288. 10.1108/14676370410546420. Peffers, K., Tuunanen, T., Rothenberger, M. A., Chatterjee, S. (2007). A design science research methodology for information systems research. Journal of management information systems, 24(3), 45-77. Pereira, J. C., de FSM Russo, R. (2018). Design thinking integrated in agile software development: A systematic literature review. Procedia computer science, 138, 775-782. Robinson, J., Ariga, A., Cameron, S., & Wang, R. (2022). Reaching the Rest: Embedding Sustainability in Undergraduate Student Learning. Journal of Integrative Environmental Sciences, 19(1), 171-187. 744 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Shapira, H., Ketchie, A., Nehe, M. (2017). The integration of Design Thinking and Strategic Sustainable Development. Journal of Cleaner Production, 140, 277-287. Sinakou, E., Boeve-de Pauw, J., Van Petegem, P. (2019). Exploring the concept of sustainable development within education for sustainable development: implications for ESD research and practice. Environment, Development and Sustainability. 21. 1-10. 10.1007/s10668-017-0032-8. Tjoa, A.M., Tjoa, S. (2016). The Role of ICT to Achieve the UN Sustainable Development Goals (SDG). In: Mata, F., Pont, A. (eds) ICT for Promoting Human Development and Protecting the Environment. WITFOR 2016. IFIP Advances in Information and Communication Technology, vol 481. Springer, Cham. https://doi.org/10.1007/978-3-319-44447-5_1 Umbrello, S., Capasso, M., Balistreri, M. et al. (2021). Value Sensitive Design to Achieve the UN SDGs with AI: A Case of Elderly Care Robots. Minds & Machines 31, 395–419. https://doi.org/10.1007/s11023-021-09561-y UNECE. Learning for the future: Competences in Education for Sustainable Development; United Nations Economic Commission for Europe (UNECE): Geneva, Switzerland, 2012. UNESCO. (2017). Education for Sustainable Development Goals: Learning Objectives. Paris: UNESCO. http://unesdoc.unesco.org/images/0024/002474/247444e.pdf UN (2015). Transforming Our World: The 2030 Agenda for Sustainable Development. Resolution Adopted by the General Assembly on 25 September 2015, 42809, 1-13. UN (2017) Work of the Statistical Commission pertaining to the 2030 Agenda for Sustainable Development. Resolution adopted by the General Assembly on 6 July 2017, (A/RES/71/313). Wiek, A., Bernstein, M., Foley, R., Cohen, M., Forrest, N., Kuzdas, C., Kay, B., & Withycombe, Keeler, L. (2015). Operationalising competencies in higher education for sustainable development. In: Barth, M., Michelsen, G., Rieckmann, M., Thomas, I. (Eds.). Handbook of Higher Education for Sustainable Development. Routledge, London. pp. 241-260. Wu, Y.-C.J. and Shen, J.-P. (2016). Higher education for sustainable development: a systematic review. International Journal of Sustainability in Higher Education. Vol. 17 No. 5, pp. 633-651. https://doi.org/10.1108/IJSHE-01-2015-0004 Žalėnienė, I., Pereira, P. (2021). Higher Education For Sustainability: A Global Perspective. Geography and Sustainability. 2. 10.1016/j.geosus.2021.05.001. RESEARCH IN PROGRESS AUTOML AS FACILITATOR OF AI ADOPTION IN SMES: AN ANALYSIS OF AUTOML USE CASES ANNA K. POLZER, JOHANNES P. ZEIRINGER, STEFAN THALMANN University of Graz, Graz, Austria anna.polzer@uni-graz.at, johannes.zeiringer@uni-graz.at, Stefan.thalmann@uni-graz.at While the uptake of AI and ML has been rising in recent years, SMEs still face various adoption challenges. In contrast to large enterprises, SMEs struggle to adopt AI as already the identification of suitable AI use cases requires substantial technical expertise. At the same time, productivity tools like AutoML promise easy access to AI capabilities to non-experts. This research-in-progress aims to investigate how AutoML tools can be utilised to facilitate the adoption of AI in SMEs. In a focus group with 11 representatives from SMEs, we identified and Keywords: discussed potential AutoML use cases in detail. Results show that SME, AI the identification of potential use cases rarely focused on existing adoption, and available data but rather repeated known use cases and AutoML, success stories from large enterprises. We argue that a paradigm Bled eConference, shift towards a data-centric approach would be beneficial to research exhaust the capabilities of AutoML for SMEs. paper DOI https://doi.org/10.18690/um.fov.4.2023.46 ISBN 978-961-286-751-5 746 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Advancements in artificial intelligence (AI) such as large language models have recently gained attention and increased awareness among the broader public. However, the utilization of AI in organisations still faces various challenges, ranging from technological and organisational to social, legal, ethical, economical and data-related ones (Dwivedi et al., 2021). As such, many organisations still struggle to apply AI in their key operations and outside of proof of concept deployments (Dzhusupova et al., 2022). Especially small and medium-sized enterprises (SMEs) have a slow pace of adopting AI (Hansen & Bøgh, 2021). In contrast to large enterprises, SMEs lack the financial resources to invest in infrastructure, recruit AI talents in a competitive labour market or buy in external expertise (Bauer et al., 2020). The resulting lack of AI capabilities hinders the identification of potential AI use cases in SMEs (Bauer et al., 2020). Additionally, current approaches and tools that address challenges in AI adoption usually presume existing AI knowledge (Kirschbaum et al., 2022). One approach that aims to provide easy access to AI based on machine learning (ML) to non-specialist users is AutoML (Zöller & Huber, 2021). Automated ML, short AutoML, subsumes the methods aiming to automate, at least to some extent, al stages of the design and development of ML-based systems (Hutter et al., 2019). AutoML simplifies the application of ML by reducing the number of steps of data preparation, model selection, model hyperparameters for the applicant and helps them with data visualisation, model comprehensibility and usage (Crisan & Fiore- Gartland, 2021). One of the main goals of AutoML is to democratise access to AI/ML Technology for people without in-depth knowledge of coding, statistics or ML (Zöller & Huber, 2021). Already a myriad of AutoML frameworks and tools exist which enable users to build and explore AI. Depending on the level of expertise, users can utilise coding libraries such as auto-sklearn or autokeras that automate only specific parts in the ML development pipeline or commercial tools that can be utilised through graphical user interphases (GUI), e.g., Google AutoML or JadBio. While the development of AI systems by non-ML expert users through the use of AutoML is often seen as problematic (Crisan & Fiore-Gartland, 2021; Wang et al., 2019), AutoML does ⁠ provide an opportunity for early exploration and easy trial runs by domain experts A. K. Polzer, J. P. Zeiringer, S. Thalmann: AutoML as Facilitator of AI Adoption in SMEs: An Analysis of AutoML Use Cases 747 (Xin et al., 2021). As such, AutoML provides professionals in different areas of expertise the opportunity to explore and identify possible AI use cases without the need for in-depth knowledge of AI/ML. However, research on the required knowledge of AutoML users or the appropriate integration of AutoML into the AI development process is very limited (Polzer & Thalmann, 2022). As of now, little research has focused on how SMEs identify and select AI use cases. Additionally, there is little knowledge of how AutoML can be utilised by domain and business experts to recognise AI opportunities in different application fields. This study combines both research gaps and tries to answer the question How can AutoML facilitate the adoption of AI in SMEs? Thus, this research in progress (RiP) is a first explorative investigation on how AutoML can be introduced to SMEs to leverage their AI capabilities and how AutoML provides an easy-to-use exploration tool to identify possible AI use cases. 2 Methodology A focus group with 11 representatives from 10 Austrian SMEs was held in February 2023. As seen in table 1, the participants came from different industry sectors and with different AI experience. The goal was to discuss SMEs’ capabilities concerning AI and to discover potential AutoML use cases the SMEs representatives saw in their own organisations. After a short introduction to AI and its chal enges, two demonstrators of exemplary use cases conducted with different AutoML tools were shown to foster the participants' creativity. The participants were grouped into two sub-groups in the second part of the workshop. In these focus groups, each participant was asked to think of possible AI use cases to be implemented using AutoML in their organisation. The participants were provided with template cards to elaborate on the identified use cases with additional information on what data would be needed, what type of analysis would be needed and what benefits and chal enges they could encounter during the implementation of such an AI use case. The participants had 20 minutes to individually think of possible AI use cases and to fil the template. Afterwards, each participant presented their use case using the template, and the group discussed requirements, feasibility, and possible challenges in the implementation. Finally, we discussed and jointly reflected by the entire group. The 748 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY focus groups were audio recorded and transcribed for analysis. The filed-in template cards were also analysed. Afterwards, al identified use cases of both focus groups were analysed and three distinct groups of AI use cases emerged. Table 1: Participants # Industry sector1 Company Size2 AI/ML experience 1 Human health Medium Sized Novice 2 Professional, scientific & Medium Sized Some experience technical activities 3 Professional, scientific & Medium Sized Expert technical activities 4 Information & Micro Expert communication 5 Other service activities Small Some experience 6 Manufacturing Small Expert 7 Other service activities not SME Expert 8 Other service activities Micro Some experience 9 Transportation and storage Small Novice 10 Manufacturing Small Some experience 11 Information & Micro Expert communication 3 Results In total, 15 distinct potential AutoML use cases were identified, discussed, and clustered into three groups depending on their purpose. As most use cases would influence either the primary value-adding activities (e.g., operations), or supporting activities (e.g. HR management) of a firm's value chain (Porter, 1998) value chain was used as the basis of the for the grouping of the use cases. As such, the first cluster focuses on improving the core activities of existing business models. The second cluster focuses on support activities, often managerial activities that are not specifically related to the value proposition of the business model. The third cluster 1 Based on United Nations (2023): International Standard Industrial Classification, URL: https://unstats.un.org/unsd/publication/seriesM/seriesm_4rev4e.pdf [last retrieved March 31st] 2 Based on Europan Commision SME Defintion (2003): https://single-market-economy.ec.europa.eu/smes/sme-definition_en [last retrived May 19th, 2023] A. K. Polzer, J. P. Zeiringer, S. Thalmann: AutoML as Facilitator of AI Adoption in SMEs: An Analysis of AutoML Use Cases 749 consists of the application of AI leading to extensions of the existing business model or business model innovations (see figure 1). In the primary activity cluster seven AI use cases to be implemented with AutoML have been identified. A common theme across these use cases was quality management. Thus, the use cases focused on improving maintenance or quality testing activities. For this cluster, the participants had a clear idea of what input data might be needed and how implementing the use cases could influence their business as wel as which chal enges they might encounter. One exemplary use case was the prediction of the deterioration of plates used to punch out different components in a production process. The participant knew the data and analysis needed to implement the use case in their company. Clear benefits, such as better decision support for maintenance activities, were also recognised. The most important chal enge for this use case was the availability of data, a common chal enge across many identified use cases. The use cases of the support activity cluster did not have a direct connection to any particular product or service offered by the SMEs. Instead, the five use cases were centred around supporting activities, such as improving project cost predictions or procurement needs. These use cases had a more general nature and were not focused on the core or value-adding processes of the SME. In contrast to the primary activity cluster, it seemed more chal enging to estimate the business value and special y to make a proper cost–benefit analysis. For instance, one discussed use case suggested the segmentation of potential customers for acquisition purposes. The main advantage in this category of use cases was seen in freeing of resources, e.g., from routine and time-intensive tasks, which can then be better utilised in other activities. The business model innovation cluster explored more innovative ways to use AI that were not variations of already broadly known use cases, like predictive maintenance. The use cases of this cluster had different purposes and application fields, such as decontamination acceleration through the support of AutoML models in healthcare facilities, emotion detectors for consumption predictions, or veracity evaluation of news. Participants had varying ideas about what input data was necessary for these use cases, and some were uncertain or had limited knowledge of what suitable input data would be necessary. However, with the novelty of the 750 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY approaches also, the chalenges and uncertainties concerning the implementation of such use cases increased. As such, aspects of privacy concerns in connection to GDPR were mentioned, but also the complexity and variety of data or the chal enges in validation and ensuring compliance were mentioned. Figure 1: Identified use cases 4 Discussion and Outlook Overall, our 11 participants acknowledged the capabilities of AutoML to start thinking about AI use cases in their SMEs. As the ideation and identification of AI use cases can either be purpose-driven or data-driven (Strum et al., 2021), AutoML, especial y regarding data-driven identification, can provide beneficial support, as it provides easy to use access for exploring existing data sets (Wang et al., 2019). Therefore, we provided first evidence that AutoML can support the identification of AI use cases. The analysis of the focus groups showcased that the identification of the use cases was based on already known success stories that can be observed in literature and are present in the public discourse, such as predictive maintenance or customer segmentation (Thalmann et al., 2018). In this regard, mostly a purpose-driven approach to finding AI use cases was utilised, which aimed at improving process steps, tasks, or decisions through AI. However, especial y a data-driven explorative approach of use case identification can lead to new ideas and approaches on how to use AI. A. K. Polzer, J. P. Zeiringer, S. Thalmann: AutoML as Facilitator of AI Adoption in SMEs: An Analysis of AutoML Use Cases 751 In this regard, participants who already have good knowledge or experience with AI, proposed novel and disruptive AI use cases. The use cases of the business model innovation cluster were proposed by participants who at least had some experience in the field of AI/ML. As such, a minimum of training regarding AI applications upfront is required, which also influences the necessary level of knowledge required to utilise AutoML effectively (Polzer & Thalmann, 2022). Thus, the identification and further the implementation of AI use cases (especial y with disruptive impact) still requires conscious effort in the development of organizational AI capabilities (Sjödin et al., 2021). The black-box character of AI was discussed as a potential barrier, especial y regarding its adoption in sensitive use cases. Participants demanded explainability features (see (Gashi et al., 2022) for an overview) and many times envisioned causal discovery (see (Vuković & Thalmann, 2022) for an overview). Similarly, also the feasibility of the proposed use cases in relation to ethical and privacy issues was discussed. Thus, especially in use cases relying on personal data, like in the emotion detection use case, considerable chal enges concerning legal but also ethical perspectives were highlighted. Therefore, there is a need for providing user guidance on AutoML tools to ensure the development of fair, accountable, and transparent AI systems (Polzer & Thalmann, 2022). This RiP paper has many limitations and serves as starting point for future research. First, we used AutoML demos to spark discussions among SME representatives and to identify use cases. So far, we do not have evidence that such use cases can be implemented in SMEs. For this purpose, we plan implementation case studies with SMEs organised as think-aloud studies. Second, the AutoML knowledge of some participants was limited, and our data suggest that more knowledge could facilitate the capabilities to "think out of the box". Also, in our case studies, we will investigate which knowledge is needed to use AutoML in a responsible way and how more knowledge affects the capability to identify new use cases. Third, so far, our sample relies on a small sample of SMEs just from Austria. For this purpose, we plan additional workshops in Austria, Germany, Portugal, and Spain to broaden our focus. This future work will especially focus on the limitations and challenges in implementing AutoML into SMEs. 752 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Acknowledgement The research was funded by the European Union as part of the Erasmus+ "Colaborative development of Al capabilities in SMEs" project (Grant number #2O22.T. ATO1.KA22O.H ED.OOOO89256). References Bauer, M., van Dinther, C., & Kiefer, D. (2020). Machine Learning in SME: An Empirical Study on Enablers and Success Factors. In AMCIS 2020 Proceedings. https://aisel.aisnet.org/amcis2020/adv_info_systems_research/adv_info_systems_research/ 3 Crisan, A., & Fiore-Gartland, B. (2021). Fits and Starts: Enterprise Use of AutoML and the Role of Humans in the Loop. In Y. Kitamura (Ed.), ACM Digital Library, Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems (pp. 1–15). Association for Computing Machinery. https://doi.org/10.1145/3411764.3445775 Dwivedi, Y. K., Hughes, L., Ismagilova, E., Aarts, G., Coombs, C., Crick, T., Duan, Y., Dwivedi, R., Edwards, J., Eirug, A., Galanos, V., Ilavarasan, P. V., Janssen, M., Jones, P., Kar, A. K., Kizgin, H., Kronemann, B., Lal, B., Lucini, B., . . . Wil iams, M. D. (2021). Artificial Intelligence (AI): Multidisciplinary perspectives on emerging chal enges, opportunities, and agenda for research, practice and policy. International Journal of Information Management, 57, 101994. https://doi.org/10.1016/j.ijinfomgt.2019.08.002 Dzhusupova, R., Bosch, J., & Olsson, H. H. (2022). The Goldilocks Framework: Towards Selecting the Optimal Approach to Conducting AI Projects. In 2022 IEEE/ACM 1st International Conference on AI Engineering–Software Engineering for AI (CAIN)) (pp. 124–135). Gashi, M., Vuković, M., Jekic, N., Thalmann, S., Holzinger, A., Jean-Quartier, C., & Jeanquartier, F. (2022). State-of-the-Art Explainability Methods with Focus on Visual Analytics Showcased by Glioma Classification. BioMedInformatics, 2(1), 139–158. https://doi.org/10.3390/biomedinformatics2010009 Hansen, E. B., & Bøgh, S. (2021). Artificial intelligence and internet of things in small and medium-sized enterprises: A survey. Journal of Manufacturing Systems, 58, 362–372. https://doi.org/10.1016/j.jmsy.2020.08.009 Hutter, F., Kotthoff, L., & Vanschoren, J. (2019). Automated Machine Learning. Springer International Publishing. https://library.oapen.org/handle/20.500.12657/23012 https://doi.org/10.1007/978-3-030-05318-5 Kirschbaum, J., Posselt, T., & Roth, A. (2022). Use-Case-Based Innovation For Artificial Intelligence– An Ontological Approach. In ECIS 2022 Research-in-Progress Papers. https://aisel.aisnet.org/ecis2022_rip/64/ Polzer, A. K., & Thalmann, S. (2022). The impact of AutoML on the AI development process. In Proceedings of the 2022 Pre-ICIS SIGDSA Symposium. https://aisel.aisnet.org/cgi/viewcontent.cgi?article=1012&context=sigdsa2022 Porter, M. E. (1998). Competitive Advantage: Creating and Sustaining Superior Performance (2. Auflage). Free Press. https://ebookcentral.proquest.com/lib/kxp/detail.action?docID=4934923 Sjödin, D., Parida, V., Palmié, M., & Wincent, J. (2021). How AI capabilities enable business model innovation: Scaling AI through co-evolutionary processes and feedback loops. Journal of Business Research, 134, 574–587. https://doi.org/10.1016/j.jbusres.2021.05.009 Strum, T., Fecho, M., & Buxmann, P. (2021). To Use or Not to Use Artificial Intelligence? A Framework for the Ideation and Evaluation of Problems to Be Solved with Artificial Intel igence. In Hawai International Conference on System Sciences 2021. University of Hawai'i at Manoa Hamilton Library. A. K. Polzer, J. P. Zeiringer, S. Thalmann: AutoML as Facilitator of AI Adoption in SMEs: An Analysis of AutoML Use Cases 753 Thalmann, S., Mangler, J., Schreck, T., Huemer, C., Streit, M., Pauker, F., Weichhart, G., Schulte, S., Kittl, C., Pollak, C., Vukovic, M., Kappel, G., Gashi, M., Rinderle-Ma, S., Suschnigg, J., Jekic, N., & Lindstaedt, S. (2018, July 11–14). Data Analytics for Industrial Process Improvement A Vision Paper. In 2018 IEEE 20th Conference on Business Informatics (CBI) (pp. 92–96). IEEE. https://doi.org/10.1109/CBI.2018.10051 Vuković, M., & Thalmann, S. (2022). Causal Discovery in Manufacturing: A Structured Literature Review. Journal of Manufacturing and Materials Processing, 6(1), 10. https://doi.org/10.3390/jmmp6010010 Wang, D., Weisz, J. D., Mul er, M., Ram, P., Geyer, W., Dugan, C., Tausczik, Y., Samulowitz, H., & Gray, A. (2019). Human-AI Col aboration in Data Science. Proceedings of the ACM on Human-Computer Interaction, 3(CSCW), 1–24. https://doi.org/10.1145/3359313 Xin, D., Wu, E. Y., Lee, D. J.‑L., Salehi, N., & Parameswaran, A. (2021). Whither AutoML? Understanding the Role of Automation in Machine Learning Workflows. In Y. Kitamura, A. Quigley, K. Isbister, T. Igarashi, P. Bjørn, & S. Drucker (Eds.), Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems (pp. 1–16). ACM. https://doi.org/10.1145/3411764.3445306 Zöller, M.‑A., & Huber, M. F. (2021). Benchmark and Survey of Automated Machine Learning Frameworks. Journal of Artificial Intelligence Research, 70, 409–472. https://doi.org/10.1613/jair.1.11854 754 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY RESEARCH IN PROGRESS RAPID SCALING OF A DANISH PUBLIC HEALTH SYSTEM UNDER COVID-19 NICKLAS FREDERIKSEN, ERIK L. MØLLER, JARL TUXEN, SARAH E. O’NEILL, MORTEN BOESEN Copenhagen School of Design and Technology, Copenhagen, Denmark nifr@kea.dk, erlm@kea.dk, jart@kea.dk, saro@kea.dk, mobo@kea.dk In recent years cloud infrastructure services have acted as engines for scaling applications when user demand spikes. A discipline typical y recognized as complex, expensive, error-prone, and time-consuming. In the field of healthcare services, data is considered sensitive under the European Union’s data protection law and are therefore under strict jurisdiction disallowing the Danish public services to utilize cloud scalability. During the COVID-19 lockdown a small group of expert practitioners was tasked with scaling public health services to Keywords: accommodate an exponential number of excess users who scalability, needed to access test results and immunity passports. An effort health further restrained by a severely limited timeframe of two weeks. information systems, By utilizing the critical incident technique this paper is an effort rapid empirical y to capture the most significant decisions in the scaling scaling, process including organizational aspects, virtualization, content sensitive data, delivery network, lazy-loading, and firewall interface Bled configuration. eConference DOI https://doi.org/10.18690/um.fov.4.2023.47 ISBN 978-961-286-751-5 756 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Scaling information systems is an important discipline for both academics and practitioners from various industries. The discipline can be defined as the process of expanding in scope or size (for example increasing the number of features or the number of end-users) (Sahay & Walsham, 2006). Scaling has been studied from various theoretical perspectives and with different aspects in focus, including the technical, organizational, and institutional aspects. It has been noted that the inability to scale is not caused by technological problems alone. Often, the difficulties are caused by organizational or managerial issues such as confusion of roles, people making bad decisions, and lack of attention to organizational and technical implementation (Abbott & Fisher, 2015). With the COVID-19 pandemic, the rapid scaling of a public health information system suddenly became critical for the Danish public. Citizens were vaccinated and tested regularly since it became required to document one’s status when traveling and physical y attending restaurants, shops, and other public areas. Consequently, a digital corona passport was introduced and the national website for health information sundhed.dk was updated with additional features to provide digital vaccination or test certification. Within two weeks, a time constraint was imposed by the government and sundhed.dk had to scale a digital infrastructure from accommodating a rough estimate of a few hundred daily user logins to 5500 concurrent users with a recorded peak load of 1.4 million daily logins. Modern online cloud services such as Amazon Web Services usual y accommodate unanticipated and immediate requirements for rapid scaling infrastructure. Still, due to regulations concerning sensitive health data – this was not an option. This leads to a research question: What critical technical and organizational decisions enabled the rapid scaling of sundhed.dk within two critical weeks during the COVID-19 pandemic? 2 Related literature In the literature, scalability covers a wide variety of subtopics. Within the organizational part of the information system literature, it is argued that scaling is not only a technical matter but also involves huge managerial efforts if a company N. Frederiksen, E. L. Møl er, J. Tuxen, S. E. O’Neil , M. Boesen: Rapid Scaling of a Danish Public Health System Under COVID-19 757 wants to scale successfuly. Abbott & Fisher (2015) argue that setting the right team with the right roles and clarified responsibilities is a critical prerequisite when scaling. Sahay & Walsham (2006) focus more on the socio-human aspects of scaling and allude to a variety of dilemmas associated with scaling. Some of them include standardized versus customized solutions, top-down versus bottom-up approaches, and appropriate versus complex technological solutions. Furthermore, they point our attention to considering scaling both concerning the increase in users of the systems and the number of members participating in the implementation team at different stages of the scaling and implementation process. In this regard, they distinguish between two approaches for scaling information systems, the cultivation approach, and the construction approach. The cultivation approach is a more incremental way of seeing scaling which favors a smooth, situated, and improvisational strategy changing smal er parts at a time while aligning those changes with the rest of the system. The cultivation approach stands against the construction approach which offers a more rational and planned approach to scaling emphasizes. It is argued that neither of the two should be prioritized over the other (Sahay & Walsham, 2006). In classical computer science, scalability has mainly been concerned with algorithm optimization, multithreading, processor optimization, or other techniques that al ow scalable performance and execution in a given context, such as a piece of hardware (Ahn et al., 2015; Rajan, 2010; Thierens, 1999; Vachharajani et al., 2005; Yeung, 1999). In other branches, scalability has been studied through the lens of software engineering and architecture, highlighting models, patterns, and processes to make a system scalable and evaluating the potential for scalability (Brataas & Hughes, 2004; Isoyama et al., 2012; Leesatapornwongsa et al., 2017; Mirakhorli et al., 2008; Pahl & Jamshidi, 2016; Rajan, 2010; Srinivas & Janakiram, 2005; Vaquero et al., 2011). The scale cube was introduced to structure at least some of the architectural scale options (Abbott & Fisher, 2015). The scale cube breaks down horizontal scaling into three dimensions. The three dimensions are defined as the x-axis, replication, y-axis, functional splitting, and z-axis, request splitting. In recent years, a main subtopic of academic interest within scalability has revolved around component-based systems or microservices (D’Antonio et al., 2004; Hasselbring, 2016; Kächele & Hauck, 2013; Lehrig et al., 2015; Márquez et al., 2018) to leverage the elastic horizontal/vertical scaling potential of online modern PAAS 758 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY or IAAS solutions in the cloud. However, in the field of healthcare services and other critical systems (Knight, 2002), data is considered sensitive and therefore under strict jurisdiction making it difficult to utilize cloud scalability (Heitmeyer, 2005; Walling, 2020). 3 Methodology The purpose of this study is to examine the successful rapid scaling of sundhed.dk that took place over two weeks during the COVID-19 pandemic as a juxtaposition to other public IS projects (Lauesen, 2020). By examining success stories from the novel context of the COVID-19 pandemic we hope to uncover and retain unheard experiences (Boéri & Giustini, 2023) and tacit knowledge (Schluter et al., 2008) of event-based IS development and scaling processes that evolved throughout the events. We focus on the period between the day when the Danish prime minister announces the partial re-opening of the Danish society until 14 days, later when the scaled health platform, sundhed.dk should be ful y functional and ready to handle the heavily increased user load. This two-week period constitutes the context of our interviews. Our selection of participants and interview design was inspired by the critical incident technique (CIT) (Flanagan, 1954), to gain entry to, and collect data about the events of the 14 days of rapid scaling (Cenfetelli & Schwarz, 2011) (Gogan et al., 2014). By framing our research interest in a CIT-inspired perspective we aim at identifying the important and relevant events of the rapid scaling process in the unexpected and unusual context of the COVID-19 pandemic. Our study draws on the technique from a phenomenological perspective and interpretative paradigm (Chell, 1998) rather than the positivist perspective from which it was original y developed. We are interested in subjective nuances that can bolster our understanding of what happened during the scaling of sundhed.dk, hence. N. Frederiksen, E. L. Møl er, J. Tuxen, S. E. O’Neil , M. Boesen: Rapid Scaling of a Danish Public Health System Under COVID-19 759 The critical incident interviews were planned and conducted according to Chel‘s (1998) eight distinguishable aspects of the method (preliminary design work, gaining access, introduction CIT, focusing the theme, controlling the interview, concluding the interview, ethical issues, and analyzing the data. Access was obtained from one of the IT architects involved in the scaling process who pointed out additional respondents inspired by the snowbal sampling method. The criteria that we gave to him was that he should choose those who were most intensively involved in the scaling process. We are aware of a potential selection bias regarding the number and criteria for selecting the respondents and we consider also including respondents with a more peripheral role in the scaling process such as managers, developers, and testers in the sample. After interviewing the four respondents, they were pointing out the same critical incidents and thereby we reached the level of saturation (Glaser & Strauss, 1967). The length of each interview was between forty-five minutes and two hours. All the interviews were conducted at the company which supported a natural and relaxed atmosphere for the respondents. A semi-structured interview guide was designed consisting of the following steps. First, the motivation, focus, and aim of the research were presented. The respondent was then asked to present himself including tasks and responsibilities related to the scaling process. To reconstruct the critical incidents in the scaling process a timeline was drawn to let the respondent point out the critical incidents graphically. The interview continued with a detailed focus on each of the identified incidents starting from the launch of the system and going backward from that. During the interview, the respondent was encouraged to describe both the organizational and technical aspects related to the specific incidents. During the interview, we gave attention to potential discrepancies in the respondents’ descriptions of the incidents and asked follow-up questions to clarify misunderstandings, thereby increasing the validity of the data. At the end of the interview, the respondent was asked whether we could contact them in case we needed to get some of the discussed topics and incidents clarified. For analyzing the data, we used analytical triangulation where each interview was processed by al four authors. When analyzing the data, we started describing and analyzing each of the identified critical situations inspired by the typical application of the CIT. However, this process revealed several themes across the incidents which were much more interesting to analyze than the individual incidents. Hence, we moved away from an event-based analysis to a more thematic analysis. 760 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 4 Preliminary results The following analysis will examine the prerequisites, diagnostics, and critical decisions made by the task force during the two weeks of intense work that transformed the critical infrastructure of sundhed.dk and enabled the platform to scale accordingly to user demand and regulatory requirements. The analysis will be divided into groups of scalability perspectives such as organizational aspects, virtualization, content delivery network, lazy-loading, and firewall interface configuration. The initial chal enge that was overcome was the formation of a task force itself. Before the rapid scaling requirements, sundhed.dk was limited to a quarterly deployment schedule and processes involving several confirmation steps, stakeholders and budgetary approval hereby presenting overhead for decision- making. Their solution was an appointment of the task force that conducted daily meetings with top-level management to present findings and approve solutions. The task force furthermore was organizational y relocated from development to operations to pre-emptively remedy chal enges caused by the principles of separation of duties. Several chal enges were identified by the task force from a technical standpoint. To horizontal y scale a platform, the platform should be able to perform and accommodate concurrent users by load-balancing requests between a dynamic set of virtual machines containing all required services. By fulfilling such requirements new hardware was introduced to scale horizontally – and proved to reduce overhead due to service communication located locally on virtual machines as opposed to network communication. Communication overhead between services interfacing was then identified as the result of an infrastructure relying on synchronous network cal s and timeouts. Timeout limits were reduced which in turn reduced response time for non-responding services and a large-scale design pattern change in the shape of circuit-breaking was proposed but was not implemented due to the time limitations presented. Consequently, their incoming and outgoing bandwidth was expanded, hardware was added to their prior setup, firewal -interface policy was configured to accommodate more users and an asymmetrical volume was identified between low-volume ingoing requests and high-volume outgoing responses. When identified, the task force implemented a content delivery network that provided additional static N. Frederiksen, E. L. Møl er, J. Tuxen, S. E. O’Neil , M. Boesen: Rapid Scaling of a Danish Public Health System Under COVID-19 761 insensitive content to clients such that bandwidth from within secured services provided the sensitive data. 5 Conclusion This research-in-progress paper explored a single unique case. In this case, a select group of Health IT expert practitioners was faced with a scaling problem. The problem progressed under a set of unprecedented circumstances caused by sudden and extensive user request increase during COVID-19. Due to the sensitive nature of health data, it was not an option to scale horizontally and elastically by utilizing cloud services meaning that infrastructure, application, platform, and network had to be configured to al ow the service to scale. It was found that the practitioners used a multiplicity of techniques to accommodate the increased number of requests such as reconfiguration of services to allow virtualization and horizontal scaling, lazy-loading, content delivery network of static content, and firewal interface configuration to name a few of which they successful y implemented within a short frame of time. References Ahn, J., Hong, S., Yoo, S., Mutlu, O., & Choi, K. (2015). A scalable processing-in-memory accelerator for paral el graph processing. Proceedings of the 42nd Annual International Symposium on Computer Architecture, 105–117. https://doi.org/10.1145/2749469.2750386 Boéri, J., & Giustini, D. (2023). Qualitative research in crisis: A narrative-practice methodology to delve into the discourse and action of the unheard in the COVID-19 pandemic. Qualitative Research. Brataas, G., & Hughes, P. (2004). Exploring architectural scalability. Proceedings of the 4th International Workshop on Software and Performance, 125–129. https://doi.org/10.1145/974044.974064 Chell, E. (1998). Critical Incident Technique. In G. Symon & C. Cassell (Eds.), Qualitative Methods and Analysis in Organizational Research (pp. 51–72). SAGE Publications. D’Antonio, S., Esposito, M., Romano, S. P., & Ventre, G. (2004). Assessing the scalability of component-based frameworks. ACM SIGMETRICS Performance Evaluation Review, 32(3), 34–43. https://doi.org/10.1145/1052305.1052311 Hasselbring, W. (2016). Microservices for Scalability. Proceedings of the 7th ACM/SPEC on International Conference on Performance Engineering, 133–134. https://doi.org/10.1145/2851553.2858659 Heitmeyer, C. (2005). Developing Safety-Critical Systems: The Role of Formal Methods and Tools. 10th Australian Workshop on Safety Related Programmable Systems. Isoyama, K., Kobayashi, Y., Sato, T., Kida, K., Yoshida, M., & Tagato, H. (2012). A scalable complex event processing system and evaluations of its performance. Proceedings of the 6th ACM International Conference on Distributed Event-Based Systems, 123–126. https://doi.org/10.1145/2335484.2335498 762 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Kächele, S., & Hauck, F. J. (2013). Component-based scalability for cloud applications. Proceedings of the 3rd International Workshop on Cloud Data and Platforms, 19–24. https://doi.org/10.1145/2460756.2460760 Knight, J. C. (2002). Safety critical systems. Proceedings of the 24th International Conference on Software Engineering - ICSE ’02, 547. https://doi.org/10.1145/581339.581406 Leesatapornwongsa, T., Stuardo, C. A., Suminto, R. O., Ke, H., Lukman, J. F., & Gunawi, H. S. (2017). Scalability Bugs. Proceedings of the 16th Workshop on Hot Topics in Operating Systems, 24– 29. https://doi.org/10.1145/3102980.3102985 Lehrig, S., Eikerling, H., & Becker, S. (2015). Scalability, Elasticity, and Efficiency in Cloud Computing. Proceedings of the 11th International ACM SIGSOFT Conference on Quality of Software Architectures, 83–92. https://doi.org/10.1145/2737182.2737185 Márquez, G., Villegas, M. M., & Astudillo, H. (2018). A pattern language for scalable microservices-based systems. Proceedings of the 12th European Conference on Software Architecture: Companion Proceedings, 1–7. https://doi.org/10.1145/3241403.3241429 Mirakhorli, M., Azim Sharifloo, A., & Shams, F. (2008). Architectural challenges of ultra large scale systems. Proceedings of the 2nd International Workshop on Ultra-Large-Scale Software-Intensive Systems, 45–48. https://doi.org/10.1145/1370700.1370713 Pahl, C., & Jamshidi, P. (2016). Microservices: A Systematic Mapping Study. Proceedings of the 6th International Conference on Cloud Computing and Services Science, 137–146. https://doi.org/10.5220/0005785501370146 Rajan, H. (2010). Building scalable software systems in the multicore era. Proceedings of the FSE/SDP Workshop on Future of Software Engineering Research, 293–298. https://doi.org/10.1145/1882362.1882423 Schluter, J., Seaton, P., & Chaboyer, W. (2008). Critical incident technique: a user’s guide for nurse researchers. Journal of Advanced Nursing, 61(1), 107–114. Srinivas, A. V., & Janakiram, D. (2005). A model for characterizing the scalability of distributed systems. ACM SIGOPS Operating Systems Review, 39(3), 64–71. https://doi.org/10.1145/1075395.1075401 Thierens, D. (1999). Scalability Problems of Simple Genetic Algorithms. Evolutionary Computation, 7(4), 331–352. https://doi.org/10.1162/evco.1999.7.4.331 Vachharajani, N., Iyer, M., Ashok, C., Vachharajani, M., August, D. I., & Connors, D. (2005). Chip multi-processor scalability for single-threaded applications. ACM SIGARCH Computer Architecture News, 33(4), 44–53. https://doi.org/10.1145/1105734.1105741 Vaquero, L. M., Rodero-Merino, L., & Buyya, R. (2011). Dynamically scaling applications in the cloud. ACM SIGCOMM Computer Communication Review, 41(1), 45–52. https://doi.org/10.1145/1925861.1925869 Walling, S. (2020). A Comprehensive Review on Cloud Computing and Cloud Security Issues. International Journal of Scientific Research in Computer Science, Engineering and Information Technology, 483–490. https://doi.org/10.32628/CSEIT206489 Yeung, D. (1999). The scalability of multigrain systems. Proceedings of the 13th International Conference on Supercomputing, 268–277. https://doi.org/10.1145/305138.305203 RESEARCH IN PROGRESS STRENGTHS, WEAKNESSES, OPPORTUNITIES AND THREATS TO BLOCKCHAIN IN INDUSTRIES BEYOND CRYPTO NILMINI WICKRAMASINGHE, NALIKA ULAPANE, AMIR ESLAMI ANDARGOLI Swinburne University of Technology, Melbourne, Australia nwickramasinghe@swin.edu.au, nalika.ulapane@gmail.com, aandargoli@swin.edu.au Despite the hype of the blockchain technology, the implementation and execution of blockchain technologies in sectors beyond cryptocurrency is lagging and below par. Understanding the reasons behind this lag is important to enable addressing any voids and enable making maximum use of the Keywords: blockchain, technology. We shed light on this void by trying to identify the catena-X, strengths, weaknesses, opportunities, and threats (SWOT) faced healthcare, by the use of blockchain technologies in industries beyond SWOT Analysis, crypto, and thereby, draw insights valuable to develop a Bled blockchain platform for healthcare. eConference DOI https://doi.org/10.18690/um.fov.4.2023.48 ISBN 978-961-286-751-5 764 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Blockchain is a technology for data sharing and has the key characteristics of transparency, anonymization, and decentralization (Sarmah, 2018). Blockchain was introduced for recording cryptocurrency transactions and today, the millions of crypto transactions being completed presents a solid use-case for blockchain technology. This has led to the exploration of prospects of blockchain technology in other data-driven sectors such as finance (Kahyaoglu & Aksoy, 2021), healthcare (Epiphaniou, et. al., 2019), government (Guarda, et. al., 2021), manufacturing and distribution (Kasten, 2020), and more. However, the implementation and execution of blockchain technologies in other sectors beyond crypto is lagging and is below par in contrast to the hype (Luthra, et. al., 2022). Therefore, understanding the reasons behind this lag is important to enable addressing key voids and thus, making maximum use of the technology. With the aim of serving to this gap, answering the following research question is attempted through this research in progress paper: What are the strengths, weakness, opportunities, and threats, faced by the use of blockchain technology in industries beyond crypto? In particular, we focus on private blockchains. While there are numerous ways that can be crafted to answer the aforesaid research question, we take a unique approach that would benefit our larger aim of designing, developing, and implementing a blockchain platform for healthcare. This project is currently ongoing, and an approach inspired by the Design Science Research Methodology (DSRM) (Hevner, et. al., 2010), (Hevner & Wickramasinghe, 2018) is followed. The focus of this paper is reporting on the outcomes of the “Problem Identification” phase our DSRM-inspired approach. Since our focus is towards development and implementation of a blockchain platform, the unique approach we take to answer our research question is by reviewing a commercial data sharing platform that has been recently rolled out and makes use of blockchain technologies. The commercial platform we have chosen for review is Catena-X (Catena-X, 2022). Catena-X is a state-of-the-art col aborative and open data ecosystem utilizing blockchain, that is currently gaining traction in Europe, especial y in the automotive sector. Catena-X stands as one of the earliest commercially available data ecosystems of its kind, and it is still in its early years of being rolled out. As such, Catena-X serves as an exemplary use case to understand N. Wickramasinghe, N. Ulapane, A. Eslami Andargoli: Strengths, Weaknesses, Opportunities and Threats to Blockchain in Industries Beyond Crypto 765 the chalenges especialy pertaining to the implementation and execution of blockchain platforms in different industries. We thus, answer our research question by performing a scoping review about the literature on Catena-X. 2 Method Followed for the Scoping Review We started by searching the keyword “Catena-X" OR "Catena X” in academic databases such as Scopus and IEEE Xplore. However, a minimal number of results were found. Therefore, we repeated the same keyword search in Google Scholar— a more inclusive database. The search was carried out between 4th and 6th of May 2022. This search resulted in 175 results. These included a mix of peer reviewed academic publications to grey literature such as industry reports and media articles. Since the search was done in Google Scholar, we had limitations in specifying where exactly the searched keywords would appear. For instance, in databases like Scopus, one can search for keywords specifical y within the article Titles and Abstract. However, Google Scholar offered limited capability to al ow such constriction. This meant that our results could include the keyword Catena-X anywhere in the text, for example even in the reference list or an Acknowledgement. Therefore, we performed an Abstract review to find out which articles were relevant. Our inclusion criteria were to include the items that discussed information systems in the Abstract. Items not related to information systems were excluded. This resulted in 16 articles being relevant, which had Catena-X mentioned within the text body of the article. These articles were taken forward for ful text review. From the full text review, five more articles were found to be irrelevant, as some of the articles had the name Catena mentioned for things other than the Catena-X platform. Thus, our review ended up with the 11 Articles. The 11 articles are the following: (Berg, et. al, 2021), (Garrido, et. al., 2022), (Johann, et. al., 2020), (Langdon and Schweichhart, Data Spaces), (Müller, et. al., 2022), (Ramesohl, et. al., 2022), (Sautter, 2021), (Schulz, et. al., 2022), (Staab, et. al., 2022), (Usländer, et. al., 2021) and (Zhongming, et. al., 2021). A flowchart of our search is available in the Results section (Figure 1). In the following step, we performed a SWOT analysis, i.e., find out the Strengths, Weaknesses, Opportunities, and Threats, related to Catena-X as reported in the available literature. The selected 11 articles were read in search of specific issues regarding Catena-X. The related text segments were tabulated, and themes were 766 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY assigned to the main points raised. These themes eventualy yielded a list of relevant issues. Since the rigor of critique was lacking nevertheless as said before, most of the points we found in these articles could be listed as “Opportunities”. Few of the issues pointed at “Threats”. However, the depth and richness of what we expected as “Strengths” and “Weaknesses” were not explicitly talked about in these articles. Therefore, for the sake of completeness, we went a step further, and looked at some other material for what was missing. Since we set out starting at a lens of fast-track learning for “implementation,” we referred to commercial and educational providers that offered Blockchain-related services rather than reading academic research. Thus, we selected IBM (IBM, 2022)—a renowned Blockchain service provider, and 101blockchains (101blockchains, 2022)—an education provider offering courses and training on Blockchain. We referred to the material published by IBM and 101blockcains and selected what they identified as Strengths and Weaknesses, thereby filling gas in our analysis. 3 Results Shown in Figure 1 is a flowchart of our literature search. The identified Strengths, Weaknesses, Opportunities, and Threats are listed in Table 1 as summary themes. Descriptions of these themes with elaborate meanings and text extracts are presented in Appendix A (provided as an external reference) along with the detailed full text analysis. We list out the Strengths as identified in the IBM article in (Benefits of Blockchain, 2022). The weaknesses are listed as identified in the 101blockchains article in (Disadvantages of Blockchain, 2022), and from the 11 works identified for the full text review. The Opportunities and Threats are listed from the findings of our literature review (i.e., the full text review of the 11 works). The findings are listed in Table 1. The frequency each theme was found in the literature is mentioned within parentheses in front of each theme in Table 1. A detailed analysis of the findings is provided in Appendix A (provided as an external reference). N. Wickramasinghe, N. Ulapane, A. Eslami Andargoli: Strengths, Weaknesses, Opportunities and Threats to Blockchain in Industries Beyond Crypto 767 Figure 1: Flowchart depicting the literature search 768 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 1: Summary of the SWOT Analysis Strengths Weaknesses Opportunities Threats • Enhanced • Node dependent • Increased • Lack of security (1) (1) efficiency (3) collaborative • Greater • Scalability • Environmental thinking (1) transparency chal enges (1) friendliness (3) • Lack of digital (1) • Interoperability • Enhanced trust (1) • Instant issues (1) col aboration (6) • Lack of financial traceability (1) • Energy • Smart factories (2) resources (1) • Increased consumption (1) • Team robotics (1) • Lack of skills (1) efficiency and • Immutability of • Transparent value • Lack of reliable speed (1) data (1) chain (1) technical • Automation • Not Completely • Optimized foundations (1) (1) Secure (1) production (1) • Lack of reliable • Responsibility on • Increased average legal the users (1) productivity (1) foundations (1) • Cost and • Increased • Lack of suitable implementation marginal business models struggle (1) productivity (1) (1) • Lack of maturity • Increased national • Requirement of (1) product (1) new • Require antitrust • Increased trust (1) infrastructure laws (1) • Digital twins (1) (1) • Require special • Increased • Requirement of corporate models transparency (1) new rules and (1) guidelines (1) • Increased security (1) • Requirement of IP protection, • Increased trade secrets (1) autonomy (2) • Enhanced interoperability (1) • Enhanced sustainability (1) • Improved infrastructure (1) • Improved guidelines (1) • Improved scalability (1) 4 Discussion This paper contributed by way of a SWOT analysis to the use of blockchain in different industries through reviewing some early works regarding Catena-X. Themes were assigned to the main issues identified and thereby lists were formed N. Wickramasinghe, N. Ulapane, A. Eslami Andargoli: Strengths, Weaknesses, Opportunities and Threats to Blockchain in Industries Beyond Crypto 769 and presented. These themes eventually yielded a list of relevant issues. Although our review contains limited sources, the importance of the results that can be found even at this level, must not be undermined as it does identify and envisage issues at an early stage is extremely important. Therefore, we consider the points found through our review of early articles related to Catena-X as a great starting point for planning the design and development of blockchain platforms for different sectors. 4.1 A Pathway for Healthcare Today, healthcare operations are generating volumes of critical and highly sensitive data including medical records in EMRs and genomic data. As such, it becomes vital for healthcare organizations and al healthcare stakeholders to consider responsible approaches to best manage these data and ensure appropriate levels of security and privacy to ensure the highest levels of trust between and within al parties involved in the healthcare sector. Becoming proactive in this pursuit is extremely important in this era of rapid digital transformation of healthcare. In this context, we see significant potential in private blockchains to increase data integrity in the healthcare sector. This paper has highlighted the key strengths, weaknesses, opportunities, and threats, of blockchain technologies being used in sectors outside crypto. The identified issues are certainly applicable to the healthcare sector as wel . 5 Conclusions The implementation and execution of blockchain technologies in sectors beyond cryptocurrency is lagging (Luthra, et. al., 2022). The reasons behind this lag must be understood to enable fil ing any voids to enable making maximum use of the technology. We set out in this paper to fil this void by identifying the strengths, weaknesses, opportunities, and threats (SWOT) faced by the use of blockchain technologies in industries beyond crypto. We take a unique approach to perform the SWOT analysis to benefit our larger aim of designing, developing, and implementing a blockchain platform for healthcare. The project of designing this blockchain platform is currently ongoing, and a DSRM-inspired approach is followed. This SWOT analysis serves as an outcome of the “Problem Identification” phase our design approach. We also discussed a 770 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY pathway suitable for the healthcare sector grounded on the findings of the SWOT analysis. Our future work will involve carrying out the subsequent steps of DSRM to develop a private blockchain prototype for a hospital and then trial it. The findings of our SWOT analysis will guide our design. Our findings are also generalizable for other industries that may hold interest in developing blockchain platforms or data ecosystems that would serve the unique needs of their sectors. Later, we expect to get the opportunity to review some other commercial data ecosystems as wel (e.g., SAP Datasphere) to confirm generality of our findings from Catena-X, or to learn differences. Furthermore, we wil attempt to discuss the identified themes in more detail; elaborating on how they relate to healthcare and perhaps other sectors as wel . Such attempts would assist in identifying ways to convert the identified weaknesses into strengths and threats into opportunities and thereby help the uptake of the blockchain technology in different sectors. Aknowledgements This project was funded by an Alexander von Humboldt award given to the first author. References Appendix A. Link: https://dit.swin.edu.au/clotsapp/Appendix_A.pdf Benefits of Blockchain. (2022). IBM Website. Link: https://www.ibm.com/au-en/topics/benefits-of-blockchain, date of last visit: July 28, 2022. Berg, H., Bendix, P., Jansen, M., Le Blévennec, K., Bottermann, P., Magnus-Melgar, M., . . & Wahlström, M. (2021). Unlocking the potential of Industry 4.0 to reduce the environmental impact of production. European Environment Agency, European Topic Centre on Waste and Materials in a Green Economy: Mol, Belgium. Catena-X. (2022). Catena-X website. Link: https://catena-x.net/en/ visited last: November 9, 2022. Disadvantages of Blockchain. (2022). 101blockchains Website. Link: https://101blockchains.com/disadvantages-of-blockchain/, date of last visit: July 28, 2022. Epiphaniou, G., Daly, H., & Al-Khateeb, H. (2019). Blockchain and healthcare. In Blockchain and Clinical Trial (pp. 1-29). Springer, Cham. Garrido, G. M., Sedlmeir, J., Uludağ, Ö., Alaoui, I. S., Luckow, A., & Matthes, F. (2022). Revealing the landscape of privacy-enhancing technologies in the context of data markets for the IoT: A systematic literature review. Journal of Network and Computer Applications, 103465. Guarda, T., Augusto, M. F., Haz, L., & Díaz-Nafría, J. M. (2021). Blockchain and Government Transformation. In International Conference on Information Technology & Systems (pp. 88-95). Springer, Cham. Hevner, A., & Chatterjee, S. (2010). Design science research in information systems. In Design research in information systems (pp. 9-22). Springer, Boston, MA. N. Wickramasinghe, N. Ulapane, A. Eslami Andargoli: Strengths, Weaknesses, Opportunities and Threats to Blockchain in Industries Beyond Crypto 771 Hevner, A. R., & Wickramasinghe, N. (2018). Design science research opportunities in health care. Theories to Inform Superior Health Informatics Research and Practice, 3-18. IBM. (2022). IBM Website. Link: https://www.ibm.com/au-en/, date of last visit: July 28, 2022. Johann, H., Klein, M., & Rieger, A. (2020). EUROPEAN COMMISSION HORIZON 2020 LC-MG-1-4-2018 Grant agreement ID: 814951. Kahyaoglu, S. B., & Aksoy, T. (2021). Survey on Blockchain Based Accounting and Finance Algorithms Using Bibliometric Approach. In Accounting and Finance Innovations. IntechOpen. Kasten, J. E. (2020). Engineering and manufacturing on the blockchain: A systematic review. IEEE Engineering Management Review, 48(1), 31-47. Langdon, C. S., & Schweichhart, K. Data Spaces: First Applications in Mobility and Industry (Part IV Solutions & Applications). Luthra, S., Janssen, M., Rana, N. P., Yadav, G., & Dwivedi, Y. K. (2022). Categorizing and relating implementation chal enges for realizing blockchain applications in government. Information Technology & People, (ahead-of-print). Mül er, T., Gaertner, N., Verzano, N., & Matthes, F. (2022). Barriers to the Practical Adoption of Federated Machine Learning in Cross-company Col aborations. In ICAART (3) (pp. 581-588). Ramesohl, S., Berg, H., & Wirtz, J. (2022). The circular economy and digitalisation-strategies for a digital-ecological industry transformation: a study commmissioned by Huawei Technologies Germany GmbH. Sarmah, S. S. (2018). Understanding blockchain technology. Computer Science and Engineering, 8(2), 23-29. Sautter, B. (2021). Shaping Digital Ecosystems for Sustainable Production: Assessing the Policy Impact of the 2030 Vision for Industrie 4.0. Sustainability, 13(22), 12596. Schulz, W. H., Franck, O., Smolka, S., & Geilenberg, V. (2022). Applicable Knowledge for Sustainability. The Status of Artificial Intelligence in Industrial Production and the Impact of Future Sustainability. In The New Digital Era: Digitalisation, Emerging Risks and Opportunities (Vol. 109, pp. 117-124). Emerald Publishing Limited. Staab, P., Pietrón, D., & Hofmann, F. (2022). Sustainable Digital Market Design: A Data-Based Approach to the Circular Economy. Usländer, T., Schöppenthau, F., Schnebel, B., Heymann, S., Stojanovic, L., Watson, K., . . & Morinaga, S. (2021). Smart Factory Web—A Blueprint Architecture for Open Marketplaces for Industrial Production. Applied Sciences, 11(14), 6585. Zhongming, Z., Linong, L., Xiaona, Y., Wangqiang, Z., & Wei, L. (2021). Automotive industry leaders partner to develop a shared approach to carbon emissions data. 101blockchains. (2022). 101blockchains Website. Link: https://101blockchains.com/, date of last visit: July 28, 2022. 772 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY RESEARCH IN PROGRESS ORGANIZATIONS IN DIGITAL TRANSFORMATION (ODT): A LITERATURE REVIEW ON THE IMPLICATIONS FOR MALLEABLE ORGANIZATION AND HRM XIAO PENG, MACHIEL BOUWMANS University of Applied Sciences Utrecht, Utrecht, Netherland xiao.peng@hu.nl, machiel.bouwmans@hu.nl The digital age has brought significant changes in the way organizations operate and compete. As a result of rapid technology development, many organizations are undergoing a digital transformation to stay relevant and competitive in the marketplace. This literature review aims to find future research topics by providing an overview of the current state of research on organizations in digital transformation (ODT), especial y on mal eable organization design and HRM aspects. The article begins by defining digital transformation (DT), and then Keywords: examines how organizations change during DT, before delving digital transformation, into the perspectives of mal eable organization design and HRM. organizations, Finally, it concludes by identifying gaps in the literature and malleable suggesting a research agenda for future. Overall, organizational organization, HRM, factors that need more investigation are highlighted to tackle Bled complexities of ODT for further research. eConference DOI https://doi.org/10.18690/um.fov.4.2023.49 ISBN 978-961-286-751-5 774 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction In recent years, rapid and continuous advancements in technologies lead to systemic changes across governmental, economic, social and organizational domains (Ivaldi et al., 2022). Organizations have been facing increasing pressure to digitally transform to stay competitive and meet the ever-evolving needs of their stakeholders. The way organizations operate and interact with their customers, partners, and employees has therefore drastical y changed. Within digital transformation (DT), emphasis is placed on human-centricity and sustainability, as finding balance between technology and humans is crucial, and promoting sustainable development goals such as socio-environmental sustainability and resilience is the priority (Ghobakhloo et al., 2022; Reiman et al., 2021; Zizic et al., 2022). The goal of this work-in-progress paper is to develop a future research agenda on the topics of human-centricity and sustainability of organizations in DT by reviewing the literature. 2 Theoretical background Although DT is a central topic in many recent studies, there is much conceptual confusion about what exactly it entails. Therefore, it is essential to clearly define DT before addressing the chal enges in relation to human-centricity and sustainability it poses to organizations. The theoretical background of this article is based on two systematic literature reviews: the definition of DT developed by Gong and Ribiere (2021) as a starting point, followed by an overview of changes in organizations in digital transformation (ODT) described by (Hanelt et al., 2021). DT requires new ways of working and thinking, such as new strategies, structures, processes, operations and organizational culture. Successful DT leads to strategical benefits for organizations, unlike digitalization and digitization which mainly impact operational work within the organization (Gong & Ribiere, 2021). DT can be defined as: X. Peng, M. Bouwmans: Organizations in Digital Transformation (ODT): A Literature Review on the Implications for Malleable Organization and HRM 775 “A fundamental change process, enabled by the innovative use of digital technologies accompanied by the strategic leverage of key resources and capabilities, aiming to radically improve an entity* and redefine its value proposition for its stakeholders. (*An entity could be: an organization, a business network, an industry, or society.)” (Gong & Ribiere, 2021, p.12) DT leads to developments in two major areas for organizations (Hanelt et al., 2021): malleable organizational designs (internally) and digital business ecosystems (external y). A malleable organizational design is one that adapts easily and quickly to change through digital technologies and agile structures. A digital business ecosystem has rapidly changed value propositions which cause participants, positions, and roles of the organization to constantly evolve in a turbulent environment. Because of the progression and complexity of DT established theoretical models do not fit anymore with the new empirical reality (Hanelt et al., 2021), a broader view on organizational changes and innovation associated with ODT is required. This study adds to developing that broader view by addressing the chal enges that DT brings in relation to human-centricity in organizations and sustainability, by specifical y focusing on human resource management (HRM) and a malleable organization for ODT. 3 Research method A literature review is conducted using the PRISMA 2020 approach (Page et al., 2021) to undercover the topics of ODT. The PRISMA 27-item checklist was applied for a transparent and thorough manner. The literature search follows screening stages with defined criteria, which is visualized in a flow diagram. See Figure 1 with explanations of screening stages and criteria in the Appendix. The search strategy is an advanced search action in Web of Science database with a final inclusion of 40 articles for this article after screening. As the article presents work in progress, it only includes the current literature search on malleable orgnization and HRM for ODT. 776 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 4 Preliminary results From selected literature, some preliminary results are found regarding malleable organization and HRM for ODT that help formulating a future research agenda. A malleable organization for ODT A stable digital malleable organization is expected to be established where stakeholders connect online and interact digitally in dynamics that create value for all (Zizic et al., 2022). Investigating a digital malleable organization is a necessary step in understanding how digital technologies can be leveraged to benefit society as a whole. A malleable orgnization requires a corresponding strategy, business model and organizational culture, and a proper alignment of them (González-Díaz et al., 2021). Our literature review on the malleable orgnization starts by examining the adoption of organizational agility as the latest approach for operating and managing ODT. Subsequently, the review explores sustainable business models, as well as people and culture within ODT. A malleable orgnization requires a new approach for operation and management with organizational agility as the core strategy. (Reuschl et al., 2022) named it as organizational elasticity, which means organizations can take quick decisive actions for changes and be flexible to either return to old or adapt to new routines after a crisis. (Hanelt et al., 2021) called this malleable organizational designs, as previously explained. Innovation is essential for agility and it enables organizations to stay sustainable, resilient and to create long-term value when dealing with the digital change and disruption (Di Vaio et al., 2021). Simultaneously with innovation, understanding and applying change management is required for successful ODT (Florek-Paszkowska et al., 2021). Change management - the process of planning, implementing, and monitoring changes in organizations and ensure that they are successful and sustainable – is needed because innovation often requires changes to an organization's processes, systems, and culture. This involves regularly reassessing various aspects of the organization, reflecting on their efficacy, evaluating their impact, and taking appropriate action. In the process of innovating and applying change management, knowledge is a key which can be not only created but also shared via digital advancement, for example the use of big data to predict the market so that the organization can respond accordingly (Di Vaio et al., 2021), or constantly X. Peng, M. Bouwmans: Organizations in Digital Transformation (ODT): A Literature Review on the Implications for Malleable Organization and HRM 777 involving all relevant stakeholders at all stages for knowledge co-creation and sharing via a liquid network (Borchardt et al., 2022; Schiavone et al., 2022). Besides organizational agility as a core strategy, both a sustainable business model and a supportive organizational culture are critical components that must be consistent and in sync with each other. In general, a dynamic, adaptive business model is suggested for a sustainable and resilient ODT (Florek-Paszkowska et al., 2021), that not only sustains current core business, but also invest in new emerging external opportunities (Li, 2020). Specifically (Gruenbichler et al., 2021) and (Andersen et al., 2022) revealed some case studies of dynamic and adaptive business models with empirical data in SMEs. (Gruenbichler et al., 2021) investigated business performance management (BPM) which serves as a measurement tool to achieve business objectives successful y while enhancing the competitiveness and sustainability of ODT. (Andersen et al., 2022) explored business model innovation (BMI), confirming that a agile and data driven nature is important for the whole business process including identifying opportunities and making decisions. Finally, it should be noted that individuals and the prevailing organizational culture are essential factors for fostering a prosperous malleable organization in DT. Innovation requires continuous re-skilling and up-skil ing of both leaders and employees, and with strong commitment and engagement during disruptive DT, e.g. an agile culture (Florek-Paszkowska et al., 2021). The next sections delves further into the human-centric aspect of ODT. Strategic HRM for ODT Although DT is driven by rapid and continuous advancements in digital technologies, it depends largely on the human factor: a combination of and synergy between top-down initiatives and bottom-up acceptance and commitment determine its success (Gong & Ribiere, 2021). Therefore, multiple studies address the human-related chal enges that organizations face as a result of DT. On the one hand, these chal enges refer to the knowledge, skil s and attitudes of employees, including lack of employees’ awareness of newly available digital technologies, lack of adequate skills, and confidence caused by the disruptive DT changes. On the other hand, these chal enges refer to the adaptation of HRM practices, such as talent acquisition, retainment and employees training (James et al., 2022). These chal enges underline that a successful DT strategy should include a refined HRM strategy that 778 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY anticipates and responds to human-related chalenges during DT (Bamel et al., 2022; Vereycken et al., 2021). The included articles show that DT requires adaptation of HRM strategy and practices aimed at the fol owing aspects. Firstly, strategic talent management - a systematic process of recruiting, developing and maintaining the best talents for the organization - should be (re-)designed to fit the DT context (Kuchciak & Warwas, 2021). Key elements in this (re-)design are recruitment, re- and upskilling, and redeployment. Recruitment should emphasize acquiring new talents who bring essential DT skills that are not yet sufficiently present in the organization, as a new target group to trigger change in the organization. An example of this are “digital talents”, who possess essential IT-related skills (Gilch & Sieweke, 2021). Regarding re- and upskilling, DT creates an urgent need for continuous development of employees, so that employees do not become obsolete, can co-exist, collaborate with, and benefit from new technologies, and can increase their confidence in working in digitalized workplace (Gjika & Pano, 2022; Kuchciak & Warwas, 2021; Vereycken et al., 2021). These skills include not only relevant technical skills but also non-technical, transformational skills, such as creativity, workplace learning, problem solving, conceptual thinking, entrepreneurship and resilience (Bamel et al., 2022; Kuchciak & Warwas, 2021). Re-and upskil ing practices should be flexible and sustainable to respond to continuous technological developments (Piwowar-Sulej, 2021). Moreover, constant support is required during these practices, such as monitoring of employees’ skil s and development needs, offering training and mentoring, and organizing work in such ways that it stimulates learning, e.g. by creating diverse and multidisciplinary teams (Kuchciak & Warwas, 2021). Redeployment - assigning employees to new, fitting, tasks - is becoming increasingly important as work can change tremendously and employees may no longer have fitting roles or tasks. Relevant practices include helping employees to plan career development paths, inform them about promotion opportunities, address job offers to diverse employees, develop work-life balance programs, and introduce wel -being audits (Kuchciak & Warwas, 2021). Secondly, HRM strategies and practices should be adapted to increase employee involvement and empowerment though job design. As new technologies take over physical, sensorial and cognitive-based tasks, many professional tasks will shift from predominantly operations-related tasks to more complex tasks such as coordination and decision-making. This is only possible when employees are granted with more X. Peng, M. Bouwmans: Organizations in Digital Transformation (ODT): A Literature Review on the Implications for Malleable Organization and HRM 779 autonomy. It requires an organizational context that is characterized by a flat hierarchy, decentralized decision making, and employee empowerment (Vereycken et al., 2021). HR practices that promote employee involvement and empowerment are expected to increase employee acceptance of the changes that DT brings, stimulate employees’ contribution to these changes (Vereycken et al., 2021), and facilitate employees’ work-life balance and wel being because they experience a high degree of control over work, can adapt to changes, pick up technologies that make their work more effective and efficient, and experience fulfil ment (Bamel et al., 2022). Finally, the complexity of DT emphasizes that HR professionals need to rethink their own roles. HR professionals should increasingly act as consultative, strategic, business partners and change agents, who makes sense of trends in digital technologies and understand the consequences for the organization and its employees, understand new ways of working and how different parts of the organization wil interact during and as a result of DT, and promote and support the needed organizational change (Dhanpat et al., 2020). 5 Conclusion and discussion The interdependencies and co-constitutive relationship between organizing and technology, the tensions between ways of knowing and the value laden nature of decision making in ODT call for research that effectively engages insights from multiple disciplines that coalesce in the questions that rise from the phenomenon of DT. Malleable organizational designs and digital business ecosystems are under development for ODT, however progression and complexity of DT stil requires a broader view on organizational change and innovation of DT (Hanelt et al., 2021). More research is needed to examine the design and testing of malleable organization for ODT. Specifically, further exploration is required to determine how a malleable organization can effectively facilitate open innovation by connecting stakeholders to share and co-create knowledge, and react quickly to disruptive changes with optimal solutions. Moreover, future research could further examine how malleable organizations impact traditional business models, including what are the successful factors for ODT business models and how people and culture of ODT should 780 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY respond accordingly. Furthermore, it is important to study the alignment between strategy, business model, and organizational culture, as current activities in digital transformation are being carried out at a rapid pace without taking this alignment into account (González-Díaz et al., 2021). Although key HRM areas for adaptation are identified in the results section, in-depth insights into the content of relevant HRM strategies and practices in response to DT remains limited and fragmented (Bamel et al., 2022). How HRM strategy and practices should be refined requires more fine-grained research into the specific changes that DT brings, and how HRM should anticipate and respond to them. For instance, the implementation of advanced robotics in the manufacturing industry wil require a different HR strategy than the need for data analytics to optimize processes in healthcare (Vereycken et al., 2021). Moreover, research should take different organizational contexts into consideration, as similar technology implementations may lead to different requirements for HRM strategies and practices depending on this context (Gjika & Pano, 2022). It is therefore important to conduct this fine-grained research in the near future to gain more insight into required differences as wel as similarities in successful HRM strategies and practices. References Andersen, T. C. K., Aagaard, A., & Magnusson, M. (2022). Exploring business model innovation in SMEs in a digital context: Organizing search behaviours, experimentation and decision-making. Creativity and Innovation Management, 31(1), 19–34. https://doi.org/10.1111/CAIM.12474 Bamel, U., Kumar, S., Lim, W. M., Bamel, N., & Meyer, N. (2022). Managing the dark side of digitalization in the future of work: A fuzzy TISM approach. Journal of Innovation and Knowledge, 7(4). https://doi.org/10.1016/J.JIK.2022.100275 Borchardt, M., Pereira, G. M., Milan, G. S., Scavarda, A. R., Nogueira, E. O., & Poltosi, L. C. (2022). Industry 5.0 Beyond Technology: An Analysis Through the Lens of Business and Operations Management Literature. Organizacija, 55(4), 305–321. https://doi.org/10.2478/orga-2022-0020 Dhanpat, N., Buthelezi, Z. P., Joe, M. R., Maphela, T. V., & Shongwe, N. (2020). Industry 4.0: The role of human resource professionals. SA Journal of Human Resource Management, 18(0), 11. https://doi.org/10.4102/SAJHRM.V18I0.1302 Di Vaio, A., Palladino, R., Pezzi, A., & Kalisz, D. E. (2021). The role of digital innovation in knowledge management systems: A systematic literature review. Journal of Business Research, 123, 220– 231. https://doi.org/10.1016/J.JBUSRES.2020.09.042 Florek-Paszkowska, A., Ujwary-Gil, A., & Godlewska-Dzioboń, B. (2021). Business innovation and critical success factors in the era of digital transformation and turbulent times. Journal of Entrepreneurship, Management and Innovation, 17(4), 7–28. https://doi.org/10.7341/20211741 X. Peng, M. Bouwmans: Organizations in Digital Transformation (ODT): A Literature Review on the Implications for Malleable Organization and HRM 781 Ghobakhloo, M., Iranmanesh, M., Mubarak, M. F., Mubarik, M., Rejeb, A., & Nilashi, M. (2022). Identifying industry 5.0 contributions to sustainable development: A strategy roadmap for delivering sustainability values. Sustainable Production and Consumption, 33, 716–737. https://doi.org/10.1016/J.SPC.2022.08.003 Gilch, P. M., & Sieweke, J. (2021). Recruiting digital talent: The strategic role of recruitment in organisations’ digital transformation. German Journal of Human Resource Management, 35(1), 53–82. https://doi.org/10.1177/2397002220952734/FORMAT/EPUB Gjika, I., & Pano, N. (2022). Human resource development AS a contributor to industry 4.0 implementation IN Albania . THE ELECTRONIC JOURNAL OF INFORMATION SYSTEMS IN DEVELOPING COUNTRIES. https://doi.org/10.1002/ISD2.12250 Gong, C., & Ribiere, V. (2021). Developing a unified definition of digital transformation. Technovation, 102. https://doi.org/10.1016/j.technovation.2020.102217 González-Díaz, R. R., Guanilo-Gómez, S. L., Acevedo-Duque, Á. E., Campos, J. S., & Cachicatari Vargas, E. (2021). Intrinsic alignment with strategy as a source of business sustainability in SMEs. Entrepreneurship and Sustainability Issues, 8(4), 377–388. https://doi.org/10.9770/JESI.2021.8.4(22) Gruenbichler, R., Klucka, J., Haviernikova, K., & Strelcova, S. (2021). Business performance management in small and medium-sized enterprises in the slovak republic: An integrated three-phase-framework for implementation. Journal of Competitiveness, 13(1), 42–58. https://doi.org/10.7441/JOC.2021.01.03 Hanelt, A., Bohnsack, R., Marz, D., & Antunes Marante, C. (2021). A Systematic Review of the Literature on Digital Transformation: Insights and Implications for Strategy and Organizational Change. Journal of Management Studies, 58(5), 1159–1197. https://doi.org/10.1111/JOMS.12639 Ivaldi, S., Scaratti, G., & Fregnan, E. (2022). Dwelling within the fourth industrial revolution: organizational learning for new competences, processes and work cultures. Journal of Workplace Learning, 34(1), 1–26. https://doi.org/10.1108/JWL-07-2020-0127/FULL/PDF James, A. T., Kumar, G., Tayal, P., Chauhan, A., Wadhawa, C., & Panchal, J. (2022). Analysis of human resource management chal enges in implementation of industry 4.0 in Indian automobile industry. Technological Forecasting and Social Change, 176. https://doi.org/10.1016/J.TECHFORE.2022.121483 Kuchciak, I., & Warwas, I. (2021). Designing a Roadmap for Human Resource Management in the Banking 4.0. Journal of Risk and Financial Management 2021, Vol. 14, Page 615, 14(12), 615. https://doi.org/10.3390/JRFM14120615 Li, F. (2020). Leading Digital Transformation: Three Emerging Approaches for Managing the Transition. International Journal of Operations and Production Management . https://doi.org/10.1108/IJOPM-04-2020-0202 Page, M. J., Moher, D., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanvil e, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., Mcdonald, S., … Mckenzie, J. E. (2021). PRISMA 2020 explanation and elaboration: Updated guidance and exemplars for reporting systematic reviews. In The BMJ (Vol. 372). BMJ Publishing Group. https://doi.org/10.1136/bmj.n160 Piwowar-Sulej, K. (2021). Human resources development as an element of sustainable HRM – with the focus on production engineers. Journal of Cleaner Production, 278. https://doi.org/10.1016/J.JCLEPRO.2020.124008 Reiman, A., Kaivo-oja, J., Parviainen, E., Takala, E. P., & Lauraeus, T. (2021). Human factors and ergonomics in manufacturing in the industry 4.0 context – A scoping review. Technology in Society, 65. https://doi.org/10.1016/J.TECHSOC.2021.101572 Reuschl, A. J., Deist, M. K., & Maalaoui, A. (2022). Digital transformation during a pandemic: Stretching the organizational elasticity. Journal of Business Research, 144, 1320–1332. https://doi.org/10.1016/J.JBUSRES.2022.01.088 782 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Schiavone, F., Leone, D., Caporuscio, A., & Lan, S. (2022). Digital servitization and new sustainable configurations of manufacturing systems. Technological Forecasting and Social Change, 176. https://doi.org/10.1016/J.TECHFORE.2021.121441 Vereycken, Y., Ramioul, M., Desiere, S., & Bal, M. (2021). Human resource practices accompanying industry 4.0 in European manufacturing industry. Journal of Manufacturing Technology Management, 32(5), 1016–1036. https://doi.org/10.1108/JMTM-08-2020-0331/FULL/XML Zizic, M. C., Mladineo, M., Gjeldum, N., & Celent, L. (2022). From Industry 4.0 towards Industry 5.0: A Review and Analysis of Paradigm Shift for the People, Organization and Technology. In Energies (Vol. 15, Issue 14). MDPI. https://doi.org/10.3390/en15145221 X. Peng, M. Bouwmans: Organizations in Digital Transformation (ODT): A Literature Review on the Implications for Malleable Organization and HRM 783 Appendix Identification of articles on DT and sustainability Identification of articles on DT and HRM Articles identified from Articles identified from ation databases: databases: Web of Science (n = 182) Web of Science (n = 146) entific Search date: 06/02/2023 Search date: 17/01/2023 Id Articles screened for eligibility (n = 182) Articles excluded in total Articles screened for eligibility Articles excluded (n = 127) (n = 146) (n = 89) Articles sought for retrieval Articles not retrieved Articles sought for retrieval Articles not retrieved (n = 55) (n = 4) (n = 57) (n = 13) ing een Scr Ful -texts assessed for eligibility Ful -texts assessed for eligibility Articles excluded: (n = 51) Articles excluded: (n = 44) o o Focus on eHRM or the DT Sustainability for the society of HRM (n = 19) (n = 22) o No specific focus on HRM strategy or practices to support DT (n = 14) Articles included in review Articles included in review (n = 29) (n = 11) Included Total articles included in review (n =40) Figure 1: Adapted PRISMA 2020 flow diagram For the search action in Web of Science multiple search terms were combined into two Boolean search operator, one for sustainability of ODT and one for HRM of ODT. This resulted in an identification of in total of 328 articles. After identification, articles were screened in two stages. In stage one, titles and abstracts of al articles were screened for the eligibility criteria mentioned below. This resulted in the exclusion of 233 articles, including articles that could not be retrieved. For each article, reasons for exclusion were provided. In stage two, the ful -texts of the remaining 95 articles were checked. This resulted in a final inclusion of 40 articles from the Web of Science database. 784 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Papers needed to meet the folowing eligibility criteria: 1. Be published in a peer-reviewed journal. 2. Be written in English. 3. DT and HRM: papers should focus on the (changing) role of HRM to support successful DT of organizations. Papers that focused on eHRM or on the DT of HRM itself (e.g. digital recruitment or HR analytics) are not yet included in this present article. 4. DT and sustainability: papers should not be specifical y focused on one field, for example supply chain, and papers with empirical application only on specific countries are not included in this present article. In stage 2, only papers with a focus on sustainability regarding organization are included. Papers on sustainability as a societal development are excluded in this present article. RESEARCH IN PROGRESS IMPROVING SAFETY AT SEAS JUKKA HEIKKILÄ, MARIKKA HEIKKILÄ University of Turku, Turku School of Economics, Turku, Finland jups@utu.fi, marikka.heikkila@utu.fi Digitalization in the maritime industry has been suggested as means to achieve safe, economically, and environmentally sound marine traffic. Despite the increasing adoption of digital services and technology in the maritime industry in Finland, the number of accidents has not reduced significantly. This research article aims to explore the potential for improving waterway safety in Finland. The study analyzes waterway accident data to identify the user group most in need of safety improvements and proposes targeted safety measures better tailored to recreational boaters. The study suggests enhancing the physical infrastructure and related digital services of sea areas to provide better Keywords: coordinated traffic at sea and situational awareness among digitalization, waterway seafarers, including non-professional seafarers. Finally, a safety, roadmap for implementing these safety measures will be recreational developed. The research provides a comprehensive framework boaters, maritime, for improving waterway safety in Finland, reducing accidents, accident and fatalities. reduction DOI https://doi.org/10.18690/um.fov.4.2023.50 ISBN 978-961-286-751-5 786 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The pace of international development in traffic system automation has been rapid, driven by the goal of reducing human errors in traffic through increased automation and autonomy of vehicles and traffic control. Seafaring is no exception: more digital services have been approved by authorities and adopted by merchant and private mariners. Remotely operated and autonomous ships are being increasingly trialed and developed in the maritime industry (Aawa, 2016; Kepesedi, 2022). Equipping ships and fairways with positioning, navigation, sensoring, and communication technology has been suggested to achieve safe, economical y, and environmental y sound marine traffic for all (socially sustainable seafaring) as communicated by the ministry in their societal objectives (Miettinen et al., 2021). However, despite these improvements we see limited progress in reducing the total number of accidents – the improvements are due to safer merchant traffic of big vessels - most accidents happened outside merchant traffic by recreational seafarers in 2017-2020 (Statistics Finland, 2022). The aim of this research article is to explore the potential for improving the safety of waterway users in Finland by leveraging the opportunities provided by digitalization for vessels and waterways. Specifically, we aim to identify the user group that require the most attention in terms of improving their safety, assess the suitability of planned safety elements for addressing their needs, and develop a roadmap for implementing these elements to ensure the safety of al waterway users. To determine the user group most in need of safety improvements, we conduct an analysis of waterway accident data. Based on this analysis, we propose targeted safety measures that are better tailored to recreational boaters. We suggest enhancing both the physical infrastructure and related digital services of territorial sea to provide better coordinated traffic at sea and situational awareness among seafarers - meaning al seafarers and even beyond official fairways. We believe that technology and its usage has reached a point where even non-professional seafarers can utilize digitally-enhanced fairway infrastructure for safer and more environmental y-friendly travel. Overall, this research-based development intiative provides a framework for improving waterway safety, leveraging the opportunities provided by digitalization for vessels and waterways. By identifying recreational boaters as the user group most J. Heikkilä, M. Heikkilä: Improving Safety at Seas 787 in need of safety improvements and tailoring safety measures to their specific needs, we aim to ensure the safety and reduce the number of accidents and fatalities in Finnish waters. 2 Analysis of accident data In recent years, several new elements have been introduced to improve the safety of merchant traffic on well maintained fairways, resulting in a significant decrease in the number of accidents involving large ships. In fact, merchant shipping accounts for only 1-3% of the total accidents, and fatalities are very rare today. However, despite this improvement, relaxed practices and an increasing number of vessels with high speed differences have led to more near-misses and limited progress in reducing the total number of accidents. More than 2000 accidents and approximately 50 annual fatal incidents happened outside merchant traffic in 2017-2020 (2022, Statistics Finland). Over 90% of these fatalities happen to recreational crafts and are a consequence of operating error while drunk leading to listing or cap sizing (ibid.). But the picture changes when looking into the non-lethal accidents: they mostly happen under good daytime conditions during the weekends (ibid.). Of the accidents, motorboats form a clear majority (70%), but rigid inflatable boats (ribs) and sailboats are over-represented in accident statistics. The root causes for accidents are technical failures (68%) such as malfunctions of manoeuvring devices or engines, and human error/misconduct in 37% the accidents. Weather conditions are the reason for approximately 5% of the accidents. As a consequence, we suggest that the most potential area for the development of safety at seas is to provide means for recreational watercrafters (motorboats, ribs and sailboats) to avoid, react and receive help for malfunctions, groundings, allusions, and collisions, and ultimately also receive search and rescue services (SAR) quickly enough on the accident site. This calls for a better joint situational awareness among all seafarers and safety organizations. 788 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3 Digitalisation of vessels and traffic At open seas, modern vessels have been semiautomatic for a long time, and this is now spreading towards regular traffic between harbours and to some extent, also recreational traffic. However, bigger vessels have to fulfil the requirements set by Convention on Standards of Training, Certification and Watchkeeping for Seafarers (STCW, 1978). The trials with remote piloting and autonomous traffic (Aawa, 2016; Kepesedi, 2022) have shown the need for enhanced information to improve the perceptive, automated capabilities of the vessel and to provide potential for better situational awareness of the traffic situation for the mariners, e.g. on fairways or in the port area or on the fairways. Categories of autonomous vessels are defined according to the operator’s role on a ship (IMO, 2021). A Maritime Autonomous Surface Ship (MASS) is “a ship which, to a varying degree, can operate independently of human interaction”. IMO identified four ’ degrees of autonomy’, of which two are manned, i.e. seafarers are on board, and two are unmanned (ibid.): Manned vessels are assisted with automated operation processes and decision support systems (type 1), even remotely control ing the vessel (type 2), but the crew can intervene and override the systems. When there is no crew on board the vessel is either remotely operated (type 3) or operating autonomously (type 4) without crew intervention (table 1). Table 1: Degrees of autonomy Manned Type 1: Automated operation and Type 2: Remote control, but crew can intervene Vessels DSS Helping merchant traffic on restricted Both in merchant and areas at harbours and piloting. increasingly in recreational traffic No Type 3: Remote operation: Regular Type 4: Autonomous operation: crew on merchant traffic in restricted Trials in merchant traffic (test beds and board areas and waterways test areas). As vessels are turning more automated and autonomous, towards types 3 and 4, there is a greater need for careful analysis of faults to improve. This involves continuously monitoring and improving the reliability and performance of the systems and their components’ quality (Chaal et al., 2022), to create Safety J. Heikkilä, M. Heikkilä: Improving Safety at Seas 789 Instrumented Systems (SIS) with relative levels of risk-reduction (Safety Integrity Levels, SIL) for continous safety improvement. In navy and merchant shipping, these ideas of automation is slowly becoming a reality towards types 3 and 4 (e.g. Kepesedi, A., 2022). The first trials show, that the behaviour of such vessels is different from traditional ships: they are following pre-planned routes, with limited set of maneuvers (Safety and Regulations for European Unmanned Maritime Systems1 and SOLAS, 1974). This is to optimize time and fuel consumption for profit, to meet environmental and emission requirments and, only secondly, to participate in the communication, interaction with other traffic, not to mention participating in rescue operations in traditional sense at waters. For recreational vessels, the first autonomous category, i.e. providing navigational aids – to increasing extent also decision support systems – are becoming common integrated with ECDIS2 for new vessels. Still, recreational vessels are mostly relying on the traditional watchkeeping and elementary safety measures at seas. However, lately improved self-services are being introduced to track merchang vessels and show maritime alerts (e.g. https://aluskartta.com; https://extranet.vayla.fi/pooki_www/merivaroitukset/list_fi.html). Summary of Research in Progress This is the outset of our research and development project aimed at improving the safety in territorial seas. We wil focus on an area where can be leveraged to address unique local conditions. In the future, Finnish waters wil witness the presence of both autonomous vessels and more accident prone recreational boats, which poses additional chal enges for seafarers in terms of situational awareness. To address these chal enges effectively, it is necessary to enhance data accessibility for al seafarers, enabling them to develop a better understanding of each other's intentions and safe navigation routes. Given the diverse capabilities of different user groups in terms of maneuvering and communication at sea, the standard fairway, 1 https://eda.europa.eu/docs/documents/SARUMS-Flyer-2012.pdf 2 Electronic Chart Display and Information System, definitions by International Hydrgraphic organization IHO S-57 (content and display) and IHO S-52 (transfer) and complying with IEC 61174 test procedures. 790 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY weather, and traffic information often fails to meet the specific needs of each group. Therefore, we propose utilizing common, multisided platforms and developing applications tailored to the requirements of different user groups, thus taking advantage of suitable applications (Heikkilä et al., 2023; Parker and Van Alstyne, 2004). The responsibility of authorities is limited to traffic on official fairways, but there is a clear safety discrepancy when it comes to providing guidance and assistance for recreational traffic outside waterways. Therefore, it is crucial to careful y examine the technical means by which safe routes in the archipelago can be identified. This may involve tracing traffic patterns, digitalizing currently fragmented oral information on traffic conditions, safety alerts, and environmental hazards, and utilizing VHF-radio for interaction with authorities and fel ow seafarers, both through broadcast and point-to-point communication. Digitalization has made these advancements possible, and the most crucial standardization efforts for improving safety have shifted from IMO to IALA3 and ITU4. This transition is driven by the increasing autonomy of merchant traffic and the telecommunications limitations faced by recreational vessels. These organizations are focused on designing more efficient digital communication systems (e.g. VDES R-mode5) and integrated navigational aids. With the assistance of machine learning and artificial intelligence models, it is anticipated that these functionalities (e.g., as described by Pedersen et al., 2020) will expand to type 1 applications and remotely assisted services (type 2), thereby benefiting recreational vessels as wel . However, it is evident that achieving this requires collaborative, co-created design efforts and the wil ingness of authorities to enhance, open, and maintain safety-critical data. Similarly, developing robust communication infrastructure at sea is of equal importance. None of these advancements can be effective if they are not integrated into the education of recreational seafarers, emphasizing enhanced digital 3 International Association of Marine Aids to Navigation and Lighthouse Authorities, “gathers together Marine Aids to Navigation authorities, manufacturers, consultants, and, scientific and training institutes from all parts of the world and offers them the opportunity to exchange and compare their experiences and achievements.” (IALA, 2023). 4 The International Telecommunication Union 5 https://www.iala-aism.org/product/g1158/ J. Heikkilä, M. Heikkilä: Improving Safety at Seas 791 seafaring skils and fostering a safety culture. Research is necessary to make informed decisions, identify effective transformational measures, and ensure the successful implementation of safer seafaring practices. References Aawa (2016). Remote and autonomous ships – the next steps. Position paper, Rols-Royce, London. 88 pages. Retrieved from: https://www.rol s-royce.com/~/media/Files/R/Rolls- Royce/documents/customers/marine/ship-intel/aawa-whitepaper-210616.pdf ABS (2020). American Bureau of Shipping. Chaal, M., Bahootoroody, A., Basnet, S., Banda, O. A. V., & Goerlandt, F. (2022). Towards system-theoretic risk assessment for future ships: A framework for selecting Risk Control Options. Ocean Engineering, 259, 111797. Heikkilä, J., Heikkilä, M., & März, G. (2023). Platforms for Smart Fairways – Enhancing Services for Autonomous Maritime Traffic and Other Emerging Uses of Territorial Sea, in the Proceedings of the 56th HICSS, 2023-01-03. 1001- 1009. https://scholarspace.manoa.hawaii.edu/items/61d49342-893b-4e4a-ba40-bc10236e4a4c IALA (2023). https://www.iala-aism.org/about-iala/ IMO (2021). Outcome of the regulatory scoping exercise for the use of maritime autonomous surface ships (MASS), IMO, 3.6.2021. available at https://wwwcdn.imo.org/localresources/en/MediaCentre/PressBriefings/Documents/MSC .1-Circ.1638 - Outcome Of The Regulatory Scoping ExerciseFor The Use Of Maritime Autonomous Surface Ships. . (Secretariat).pdf Kepesedi, A. (2022) Maritime Autonomous Surface Ships: A critical ‘MASS’ for legislative review, UNCTAD Transport and Trade Facilitation Newsletter, article no 97, 13 Dec 2022, Retrieved from: https://unctad.org/news/transport-newsletter-article-no-97-fourth-quarter-2022 Miettinen K., Miettinen A., Hauta J., Töyrylä S. & Reinimäki S., (2021). Liikenteen automaation lainsäädäntö- ja avaintoimenpidesuunnitelma, (Action plan on legislation and key measures of transport automation ). Li kenne- ja viestintäministeriö Helsinki 2021, 243 pages (in Finnish). Parker G.G. and Van Alstyne M.W., (2005). Two-Sided Network Effects: A Theory of Information Product Design. Management Science 51(10),1494-1504. https://doi.org/10.1287/mnsc.1050.0400 Pedersen, T. A., Glomsrud, J. A., Ruud, E. L., Simonsen, A., Sandrib, J., & Eriksen, B. O. H. (2020). Towards simulation-based verification of autonomous navigation systems. Safety Science, 129, 104799. https://doi.org/10.1016/j.ssci.2020.104799 SOLAS, (1974). International Convention for the Safety of Life at Sea. https://www.imo.org/en/About/Conventions/Pages/International-Convention-for-the- Safety-of-Life-at-Sea-(SOLAS),-1974.aspx Statistics Finland (2022). Vesiliikenneonnettomuustilasto - Statistik över sjöolyckor Ennakkotieto tammi- joulukuu 2021. (Statistics on accidents on seas and lakes, pre-information for Jan-Dec 2021). In Finnish and Swedish; available at https://www.traficom.fi/sites/default/files/media/file/VESILONN-2021-1-12- ennakko.pdf. STCW Convention (1978). International Convention on Standards of Training, Certification and Watchkeeping for Seafarers. Available at https://www.imo.org/en/OurWork/HumanElement/Pages/STCW-Convention.aspx 792 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY RESEARCH IN PROGRESS HUMAN-AI COLLABORATION IN THE CONTEMPORARY WORKPLACE: THE JOB DEMANDS-RESOURCES MODEL PERSPECTIVE TAMILLA TRIANTORO,1 ALEKSANDRA PRZEGALINSKA,2 ANNA KOVBASIUK2 1 Quinnipiac University, Hamden, Connecticut, United States of America tamilla.triantoro@qu.edu 2 Kozminski University, Warsaw, Poland aprzegalinska@kozminski.edu.pl, akovbasiuk@kozminski.edu.pl Evaluating the impact of AI in the context of the Job Demands- Resources (JD-R) model is useful for understanding the factors that influence employee wel -being in the modern workplace. The JD-R model highlights the importance of job demands and resources in predicting work-related outcomes, such as job satisfaction and burnout. However, with the increasing adoption of AI in the workplace, it is unclear how these factors interact Keywords: with AI use. Evaluating the impact of AI on job demands and generative AI, resources in the JD-R context can shed light on how this ChatGPT, technology can affect employee wel -being. Additionally, by JD-R model, understanding the factors that influence employee wel -being in future of the presence of AI, organizations can create effective strategies work, to enhance human-AI collaboration, leading to increased large language productivity, efficiency, and ultimately, a healthier and more model satisfied workforce. (LLM) DOI https://doi.org/10.18690/um.fov.4.2023.51 ISBN 978-961-286-751-5 794 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The contemporary world presents unique chal enges to individuals in the workforce, with increasing demands leading to stress, burnout, and reduced productivity. The recent COVID-19 pandemic has also brought about significant changes in the way we work, with many employees shifting to remote work or facing new demands related to health and safety. In this context, the Job Demands-Resources (JD-R) model has become increasingly relevant as a framework for understanding how job demands and resources impact employee wel -being and performance. By examining the factors that affect employee wel -being, the JD-R model helps employers identify ways to promote employee health and productivity, which is especial y important in light of the many chal enges facing workers today. One potential resource for addressing these chal enges is artificial intel igence (AI). AI has served as an effective assistant to human workers by carrying out mundane and time-intensive tasks, while concurrently offering insightful feedback and suggestions. By offloading these tasks, human workers can focus on more complex and rewarding aspects of their jobs, leading to greater job satisfaction and potential y reducing the risk of burnout. Additionally, AI can provide valuable insights and recommendations that can help human workers make more informed decisions and improve their overal job performance. Our study seeks to explore the potential of AI as an effective resource to human workers. By investigating how AI can be integrated into the workplace, we hope to identify ways in which AI can enhance collaboration between humans and machines. Specifical y, we are examining whether AI can perform routine and time-consuming tasks while also providing valuable insights and recommendations. We believe that the findings of our study wil contribute to enhancing collaboration between humans and AI in professional settings, leading to greater efficacy and output. Ultimately, this could benefit both individual workers and the organizations they work for by promoting greater productivity, job satisfaction, and overall well-being. T. Triantoro, A. Przegalinska, A. Kovbasiuk: Human-AI Col aboration in the Contemporary Workplace: The Job Demands-Resources Model Perspective 795 2 Job Demands-Resources (JD-R) Model The Job Demands-Resources (JD-R) model is a theoretical framework developed to explain how certain job characteristics influence employee wel -being and work outcomes. The model suggests that there are two broad categories of job characteristics: job demands and job resources (Demerouti et al. 2001). Job demands refer to the physical, psychological, social, or organizational aspects of work that require sustained physical or mental effort, and that can be a source of stress or strain for employees. Examples of job demands include workload, time pressure, emotional demands, and role conflict. Job resources, on the other hand, are the physical, psychological, social, or organizational aspects of work that help employees to achieve work goals, reduce job demands, and promote wel -being. Examples of job resources include social support, autonomy, feedback, and opportunities for development. The JD-R model posits that the interaction between job demands and job resources influences employee wel -being and work outcomes. Specifical y, high job demands combined with low job resources can lead to negative outcomes such as burnout, whereas high job resources can help to buffer the negative effects of high job demands. Conversely, low job demands combined with high job resources can lead to positive outcomes such as motivation and engagement, whereas low job resources can limit opportunities for growth and development (Bakker and Demerouti 2007). The Job Demands-Resources (JD-R) Model has been widely used to investigate how different job characteristics affect employee wel -being and job performance (Bauer et al. 2014). We are interested in applying this model to research on generative AI to understand how this technology can support human workers in the workplace. Generative AI is a relatively new tool that can provide employees with a new resource to enhance their work. This technology can assist employees by summarizing large amounts of data, generating reports and proposals, and writing emails. By using generative AI to automate these tasks, employees can free up time to focus on more complex and creative aspects of their work. In the next section we will describe generative AI and its use in virtual assistant applications. 796 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3 Generative AI and Virtual Assistants Generative AI is a subset of artificial intel igence that has been widely used to create new content through machine learning techniques. This technology is increasingly being used in the workplace, particularly in areas such as marketing, design, and content creation. However, little is known about how generative AI affects the wel - being and job performance of human workers, which is where the JD-R model can play an important role. The JD-R model proposes that job demands, such as workload and time pressure, can lead to job strain and burnout, while job resources, such as autonomy and social support, can enhance job satisfaction and motivation. In the context of generative AI, the technology can be seen as a job resource that al eviates problems and reduces job demands, thereby improving employee wel - being and job performance. The virtual assistant is a notable expression of generative AI in the workplace that utilizes natural language processing and machine learning technology (Triantoro and Przegalinska 2022). Designed to automate tasks and enhance employee productivity, virtual assistants can be activated through either text or voice commands. According to Gartner's projection, by 2025, half of all knowledge workers will utilize a virtual assistant on a daily basis (Bradley 2020). As an example, a virtual assistant powered by generative AI could help human workers in a marketing team design and launch an advertising campaign for a new product. The assistant could generate ideas and content for the campaign, reducing the workload and time pressure on human workers. In this way, generative AI can act as a job resource that supports human workers, enhancing their job satisfaction and motivation. To investigate the role of generative AI as a job resource, the JD-R model can be applied to analyze the extent to which this technology can reduce job demands and enhance job resources. The model can also be used to investigate the relationships between generative AI, employee wel -being, and job performance. By doing so, we can gain insights into how generative AI can be leveraged to improve the performance and wel -being of human workers in the workplace. T. Triantoro, A. Przegalinska, A. Kovbasiuk: Human-AI Col aboration in the Contemporary Workplace: The Job Demands-Resources Model Perspective 797 4 Methodology Our research aims to investigate the potential of AI in enhancing human performance in a business context. The study involves participants who wil use an AI chatbot, which is being trained and refined with the GPT-3.5 and GPT-4 language models by OpenAI, to create a marketing campaign for a fictional product. Marketing has been selected as the domain for our study because it provides a diverse range of tasks that vary in their level of difficulty and creativity. This allows us to explore the potential benefits of human-AI collaboration across different types of job demands. Furthermore, marketing campaigns often involve language-based communication, making them an ideal domain to test the capabilities of the AI application being used in our study, which is based on the large language models by Open AI. To investigate the effects of AI resources in the context of the JD-R model, we wil conduct an experimental study. The participants wil be randomly assigned in one of two conditions. In the first condition, they wil work on the task alone, using traditional resources available to marketers, such as search engines and databases. In the second condition, participants wil collaborate with an AI chatbot trained in a marketing domain to work together on the same task. By comparing the results of the two conditions, the study aims to identify differences or advantages that may arise from using AI as a collaboration resource in a workplace. By examining the effect of AI assistance on tasks with varying degrees of difficulty and creativity, we can determine whether AI is more effective in aiding humans with certain types of job demands. This information can be valuable in designing future AI systems that can better support and enhance human performance and wel -being in the workplace. The study is currently in the pilot phase, and we expect to present the preliminary findings at an upcoming conference. 798 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 5 Conclusion The Job Demand-Resource (JD-R) model is a useful framework to understand the implications of Artificial Intelligence (AI) in the modern workplace. AI has the potential to revolutionize various aspects of organizational tasks and enable humans and machines to work together in new ways, changing how work is done now and in the future. However, the impact of AI-driven changes on organizations and employees is not thoroughly examined. In this study, we attempt to evaluate the fit between resources and work demands and explore how generative AI can provide support in tasks of various complexity. To our knowledge, this research is the first to apply the JD-R model to the context of generative AI and measure the impact of generative AI on organizational and individual outcomes. By doing so, the study will provide valuable insights for organizations seeking to enhance their existing processes and individuals looking for technology aids to assist them in various tasks. It is important to note that while generative AI models can produce useful results, they also have limitations. Therefore, the study will closely monitor and address ethical concerns that may arise during the experiments. By evaluating the performance and ethical implications of generative AI models, this research wil contribute to the development of responsible AI practices in the workplace. References Bakker, A. B., & Demerouti, E. (2007). The job demands‐resources model: State of the art. Journal of managerial psychology, 22(3), 309-328. Bauer, G. F., Hämmig, O., Schaufeli, W. B., & Taris, T. W. (2014). A critical review of the job demands-resources model: Implications for improving work and health. Bridging occupational, organizational and public health: A transdisciplinary approach, 43-68. Bradley, A.J. (2020, August). Brace Yourself for an Explosion of Virtual Assistants. Gartner. https://blogs.gartner.com Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2001). The job demands-resources model of burnout. Journal of Applied psychology, 86(3), 499. Triantoro, T. & Przegalińska, A. (2022). Why Should I Like AI? The Effect of Personality Traits on the Acceptance of AI-Driven Virtual Assistants in the Workplace. Pacific Asia Conference on Information Systems (PACIS). Taipei-Sydney. RESEARCH IN PROGRESS TOWARDS AN ADAPTIVE IMPLEMENTATION TOOL FOR DEVICES IN COMPLEX HOSPITAL DEPARTMENTS NAVIN SEWBERATH MISSER,1 JORIS JASPERS,2 JOHAN VERSENDAAL3 1 HU University of Applied Sciences, Institute for Design & Engineering, Utrecht, Netherland Navin.SewberathMisser@hu.nl 2 University Medical Center Utrecht, Medical Technology and Clinical Physics, Utrecht, Netherland J.Jaspers@umcutrecht.nl 3 HU University of Applied Sciences, Research Group Digital Ethics, Utrecht, Netherland Johan.Versendaal@hu.nl Implementation of technological digital devices in existing complex hospital environments remains a chal enge. We constructed a framework for the implementation of technological devices in operating rooms. In this research, we address users' needs for an adaptive (digital) implementation tool or app. We use a requirements engineering method to identify Keywords: stakeholders and to identify steps to define requirements for this implementation framework, digital tool. We wil construct personas to identify requirements technological and based on these findings a market search wil follow to decide, devices, either to buy and configure an existing tool or to develop a new digitization, hospitals, tool that facilitates the implementation of devices in complex operating hospital departments. room DOI https://doi.org/10.18690/um.fov.4.2023.52 ISBN 978-961-286-751-5 800 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Digitization of healthcare occurs in many hospitals and affects many stakeholders in hospitals. Many technological devices are used in hospitals to facilitate health care processes and these devices are connected with various information systems. For example, when vital signs are recorded these can be transferred to a patient electronic health care record (EHR) automatical y; many other devices used in operating rooms are connected to EHR's, and thus increase digitization in operating rooms in hospitals. Not only because of this dependence on and connectivity to information systems, but also due to many other situational and contextual variables, implementation of new devices into an existing and complex environment such as an operating room in hospitals, is a chal enge. Success rates of implementations of information systems and other technological devices vary and numerous implementations fail (Damschroder et al., 2022; Jacobs et al., 2015; Rafferty et al., 2013). Less successful implementations of systems and devices in healthcare are reported less often, and many learnings for successful implementations can be derived from less successful implementation projects (Ebad, 2020). Successful implementations require well-prepared implementation activities in preparation to the introduction of the device and to influence adoption for use of the new device (Fennelly et al., 2020). To facilitate the implementation of new technological devices in OR's, we conducted several studies to construct a framework for implementation of new technological devices (Sewberath Misser, 2023). Based on the findings and the evaluation of this research, potential users of this framework identified a need for a (digitized) tool or app that enables users to be equipped with situational relevant implementation activities In this study, we adress this need and we composed the following research question: How can requirements be identified for a user friendly and adaptive digital tool or app for implementation activities for new devices and software in complex hospital environments such as operating rooms? In the next section, we in short describe the framework for implementation for technological (digital) devices in operating rooms (Sewberath Misser, 2023). In section three, we describe an approach to develop an adaptive tool for the implementation of activities. N. Sewberath Misser, J. Jaspers, J. Versendaal: Towards an Adaptive Implementation Tool for Devices in Complex Hospital Departments 801 2 Background 2.1 Introducing technological and digital devices Implementation of new technological devices in hospitals appears to be a chal enge, as many stakeholders and departments are involved when introducing these new tools. Edmondson et al. refer to a successful implementation of a new tool when the tool is used in day-to-day activities of users (Edmondson et al., 2001). This implementation remains a chal enge and wil need to be addressed careful y. To address this chal enge, we used characteristics from design science research (Peffers et al., 2007). In previous research, we identified in three phases to develop and evaluate an implementation framework, starting with identifying a problem definition. Based on these findings, we developed an artifact in the second phase of research. In the third and last phase of our research we evaluated this artifact. We constructed a problem definition by providing an overview of the research context and by identifying the problem. Schovil e et al. state that there is no overarching implementation theory but only a number of models and processes to facilitate implementation (Schoville & Titler, 2015). For the implementation of electronic health care records (EHR) various studies have been conducted to identify factors or barriers for the implementation and adoption of EHR (Fennelly et al., 2020). Based on existing literature an implementation framework or process model for the implementation of new technological devices in an existing operating room environment appeared not available (Damschroder et al., 2009; Moullin et al., 2015; Nilsen, 2015). Based on these findings, an artefact to implement new technological devices in OR’s was needed. With that knowledge, we conducted several studies to develop the artifact: a framework for implementation. In this framework we distinguished determinants for implementation, and we identified activities and instructions for implementations. To construct this framework for implementation we conducted an explorative study with scrub nurses and circulating nurses in Operating Rooms to identify relevant factors for implementation (Sewberath Misser, Jaspers, et al., 2018). In a systematic literature review we focused on relevant factors and activities for implementations based on existing studies and theories (Sewberath Misser, Zaane, et al., 2018). Based on the findings of these two studies, we composed a baseline framework for implementation, consisting of implementation factors, implementation activities and implementation instructions (Sewberath Misser et al., 2020). We conducted two studies to validate this framework. In the first validation 802 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY study we organized three focus group sessions with implementation experts from varying backgrounds and hospital departments. These experts were surgeons, anesthesiologists, logistics employees, scrub nurses, OR-management, and a methodologist. Based on these findings suggestions for revisions were proposed and analyzed (Sewberath Misser, Jaspers, Van Zaane, et al., 2021). In the second validation study, our base line implementation framework was used to introduce an exoskeleton in the OR (Sewberath Misser, Jaspers, van Zaane, et al., 2021). An exoskeleton is a wearable, mechanical external structure that enhances or supports the power of someone when performing specific activities (De Looze et al., 2016). Based on the data of this case study and the outcomes of the validation study with experts, revisions for the baseline protocol were identified and processed. The final and revised framework for implementation consists of four implementation factors, with related activities and instructions. In the next subsections these factors are explained. • Set up a project plan. The first factor for implementation focuses on the composition of a project plan. Activities involve identifying topics, goals, performance indicators, risks and relevant implementation activities based on the complexity of the device. • Organizational preparation The second implementation factor involves preparation of the organization. Implementation activities include activities to assemble a multi-disciplinary implementation team and to prepare the existing stakeholders of the organization. Processes and activities that need adjustment due to the introduction of a new technological device are redesigned and shared with relevant users and departments. Simulations using the new device are prepared and executed and communication plans are identified and deployed. • Technical preparation The third factor for implementation involves the technological preparation of the environment and the configurational preparations of the device. These configurational preparations include positioning of the device as well as assembly N. Sewberath Misser, J. Jaspers, J. Versendaal: Towards an Adaptive Implementation Tool for Devices in Complex Hospital Departments 803 and disassembly procedures. Technical preparation also entails composing and deploying data management plans, maintenance plans, and updating (safety) regulations. • Training and evaluation Training and evaluation activities include composing and deploying a training and evaluation plan. Training involves technical and non-technical skills and technical skills also include troubleshoots and interpretation of necessary data and alarms. This training and evaluation plan also covers assessment of skills, as well as evaluation of the device and evaluation of the implementation project. Our implementation framework is included in appendix 1. 3 Towards an adaptive tool for implementation activities In the previous chapter, the implementation framework and its activities are described. The purpose of this research is to identify requirements to develop an adaptive tool or app for implementation activities, depending on the user requirements and the device. There are different ways to develop or to configure a tool, such as the waterfal method or more agile methods to develop software. A waterfal method of software development general y consists of sequential phases: requirement analysis, system design, implementation, system testing, system deployment and system maintenance (Zayat & Senvar, 2020). The advantages of this method are that this method is described as structured, simple to understand and to implement. However, this method is described as less flexible and difficult when changes in requirements need to be considered and processed (Kramer, 2018). Similarly, the Software Development Life Cycle (SDLC) consists of the following phases: surveying and assessing the feasibility of a software development project, analyze existing information systems, determine requirements for the system, selecting the best solution, determine hardware and software, designing the system, building the system, implementing the system, and maintain the new system (Adi Guna Permana, 2015; Zayat & Senvar, 2020). A more flexible method for developing a software tool is for example the agile development method SCRUM. In agile development methods a development team is selected with a common goal to develop a holistic product. Different team roles are identified. One team member (product owner) defines business goals and requirements for the tool and this team 804 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY member initiates the development process by prioritizing activities. These activities are scheduled and executed in development cycles (sprints) (Adi Guna Permana, 2015; Zayat & Senvar, 2020). To develop a tool for our framework for implementation, we need to consider these development frameworks. As the requirements for a tool are not clear yet, one of the steps in the development process is to identify requirements for the tool. Panday et al. introduced a requirements engineering process model, in which they consider four steps to identify and select requirements for a software tool (Pandey et al., 2010). In the first step, requirements elicitation and development, stakeholders are identified to provide requirements. In the second step, requirements are documented and in the third step requirements are verified and validated. In the last step, they describe how the agreed requirements can be tracked and controlled during the development process of a software tool (Pandey et al., 2010). When focusing on the identification of requirements, Mayas et al. introduces the use of personas to identify requirements for users (Mayas et al., 2016). They describe personas as archetypes: distinguishable characters based on behavior and dispositions of real people. To identify requirements for an adaptive tool or app for implementation activities we wil use personas, as we expect the tool to be used by different stakeholders in OR’s. When executing the first step of the requirements engineering process we wil identify personas based on potential users with stakeholders (Mayas et al., 2016; Pandey et al., 2010). We wil then use these constructed and validated personas to define requirements for an adaptive tool or app for implementation activities. These requirements wil be documented and after documentation these requirements will be verified and validated with stakeholders. Following the SDLC method, the best solution needs to be selected based on validated requirements. In this phase, it should become clear whether a new tool needs to be developed or whether an existing tool used for other purposes needs to be considered. If a suitable tool is not available on the market, tool development can be prepared. Agile methods of tool development are preferred, as with our current knowledge we consider the tool to be an incremental tool which would enable rapid development cycles. To develop this adaptive tool, we would like to include users and developers to increase development success and adoption (Abrahamsson et al., 2017). N. Sewberath Misser, J. Jaspers, J. Versendaal: Towards an Adaptive Implementation Tool for Devices in Complex Hospital Departments 805 4 Conclusion, limitations and further research Implementation of technological digital devices in complex environments such as operating rooms remain a chal enge. To address this chal enge, we developed and evaluated a framework for implementation consisting of implementation factors, related implementation activities and implementation instructions. These factors are: setting up a project plan, organizational preparation, technical preparation and training and evaluation. This research adresses the need of potential users of this framework. These users identified a need for an adaptive tool or app enabling users to select and prioritize implementation activities. To respond to the research question we wil use process steps of requirements engineering to identify requirements for this tool. Based on the findings up to now and our current knowledge, we need to select stakeholders and develop personas. Based on these personas we wil identify requirements. A market search on available tools needs to be executed to decide whether to buy and adjust an existing vendor tool or to develop a new tool. In case of a development of a new tool, we prefer an agile method for development. Limitations and further research In this research in progress we initiate the process to fol ow up on users needs to develop or acquire a tool or app that facilitates implementation of new technological equipment. Up to now did a provisonal and limited search, on development methodologies for software to develop this tool. We use the body of knowledge of requirements engineering to identify requirements with use of personas. After these requirements have been specified, further research needs to be conducted on the preferred agile development method. References Abrahamsson, P., Salo, O., Ronkainen, J., & Warsta, J. (2017). Agile Software Development Methods: Review and Analysis. Chemistry (Weinheim an Der Bergstrasse, Germany), 19(21), 6641–6649. https://doi.org/10.1002/chem.201203966 Adi Guna Permana, P. (2015). Scrum Method Implementation in a Software Development Project Management. International Journal of Advanced Computer Science and Applications, 6(9), 198–204. https://doi.org/10.14569/IJACSA.2015.060927 Damschroder, L. J., Aron, D. C., Keith, R. E., Kirsh, S. R., Alexander, J. A., & Lowery, J. C. (2009). Fostering implementation of health services research findings into practice: A consolidated 806 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY framework for advancing implementation science. Implementation Science, 4(1), 1–15. https://doi.org/10.1186/1748-5908-4-50 Damschroder, L. J., Reardon, C. M., Widerquist, M. A. O., & Lowery, J. (2022). The updated Consolidated Framework for Implementation Research based on user feedback. Implementation Science, 17(1), 1–16. https://doi.org/10.1186/s13012-022-01245-0 De Looze, M. P., Bosch, T., Krause, F., Stadler, K. S., & O’Sul ivan, L. W. (2016). Exoskeletons for industrial application and their potential effects on physical work load. Ergonomics, 59(5), 671–681. https://doi.org/10.1080/00140139.2015.1081988 Ebad, S. A. (2020). Healthcare software design and implementation—A project failure case. Software - Practice and Experience, 50(7), 1258–1276. https://doi.org/10.1002/spe.2807 Edmondson, A. C., Bohmer, R. M., & Pisano, G. P. (2001). Disrupted Routines: Team Learning and New Technology Implementation in Hospitals. Administrative Science Quarterly, 46(4), 685. https://doi.org/10.2307/3094828 Fennel y, O., Cunningham, C., Grogan, L., Cronin, H., O’Shea, C., Roche, M., Lawlor, F., & O’Hare, N. (2020). Successfully implementing a national electronic health record: a rapid umbrella review. International Journal of Medical Informatics, 144(September), 104281. https://doi.org/10.1016/j.ijmedinf.2020.104281 Jacobs, S. R., Weiner, B. J., Reeve, B. B., Hofmann, D. A., Christian, M., & Weinberger, M. (2015). Determining the predictors of innovation implementation in healthcare: A quantitative analysis of implementation effectiveness. BMC Health Services Research, 15(1), 1–13. https://doi.org/10.1186/s12913-014-0657-3 Kramer, M. (2018). Best Practices in Systems Development Lifecycle: An Analyses Based on the Waterfal Model. Review of Business & Finance Studies, 9(1), 77–84. Mayas, C., Hörold, S., & Krömker, H. (2016). Personas for Requirements Engineering. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intel igence and Lecture Notes in Bioinformatics): Vol. 9312 LNCS (pp. 34–46). https://doi.org/10.1007/9783-319-45916-5_3 Moul in, J. C., Sabater-Hernández, D., Fernandez-Llimos, F., & Benrimoj, S. I. (2015). A systematic review of implementation frameworks of innovations in healthcare and resulting generic implementation framework. Health Research Policy and Systems, 13(1), 1–11. https://doi.org/10.1186/s12961-015-0005-z Nilsen, P. (2015). Making sense of implementation theories, models and frameworks. Implementation Science, 10(1), 1–13. https://doi.org/10.1186/s13012-015-0242-0 Pandey, D., Suman, U., & Ramani, A. K. (2010). An Effective Requirement Engineering Process Model for Software Development and Requirements Management. 2010 International Conference on Advances in Recent Technologies in Communication and Computing, 287–291. https://doi.org/10.1109/ARTCom.2010.24 Peffers, K., Tuunanen, T., Rothenberger, M. A., & Chatterjee, S. (2007). A design science research methodology for information systems research. Journal of Management Information Systems, 24(3), 45–77. https://doi.org/10.2753/MIS0742-1222240302 Rafferty, A. E., Jimmieson, N. L., & Armenakis, A. A. (2013). Change Readiness: A Multilevel Review. Journal of Management, 39(1), 110–135. https://doi.org/10.1177/0149206312457417 Schoville, R. R., & Titler, M. G. (2015). Guiding Healthcare Technology Implementation. CIN: Computers, Informatics, Nursing, 33(3), 99–107. https://doi.org/10.1097/CIN.0000000000000130 Sewberath Misser, N. (2023). Introducing technological innovations in Operation Rooms in hospitals : an implementation framework for technological devices LK - https://hu.on.worldcat.org/oclc/1368026515. Open Universiteit. Sewberath Misser, N., Jaspers, J., van Zaane, B., Gooszen, H., & Versendaal, J. (2018). Transforming operating rooms: factors for successful implementations of new medical equipment. Digital Transformation – Meeting the Chal enges, June, 279–289. https://doi.org/10.18690/978-961-286-170-4.18 N. Sewberath Misser, J. Jaspers, J. Versendaal: Towards an Adaptive Implementation Tool for Devices in Complex Hospital Departments 807 Sewberath Misser, N., Jaspers, J., van Zaane, B., Gooszen, H., & Versendaal, J. (2021). Evaluating an Implementation Protocol for Digitization and Devices in Operating Rooms: a Case Study. 34th Bled EConference Digital Support from Crisis to Progressive Change: Conference Proceedings, 351–364. https://doi.org/10.18690/978-961-286-485-9.26 Sewberath Misser, N., Jaspers, J., Van Zaane, B., Gooszen, H., & Versendaal, J. (2021). Evaluation of an implementation protocol for digitization and devices in Operating Rooms. AMCIS 2021 Proceedings, 0–10. https://aisel.aisnet.org/amcis2021/healthcare_it/sig_health/2 Sewberath Misser, N., Jaspers, J., Zaane, B. Van, Gooszen, H., & Versendaal, J. (2020). A protocol for the implementation of new technology in a highly complex hospital environment: the operating room. International Journal of Networking and Virtual Organisations, 22(2), 199. https://doi.org/10.1504/IJNVO.2020.105543 Sewberath Misser, N., Zaane, B. Van, Jaspers, J. E. N., Gooszen, H., & Versendaal, J. (2018). Implementing Medical Technological Equipment in the OR : Factors for Successful Implementations. Journal of Healthcare Engineering, 2018. https://doi.org/https://doi.org/10.1155/2018/8502187 Zayat, W., & Senvar, O. (2020). Framework Study for Agile Software Development Via Scrum and Kanban. International Journal of Innovation and Technology Management, 17(04). https://doi.org/10.1142/S0219877020300025 808 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Appendix 1: implementation framework for technological devices in operating rooms A revised framework for implementation for technological devices was constructed and is presented in the next table (Sewberath Misser, 2023). We intend to use the contents of this framework as first data set for an adaptive digital tool or app to help users to implement new digital and technological devices. Table 1: Revised implementation framework (Sewberath Misser, 2023) Id Implementation Activities Instructions for implementation 1 Set up a project plan Identify strategic and tactical Operationalize overall strategic and tactical goals 1.1 topics for the implementation stage. Identify performance indicators to define the performance of the implementation stage and 1.2 Identify performance define how these variables are measured and analyzed. Performance metrics for success could be efficiency, finance, and ergonomics. Identify (groups of) stakeholders, which are responsible, accountable, consulted and 1.3 Identify stakeholders informed such as sponsors, key-representatives, staff, teams. Identify a project manager for implementation Perform a risk assessment to identify risks and Identify risks related to 1.4 identify unintended outcomes as new technology implementation may have unforeseen consequences. Identify relevant activities for implementation, Identify activities for 1.5 based on listed activities. Generate a planning or implementation timeline for execution of these activities. 2 Organizational preparation Assemble a team which includes various members of involved departments and stakeholders such as scrub nurses, circulating Assemble a multidisciplinary 2.1 nurses, anesthesiologists, perioperative implementation team technicians, surgeons, administrators, IT specialists, and schedulers. Consider assigning an extra team member during implementation to N. Sewberath Misser, J. Jaspers, J. Versendaal: Towards an Adaptive Implementation Tool for Devices in Complex Hospital Departments 809 Id Implementation Activities Instructions for implementation increase familiarity with procedures, e.g. setup procedures. Team familiarity and stability impacts teamwork, communication, and satisfaction during 2.2 Foster team familiarity implementation. Assign a dedicated implementation team. Involve and inform this team well. Introducing new (medical) equipment influences existing activities and work processes. Identify Identify affected activities 2.3 these and analyze how these processes are and/or processes affected and which identified stakeholders are involved. Checklists improve safety and reliability prior to, Update checklists and/or and during surgical procedures. Update 2.4 protocols operating procedures or protocols. If necessary, update existing check lists. Simulate with stakeholders (and departments) how processes and work activities are executed 2.5 Perform simulations prior to introducing (medical) equipment. Practice with a new tool or new (prototype) equipment on trial basis. Participation of employees when introducing new (medical) equipment increases employees’ Identify and deploy activities engagement in the OR. Deploy activities to 2.6 to increase employees’ engage employees in the OR, e.g., involvement engagement of work councils, create a communications council. Embedding information systems or new (medical) equipment in day-to-day activities as an accepted routine is a chal enge. Identify and Identify and deploy activities deploy activities to increase adoption with 2.7 to increase employees’ stakeholders such as demonstrating relative adoption advantages, possibilities to observe and experiment, demonstrate benefits, use training and assign key users or champions. Communication with stakeholders increases Communicate with 2.8 engagement and involvement of stakeholders. stakeholders Set up a communications plan, consisting of 810 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Id Implementation Activities Instructions for implementation communication activities over time. Involved stakeholders should be aware what their role is relating to the new (medical) equipment. For example, nursing personnel should be familiar with the instrumentation needs and they should be proficient in properly connecting, calibrating, set up and use (medical) equipment. Communication activities can be: (pre-operative) group briefings, interviewing stakeholders, using videos and newsletters, developing patient centered information. 3 Technological preparation Prepare technical facilities related to the use of 3.1 Prepare equipment the information system or device in the OR e.g., power and plugs (if needed) Introducing a new tool or system may affect ergonomic aspects of staff in the OR. Consider these aspects in an early stage of the project, prior 3.2 Consider ergonomic aspects to implementations. Simulations may lead to ergonomic changes and positioning of tools in the OR. Introducing new equipment requires integration Prepare interfaces with other in and with other devices in the OR. Consider the 3.3 information systems connectivity to the clinical networks to ensure safety and reliability. The introduction of new equipment affects current workflows and processes. These Integrate device within 3.4 workflows need to updated, and existing existing environment standard operating procedures need to be updated accordingly. When introducing equipment data can be generated and/or stored, e.g. when introducing a 3.5 Manage generated data new information system. Consider data processing and security aspects and develop or update procedures. New equipment in use should be maintained 3.6 Set up maintenance plan periodical y and in case of problems, support N. Sewberath Misser, J. Jaspers, J. Versendaal: Towards an Adaptive Implementation Tool for Devices in Complex Hospital Departments 811 Id Implementation Activities Instructions for implementation should be available. To address and facilitate this, a maintenance plan should be set up. Provide instructions how to maintain (clean) tools/equipment such as screens in the OR and confirm who is responsible for this activity. The introduction of new equipment may affect work activities of personnel. Assess the safety 3.7 Update safety (regulations) procedures and if needed, update these procedures accordingly. 4 Training and evaluation (Recurrent) training is crucial for correct and safe use of the system or tool and affects adoption and success of an implementation. Training focuses on technical skil s and non-technical skills. Technical skills may include cognitive, integrative, and automatic skil s such as congress 4.1 Train involved staff visits, demonstrations, research results, online courses, knowledge training, expert opinions, and simulation trainings. Specific trainings on changing ICT and updated workflows and activities should be included as well. Non- technical skills may include decision making, communication and leadership skills. In case of electronic equipment, notifications may occur visibly on screens, lights, or audible 4.2 Interpret screens and (alarms). Involved personnel should be able to troubleshooting interpret these notifications and should be able to troubleshoot in case of occurring problems. To assess the readiness for use, a skil s assessment plan needs be developed and executed, tailored to the stakeholders. This plan may include supervision by co-workers. An 4.3 Assess Skil s assessment plan can be determined and executed by a manufacturer, the hospital or a department. (If applicable) assess whether skil s need to be recorded and tracked. Evaluate experiences and gather feedback 4.4 Evaluate experiences regarding the use of the new device, provide 812 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Id Implementation Activities Instructions for implementation input to optimize the device, the use of the device or the workflow. Evaluate the implementation process and relate 4.5 Evaluate implementation results to the performance indicators mentioned process in the implementation plan. RESEARCH IN PROGRESS COMPONENT MODELS FOR IOT SEARCH ENGINE VAIDAS GIEDRIMAS Panevėžio kolegija/State Higher Education Institution, Panevežys, Lithuania vaidas.giedrimas@panko.lt The more distributed system expands, the higher need for search engines of its elements we have. The Internet of Things (IoT) systems become very complex, and the number of devices is growing exponential y. The demand for the search engine of things now reached the demand for the search engine of web Keywords: pages as it was in the 1990s. We propose to use component- search based architecture for the search engine of things (IoTSE). As engine, internet of IoT systems are heterogeneous, and the interoperability of things, various component models is problematic, we must focus on component selected component models only. This paper surveys existing model, component-based component models and discuss their feasibility to be used in IoT software search engine. engineering DOI https://doi.org/10.18690/um.fov.4.2023.53 ISBN 978-961-286-751-5 814 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The Internet of Things (IoT) is an example of a complex computational system. It evolved from the idea in the 1990s to the infrastructure of everyday life nowadays. The number of IoT devices is growing exponential y, so the demand for the search engine of "things" is as big as for the search engine of webpages 3 decades ago (Gubbi et al. 2013; Fathy, Y., Barnaghi, P., Tafazolli, 2018; Tran et al. 2017; Tang et al. 2023). They are a pilot products already such as IoTCrawler and Shodan. However, the development of IoT search engines is not entirely the same as the development of "ordinary" search engines. It exists some chal enges and threats. When webpages are uniform to some extent, the IoT network is highly heterogeneous and dynamic. A typical web search engine's data source is the web pages, metadata, and content. IoT content is far more complex as they provide very dynamic data. Instead of static content they usual y provide the data streams. On the other hand, even the type of data can change. For example, a temperature sensor can provide floated point values (temperature), when it is working, Boolean value (false) when it is out of order, or the textual content (via its representative). Because of the nature of IoT, we cannot just reuse the principles of web page search engines. In software engineering, it is common practice to use reusable software components and/or software services when it is required to address dynamism and agility. Component-based software engineering (CBSE) uses modules of higher granularity of the components. This enables to reduction the time required for the development, testing of component-based software significantly. Components are "hot swap" modules in software similar to the storage disks, power supplies, and other modules in contemporary hardware (mainly servers). As al the communication between software components is organized via Interfaces, each software component can be easily and quickly replaced by another one, which implements the same interface. After this, there is no need to perform testing of the overal system. A component-based system can be considered highly reliable and easily maintainable (Vale et al., 2016). V. Giedrimas: Component Models for IoT Search Engine 815 The goal of this paper is to survey existing component models and discuss their feasibility to be used in IoT search engines (IoTSE). The rest of the paper is organized as follows: Section 2 presents the main architectural parts of search engines and the state-of-art taxonomy of IoT search engines. Section 3 surveys the most popular software component models (CORBA, EJB, .NET, OSGi) and discusses their place in the overal architecture of dynamic Internet of things search engine. Section 4 shortly introduces related work in this area. Finally, the conclusions are made, and open questions are discussed. 2 Search engine concept in the IoT context As presented by Brin and Page (1998), and by Bruce et al. (2010), main parts of the search engine are as fol ows: 1. A Crawler, or a bot, is a program that traverses the web to discover and download web pages. This is a critical component of a search engine because it enables the system to discover new web pages and update its index of existing ones. 2. Indexer : Once the crawler has downloaded a web page, the next step is to extract its content and store it in an index. An indexer is a program that processes the content of web pages and creates an index of words and their locations. This allows users to find relevant pages quickly by searching for specific words. 3. The Query processor is responsible for interpreting user queries and finding the most relevant web pages. This component usual y involves a combination of natural language processing, machine learning, and other algorithms to match user queries with relevant pages in the index. 4. A Ranking algorithm is used to determine the order in which web pages are presented to users. Different search engines use different algorithms, but the basic idea is to assign a score to each page based on factors such as relevance, authority, and popularity. 5. The User interface is the part of the search engine that users interact with. This includes the search box, search results page, and other elements that make it easy for users to find what they're looking for. 816 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Each IoT search engine must be able to perform at least two activities: discovery activity and search activity. During the discovery activity engine retrieves the list of IoT devices by scanning the local/global environment, IoT device-related websites, or cached databases. According to (Tran et al. 2019), more than 90% of the engines include discovery activities. After this list is made, the second activity – searching – is running. Its goal is to filter the former list by rejecting all the devices whose properties do not conform with the initial query. IoT search engines can be classified according to their meta-path. Meta-path reveals the source of the data for discovery and search and the target of these activities. Nguyen et al. (2019) distinguish 8 classes of IoTSE: 1. R → R class IoTSE 2. D + R → T → R class 3. D → D 4. R → T → R + D 5. D → T → R 6. F → F 7. R → T → F 8. S → S where R is Representative of IoT device (e.g. dedicated web page for its monitoring and control), D - Dynamic IoT content (e.g. not usual data stream), T - IoT device (aka Thing). 1st class (R-R) of IoT search engines is highly influenced by web search engines. The information about devices is taken from representatives only. No direct data gathering from IoT devices at the moment of search is performed. The second class (D+R-TR) of search engines gets data not only from representatives of things but (what is more specifical y) from the IoT devices directly as dynamic streams. The result of 2nd class IoTSE is the representatives. In other words, R-R class can be considered as a subset of D+R-TR, whit zero D component. 3rd class (D-D) goes beyond of D+R-TR class and gets the data from the very end sensors of Things, not just from the public observable states of things. 4th class (R-T-R+D), according to (Tran et al., 2019), is an extension of the R-R class. What is interesting, is that search engines from this class return not only links to IoT V. Giedrimas: Component Models for IoT Search Engine 817 representatives but the data stream from them as wel. 5th class (D-T-R) is a little bit similar to 2nd and 3rd classes. Typically they work using SPARQL queries over contextual information. As (Tran et al., 2019) observed in contrast with 2nd class IoT search engines, 5th class engines select things only by the data streams, not considering any other features (e.g. representatives). 6th class (F-F) is very interesting as the IoT devices are analyzed not by their representatives or data stream, but by their functionality. The functionality of things is exposed in a shared ontology. 7th class (RTF) of engines selects devices by its representatives. Its results include the functionality of selected things. As pointed out by Tran et al. (2019), this functionality is presented as RESTful Web services and is capable of self-describing with specifications written in the Web Application Description Language (WADL). In our context (CBSE) it is a very promising fact. And 8th class of IoTSE (SS) query on cached static information from IoT devices. In contrast to 3rd class, these IoTSE, do not acquire the newest data from IoT devices nor perceive dynamic data streams. However, then can be easily applied in the fields where the device state does not change often. The examples of IoT search engines are listed in Table 1. Table 1: IoT search engines Engine Class Source URL Shodan N/A https://www.shodan.io ThingFul DRTR? https://www.thingful.net IoTCrawler DD? https://iotcrawler.eu ForwarDS-IoT RR (Gomes et al., 2015) DiscoWoT RR (Mayer, Guinard, 2011) ThingSeek N/A (Shemshadi, Sheng, Qin, 2016) IoT-SVK DRTR (Jin et al., 2011) CASSARAM DD (Perera, 2013) Snoogle RT-R+D (Wang, Tan, Li, 2010) ASAWoO FF (Mrissa, Médini, https://liris.cnrs.fr/asawoo Jamont, 2014) N/A RTF (Kamilaris, Papakonstantinou, Pitsillides, 2014) Microsearch SS (Tan et al., 2009) 818 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 2: The coverage of IoTSE functions Function CORBA EJB OSGi .NET WS Crawler (Fathy, (Arel anes, Barnaghi, Indexer Lau, 2020) Tafazolli, 2018) QP (Chang, RA Yuan, Lo, 2000) (Chang, UI Yuan, Lo, 2000) 3 Component-based approach The component in our context is a reusable software module, having explicit interfaces, capable to discover IoT devices and/or getting their data. It is the subject of the third part of the composition and is strictly dependent on the middleware/framework. The idea to combine IoT with software components and services is not very new. The group of Zhiming Liu (2010) has been working on software-hardware components fusion for a long time before the rise of IoT. In another reference, Ruppen et al. (2015) present a system based on model-driven architecture (MDA) which helps to bind IoT devices and RESTful web services in a semi-automated way. The component in (Ruppen et al., 2015) work is the pair of IoT devices - RESTFull service. In both cases, the component is a "piece" of the system having its hardware and software parts. In CBSE each component must have at least one interface (or Provided interface) in which it is implemented, and (optional y) can have a Required interface, which declares what interfaces are needed from other components to work properly. Distributed components communicate using a Framework, or Middleware, which encompasses main data structures, and services (e.g. networking, DB) as well. V. Giedrimas: Component Models for IoT Search Engine 819 In this section, we continue the architectural approach described in (Gula, Flakova, 2017; Giedrimas, Backys, 2022; Saari, Nurminen, Rantanen, 2022) and will analyze the most popular component-based technologies: CORBA, EJB, OSGi, and .NET. 3.1 CORBA CORBA (Common Object Request Broker Architecture) is a middleware technology that provides a way for distributed applications to communicate with each other. CORBA was an industrial standard for the past two decades. They are stil several legacy systems using this technology. It can be used in the development of search engines to connect different components of the system and provide a standard way for them to communicate. The use of the CORBA component model for web search engines is described in (Chang, Yuan, Lo, 2000). CORBA is supported by OMG consortium, which gives grates opportunities to embed CORBA components into any existing system. However, CORBA is only one technology that does not separate the processes of component development and the use of the components. CORBA components usual y are not a subject of third-party composition. This does not let to use the "hot swap" advantage of CBSE. 820 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 3: The coverage of IoTSE functions Class CORBA EJB and OSGi .NET WS (Chang, RR Yuan, Lo, 2000) (Dang, (Chang, Pham, DRTR Yuan, Lo, Duong, 2000) 2018) (Arel anes, Lau, 2020; (Arel anes, Dang, DD Lau, 2020) Pham, Duong, 2018) (Chang, RTRD Yuan, Lo, 2000) (Arel anes, Lau, 2020; (Arel anes, Dang, DTR Lau, 2020) Pham, Duong, 2018) (Dang, Pham, FF Duong, 2018) (Arel anes, (Arel anes, RTF Lau, 2020) Lau, 2020) (Tan et al., SS 2009) V. Giedrimas: Component Models for IoT Search Engine 821 3.2 EJB and OSGi EJB (Enterprise Java Beans) is a server-side component architecture for Java that can be used for building distributed applications, including search engines. EJB provides a set of services such as transaction management, security, and persistence, which can be useful in building complex distributed applications. OSGi (Open Services Gateway initiative) is a modular framework for Java that provides a way to build complex applications using a set of reusable components. It can be used in the development of search engines to create a modular architecture that al ows components to be developed and tested independently. 3.3 .NET .NET is a software framework developed by Microsoft that can be used for building web applications, including search engines. It includes a variety of libraries and tools for developing web applications, such as ASP.NET for building web pages and MVC for creating web applications. 3.4 Web services Web services provide a standardized way for software components to communicate with each other over the internet, using technologies such as XML, SOAP, JSON. This makes it easier to integrate different components into a single system, which is important for building complex IoT search engines. Microservices are a more recent approach to software architecture, where a system is broken down into smal , independently deployable components that communicate with each other via lightweight protocols such as HTTP or messaging systems. This approach al ows for greater flexibility and scalability in building IoT search engines, as different microservices can be developed and deployed independently and can be easily replaced or scaled up or down based on demand. We refer to both technologies as to WS. In analyzed papers most of pilot projects on IoTSE was made using web services and microservices. 822 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3.6 Frameworks As mentioned earlier, in component-based software engineering we rely not only on the components but on the frameworks as wel . The most often used frameworks are: 1. Lightweight M2M (LwM2M): LwM2M is a protocol that is designed for use in IoT devices. It provides a standardized way for devices to communicate with servers, which can be useful for integrating data from IoT devices into a search engine. 2. CoAP protocol that is designed for use in IoT devices. It is a lightweight protocol that is wel -suited for constrained environments and can be used to integrate IoT device data into a search engine. 3. Apache Kafka is a distributed streaming platform that is commonly used in IoT contexts. It can be used to ingest data from IoT devices in real-time and can be integrated with search engines to provide real-time search capabilities. 4. Apache Spark is a popular distributed computing platform that can be used for processing large volumes of data in near-real-time. It can be used to process data from IoT devices and integrate it with search engines. 5. Elasticsearch for IoT : Elasticsearch is a popular search engine that has been adapted for use in IoT contexts. It can be used to store and search data from IoT devices, and provides a variety of features that are optimized for IoT use cases, such as geospatial search and time series data. 4 Related work The architectures of component-based IoTSE are described only in (Giedrimas, Backys, 2022; Gula, Flakova, 2017). Other papers cover this topic only partially. The idea to use the (reusable) workflow engines in the overal science gateway is architecturally like our approach. Glatard et al. (2017) refer to workflow engines as software components while presenting 6 software architectures: Tight integration, Service invocation, Task encapsulation, Pool model, Nested workflows with service invocation and Conversion, and Workflow conversion with service invocation. We assume that component-based IoTSE wil have a backbone (instead of the science V. Giedrimas: Component Models for IoT Search Engine 823 gateways), and smaller parts - software components implementing different IoT search and discovery algorithms. Other papers have weak (or negative) relations with software components; however, they are highly related to software architectures (Cambazoglu et al., 2007; Ozcan et al., 2012) and/or search engines (Kejriwal, 2021; Cambazoglu et al., 2007; Ozcan et al., 2012). 5 Conclusions and Future work After the analysis, we came to the fol owing conclusions: 1. Even though the market of IoTSEs is new and growing, enough different search engines is developed already. Fortunately, IoTSEs are not unique and can be classified into a limited number of classes. Each class of the IoTSE can deny an interface and each IoTSE from this class can be considered as a component implementing this interface. 2. The area of research (component models for IoTSE), is almost not covered by the scientific community nor the business applications. We found only a few sources. Surprisingly search engines provide some results about the topic, which led to broken links and non-existing pages. We assume that the papers were withdrawn because of the lack of maturity or experimental data. 3. The survey of the 5 most popular component-based models shows that most of the pilot projects on IoTSE were made using web services and microservices. More hardware-related component models such as OSGi and CORBA are least used for this purpose. 4. We do not find any evidence of component-based technology used for the development of the Crowler and Query processing elements of search engines. Our future work on this topic includes the selection of one component model and experimental development of the IoTSE system. 824 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY References Arellanes, D., Lau, K.K.: Evaluating iot service composition mechanisms for the scalability of iot systems. Future Generation Computer Systems 108, 827848 (2020). https://doi.org/https://doi.org/10.1016/j.future.2020.02.073, Brin, S., Page, L.: The anatomy of a large-scale hypertextual web search engine. Computer Networks and ISDN Systems 30(1-7), 107117 (1998) Cambazoglu, B.B., Karaca, E., Kucukyilmaz, T., Turk, A., Aykanat, C.: Architecture of a grid-enabled web search engine. Information Processing and Management 43(3), 609623 (2007). https://doi.org/https://doi.org/10.1016/j.ipm.2006.10.011, Special Issue on Heterogeneous and Distributed IR Chang, Y.S., Yuan, S.M., Lo, W.: A new multi-search engine for querying data through an internet search service on corba. Computer Networks 34(3), 467480 (2000). https://doi.org/https://doi.org/10.1016/S1389-1286(00)00131-6, Croft, W.B., Metzler, D., Strohman, T.: Search Engines: Information Retrieval in Practice. Addison-Wesley Professional (2010) Dang, T., Pham, L.T., Duong, T.Q.: Building an intelligent search engine for iot using microservices architecture. Journal of Computer Networks and Communications 2018 (2018). https://doi.org/10.1155/2018/9259473, Fathy, Y., Barnaghi, P., Tafazol i, R.: Large-scale indexing, discovery, and rankingfor the internet of things (iot). ACM Comput. Surv. 51(2) (Mar 2018).https://doi.org/10.1145/3154525 Giedrimas, V., Backys, G.: Component-based architecture of IoT search engine. Taikomieji tyrimai studijose ir praktikoje – Applied research in studies and practice 18(1), 100105 (Dec 2022),https://ojs.panko.lt/index.php/ARSP/article/view/183 Glatard, T., Étienne Rousseau, M., Camarasu-Pop, S., Adalat, R., Beck, N., Das, S., da Silva, R.F., Khalili-Mahani, N., Korkhov, V., Quirion, P.O., Rioux, P., Olabarriaga, S.D., Bel ec, P., Evans, A.C.: Software architectures to integrate workflow engines in science gateways. Future Generation Computer Systems 75, 239255 (2017). https://doi.org/https://doi.org/10.1016/j.future.2017.01.005 Gomes, P., Cavalcante, E., Rodrigues, T., Batista, T., Delicato, F.C., Pires, P.F.: A federated discovery service for the internet of things. In: Proceedings of the 2nd Workshop on Middleware for Context-Aware Applications in the IoT. p. 2530. M4IoT 2015, Association for Computing Machinery, New York, NY, USA (2015). https://doi.org/10.1145/2836127.2836129 Gubbi, J., Buyya, R., Marusic, S., Palaniswami, M.: Internet of Things (IoT): A vision, architectural elements, and future directions. Future Generation Computer Systems-The International Journal of Escience 29(7), 16451660 (SEP 2013). https://doi.org/{10.1016/j.future.2013.01.010} Gula, M., Fláková, K.: Proposal of component based architecture for internet of things: online laboratory case study. IFAC-PapersOnLine 50(1), 337342 (2017). https://doi.org/https://doi.org/10.1016/j.ifacol.2017.08.153, 20th IFAC World Congress Jin, X., Zhang, D., Zou, Q., Ji, G., Qian, X.: Where searching wil go in internet of things? In: 2011 IFIP Wireless Days (WD). pp. 13 (2011). Kamilaris, A., Papakonstantinou, K., Pitsil ides, A.: Exploring the use of dns as a search engine for the web of things. In: 2014 IEEE World Forum on Internet of Things (WF-IoT). pp. 100105 (2014). https://doi.org/10.1109/WFIoT. 2014.6803128 Kejriwal, M.: A meta-engine for building domain-specific search engines. Software Impacts 7, 100052 (2021). https://doi.org/https://doi.org/10.1016/j.simpa.2020.100052 Liu, Z., Morisset, C., Stolz, V.: rcos: Theory and tool for component-based model driven development. In: Arbab, F., Sirjani, M. (eds.) Fundamentals of Software Engineering. pp. 6280. Springer Berlin Heidelberg, Berlin, Heidelberg (2010) Mayer, S., Guinard, D.: An extensible discovery service for smart things. In: Proceedings of the Second International Workshop on Web of Things. WoT '11, Association for Computing Machinery, New York, NY, USA (2011). https://doi.org/10.1145/1993966.1993976 V. Giedrimas: Component Models for IoT Search Engine 825 Mrissa, M., Médini, L., Jamont, J.: Semantic discovery and invocation of functionalities for the web of things. In: 2014 IEEE 23rd International WETICE Conference. pp. 281286 (2014). https://doi.org/10.1109/WETICE.2014.50 Ozcan, R., Sengor Altingovde, I., Barla Cambazoglu, B., Junqueira, F.P., Özgür Ulusoy: A ve-level static cache architecture for web search engines. Information Processing and Management 48(5), 828840 (2012). https://doi.org/10.1016/j.ipm.2010.12.007, large-Scale and Distributed Systems for Information Retrieval Perera, C., Zaslavsky, A., Christen, P., Compton, M., Georgakopoulos, D.: Contextaware sensor search, selection and ranking model for internet of things middleware. In: 2013 IEEE 14th International Conference on Mobile Data Management. vol. 1,pp. 314322 (2013). https://doi.org/10.1109/MDM.2013.46 Ruppen, A., Pasquier, J., Meyer, S., Rüedlinger, A.: A component based approach for the web of things. In: Proceedings of the 6th International Workshop on the Web of Things. WoT '15, Association for Computing Machinery, New York, NY, USA (2015). https://doi.org/10.1145/2834791.2834792 Saari, M., Nurminen, M., Rantanen, P.: Survey of component-based software engineering within iot development. In: 2022 45th Jubilee International Convention on Information, Communication and Electronic Technology (MIPRO). pp. 824828 (2022). https://doi.org/10.23919/MIPRO55190.2022.9803785 Shemshadi, A., Sheng, Q.Z., Qin, Y.: Thingseek: A crawler and search engine for the internet of things. In: Proceedings of the 39th International ACM SIGIR Conference on Research and Development in Information Retrieval. p. 11491152. SIGIR '16, Association for Computing Machinery, New York, NY, USA (2016). https://doi.org/10.1145/2911451.2911471 Tan, C.C., Sheng, B., Wang, H., Li, Q.: Microsearch: When search engines meet smal devices. In: Proceedings of the 6th International Conference on Pervasive Computing. p. 93110. Pervasive '08, Springer-Verlag, Berlin, Heidelberg (2009). https://doi.org/10.1007/978-3-540-79576-6-6 Tang, J., Lu, X., Xiang, Y., Shi, C., Gu, J.: Blockchain search engine: Its current research status and future prospect in internet of things network. Future Generation Computer Systems 138, 120 141 (2023). https://doi.org/https://doi.org/10.1016/j.future.2022.08.008, Tran, N., Sheng, Q., Babar, M., Yao, L.: A kernel-based approach to developing adaptable and reusable sensor retrieval systems for the web of things. In: 18th International Conference on Web Information Systems Engineering (WISE 2017) : proceedings. Lecture Notes in Computer Science, vol. 10569, pp. 315329. Springer, Springer Nature, United States (2017). https://doi.org/10.1007/978-3-319-68783-4-22 Tran, N.K., Sheng, Q.Z., Babar, M.A., Yao, L., Zhang, W.E., Dustdar, S.: Internet of things search engine. Commun. ACM 62(7), 6673 (Jun 2019). Vale, T., Crnkovic, I., de Almeida, E.S., Silveira Neto, P.A.d.M., Cavalcanti, Y.a.C., Meira, S.R.d.L.: Twenty-eight years of component-based software engineering. J. Syst. Softw. 111(C), 128148 (Jan 2016). https://doi.org/10.1016/j.jss.2015.09.019 Wang, H., Tan, C.C., Li, Q.: Snoogle: A search engine for pervasive environments. IEEE Transactions on Parallel and Distributed Systems 21(8), 11881202 (2010). https://doi.org/10.1109/TPDS.2009.145 826 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY RESEARCH IN PROGRESS AN EXAMINATION OF ANTECEDENTS AND CONSEQUENCES OF TECHNOSTRESS AMONG UNIVERSITY STUDENTS: TASK -TECHNOLOGY FIT PERSPECTIVE ABDUL KARIM KHAN, SAMINA QURATULAIN College of Business and Economics, United Arab Emirates University, Al-Ain, United Arab Emirates abdul.karim@uaeu.ac.ae, squratulain@uaeu.ac.ae Drawing on the task-technology fit framework, this study examines the antecedents and consequences of technostress students face in an online/blended-learning environment. The effect of task, technology, and individual students' characteristics was hypothesized as predictors of technostress. The impact of technostress was also examined on students' satisfaction with online learning and expected performance. Preliminary findings based on a sample of 261 university students suggest that tasks requiring interdependence and cooperative learning caused higher technostress, while technology characteristics such as Keywords: perceived ease of use and usefulness negatively effect technostress, interdependence, technostress. Students' conscientiousness positively affect ease of perceptions of technostress. Unlike previous research, which use, examined a few antecedents of technostress, this study provides conscientiousness, self-efficacy a more nuanced understanding of the causes of technostress in online an online higher educational context. learning DOI https://doi.org/10.18690/um.fov.4.2023.54 ISBN 978-961-286-751-5 828 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Universities are pursuing technology-enhanced learning as an important agenda for the upgradation of students' learning experiences. Newer learning methods (e.g., flipped classrooms and blended learning) are ways to incorporate the digitalization of teaching and learning resources. This technology-enhanced delivery of instructions enriches students’ learning experience and improves their knowledge acquisition and access to learning resources (Brooker, Corrin, De Barba, Lodge, & Kennedy, 2018; Tuapawa, 2017). However, these new forms of learning require time, skil s, and effort investments and may create strain for the students (Mehta et al., 2019; Paul & Glassman, 2017). As a result, students can experience technostress –distress associated with the need to use and adapt to new digital technology (Gaudioso, Turel, & Galimberti, 2017; Vuori, Helander, & Okkonen, 2019). Although the use of technology itself can be a source of technostress, the optimal fit among tasks, technology, and individual characteristics can alleviate this stress (Ayyagari et al., 2011). In addition, technology can be a source of motivation for individuals when it eases the task and enables them to achieve the desired outcome in an efficient manner (Cascio & Montealegre, 2016). There can be situations where task, technology, and individual capabilities present a poor fit, which causes technostress, and individuals engage in a coping process to manage that stress. The following sections explain the main constructs of this study, the methodology employed to assess the hypotheses, and the preliminary findings. 2 Literature Review and Theoretical Framework 2.1 Task Technology Fit (TTF) Task-technology fit (TTF) theory suggests an interrelationship of three components. First, task requirements, technological functionality, and individual capabilities should be matched to achieve optimal performance (Goodhue, 1995; Goodhue & Thompson, 1995). For example, a task may require a different level of detail, and individuals performing this task will use different cognitive and physical resources (capabilities) to complete the task when technological functionality (e.g., tools used, support services available) matches with the task requirements, individual experience high motivation. Furthermore, individual characteristics represent attributes of their technological proficiency and capabilities to perform the assigned task. Internal A. Karim Khan, S. Quratulain: Component An Examination of Antecedents and Consequences of Technostress among University Students: Task -Technology Fit Perspective 829 resources such as self-efficacy, conscientiousness motivation, and experience using different technologies can contribute to high TTF perceptions. A high level of TTF perceptions wil increase technology utilization as it matches the task requirements and individual characteristics. On the contrary, distress and frustration are experienced when technology characteristics do not match task requirements and individual capabilities. 2.1.1 Task Requirements Tasks requiring collaboration and interaction among students are essential for students’ learning (Laurillard, 1993; Ramsden, 1992). Although prior research on collaborative learning has mainly focused on students' face-to-face interactions, it is unclear how technology and teamwork can be integrated. Furthermore, the extent and depth of interaction are expected to be limited in an online learning environment, and students experience anxiety and frustration in collaborative learning activities (Bakhtiar, Webster, & Hadwin, 2018). Causes of anxiety stem from delayed responses from group members, misinterpretation, and worrying about grades affected by other members' performance (Donelan and Kear, 2018). 2.1.2 Technological Characteristics Technological characteristics related to ease of use and usefulness determine the utilization of any technological change. However, the adoption of technology is usual y beyond the control of a user because the use of technology is based on compliance (decided by the university). Thus, involuntary adoption of not-so-useful technology creates a perception of demands-resources misfit. Individuals perceive that technology does not help facilitate the completion of required tasks and believe there could be better ways to accomplish the task. These perceptions increase stress and anxiety (Sami & Pangannaiah, 2006). 2.1.3 Individual Characteristics Self-efficacy is people's self-appraisal of their abilities to perform the designated tasks (Bandura, 1986). Positive self-appraisal in any situation significantly affects perceived stress. In addition, individuals with high self-efficacy can learn new skil s and adapt to technological changes (Ellen, Bearden, & Sharma, 1991). 830 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Conscientiousness is a personality trait that directs an individual's attention toward personal growth and work accomplishment. Personality predisposes individuals to pursue certain goals, espouse particular values, and behave differently (Lazarus & Folkman, 1984). Individuals high in conscientiousness place greater emphasis on personal growth and success. Thus, any changes that can reduce their opportunities for success and growth make them stressed (Rodell & Judge, 2009). Changes in the learning environment by introducing technology and collaborative task requirement can be stressful for conscientious students as it may reduce their prospects of growth and accomplishments. 2.2 Technostress Technostress is described as “a modern disease of adaptation caused by an inability to cope with new computer technologies in a healthy manner” (Brod, 1984, p.16). Technostress captures five dimensions: 1. techno-overload (a perception of overload experienced due to excessive use of technology), 2. techno-invasion (feeling of no boundaries between personal and work life), 3. techno-complexity (difficulties in learning technology and related features), 4. techno-insecurity (job insecurity due to technology) and 5. techno-uncertainty (difficulty in keeping pace with technological changes). 2.3 Satisfaction with online learning and expected performance. Students can find online learning an exhausting experience because the deluge of information, the expectation of fast response, and the integration of learning into their daily lives affect their satisfaction with learning and performance (Yin et al., 2018). Exhaustive experience can deplete students' mental resources, thus reducing their willingness to participate (Ayyagari et al., 2011) actively, lowering their satisfaction (Kim et al., 2015), and weakening performance (Ayyagari et al., 2011) 2.4 Hypotheses. H1: Collaborative learning and goal interdependence positively affect the perceptions of technostress. A. Karim Khan, S. Quratulain: Component An Examination of Antecedents and Consequences of Technostress among University Students: Task -Technology Fit Perspective 831 H2: Technology characteristics (ease of use and usefulness) negatively affect the perceptions of technostress. H3: Self-efficacy belief negatively affects the perceptions of technostress. H4: Conscientiousness positively affects the perceptions of technostress. H5: Technostress negatively affects satisfaction with online learning. H6: Technostress negatively affects students' expected performance. 3 Method Data for this study was col ected by distributing an online survey to undergraduate students of one of the United Arab Emirates' private sector universities. Students were enrol ed in a management course that was delivered online. Participation in this study was voluntary, and anonymity of responses was ensured. Nevertheless, 470 students were invited, and 261 completed responses were received. 3.1 Measures. • Cooperative Learning (7-items) and goal interdependence (4-items) scales by Johnson and Johnson (1983) were used to operationalize task requirements. • Technology characteristics of perceived ease of use and usefulness of online learning were assessed using 12 items measure of Davis (1989). • Self-ef icacy belief was measured by the 7-item scale of Bandura et al. (1996). • Conscientiousness was measured by Hendriks et al. (1999) 20 traits inventory. • Technostress was measured by Brooks et al. (2017) on 23 items scale. • Learning Satisfaction was measured by Cao et al. (2018) 4-item scale. • Expected performance was measured by Blasco-Arcas et al. (2013) 3-item scale. 832 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Figure 1: Hypothesized research model 3.2 Analyses and Results. Path analysis was conducted using AMOS 25. Results are presented in Table -1. Table 1: Hypotheses testing results. Hypotheses and Paths Coefficient T Statistics P-values Supported? H1: Collaborating learning -> 0.127** 3.93 0.000 Yes Technostress Goal Interdependence -> Technostress 0.253** 7.93 0.000 Yes H2: Ease of use -> Technostress -0.103** -3.70 0.000 Yes Usefulness -> Technostress -0.100** -2.77 0.000 Yes H3: Self-efficacy -> Technostress -0.057 -1.33 0.184 No H4: Conscientiousness -> Technostress 0.303** 9.85 0.000 Yes H5: Technostress -> Learning Satisfaction -0.228* -2.33 0.020 Yes H6: Technostress -> Expected -0.134 -1.38 0.166 No Performance ** p<.01 , * p<.05. A. Karim Khan, S. Quratulain: Component An Examination of Antecedents and Consequences of Technostress among University Students: Task -Technology Fit Perspective 833 These preliminary findings support our hypothesized model and highlight that technostress results from three interrelated components: technology, task, and individual characteristics. In addition, this study contributes to the existing literature on technostress by identifying task characteristics (collaborative learning and goal interdependence) and individual characteristics (self-efficacy and conscientiousness), which are rarely examined along with technological characteristics as predictors of technostress. Our findings indicate that when institutions are rolling out technology-enhanced learning, they need to be aware of the ensuing technostress which may impact the academic performance of students, resulting in more dropouts. In order to reduce the technostress, the institutions should focus on all three sets of factors i.e., personal dispositions, environment, and technology-related factors. This calls for a holistic approach to the management of technostress among the student population. These findings represent a work in progress due to the smal sample size, and efforts are underway to invite more respondents to replicate these findings with a bigger sample. The next step would be to study the organizational and group-level variables (such as organizational strategy and unit-level goals) to assess their moderating influences on students' technostress and related outcomes. References Ayyagari, R., Grover, V., & Purvis, R. (2011). Technostress: Technological antecedents and implications. MIS quarterly, 831-858. Bakhtiar, A., Webster, E. A., & Hadwin, A. F. (2018). Regulation and socio-emotional interactions in a positive and a negative group climate. Metacognition and Learning, 13, 57-90. Bandura, A. (1986). Fearful expectations and avoidant actions as coeffects of perceived self-inefficacy. Bandura, A., Barbaranel i, C., Caprara, G. V., & Pastorel i, C. (1996). Mechanisms of moral disengagement in the exercise of moral agency. Journal of personality and social psychology, 71(2), 364. Blasco-Arcas, L., Buil, I., Hernández-Ortega, B., & Sese, F. J. (2013). Using clickers in class. the role of interactivity, active collaborative learning and engagement in learning performance. Computers & Education, 62, 102–110. Brod, C. (1984). Technostress: The human cost of the computer revolution. Basic books. Brooks, S., Longstreet, P., & Califf, C. (2017). Social media induced technostress and its impact on internet addiction: A distraction-conflict theory perspective. AIS Transactions on Human-Computer Interaction, 9(2), 99-122. Brooker, A., Corrin, L., De Barba, P., Lodge, J., & Kennedy, G. (2018). A tale of two MOOCs: How student motivation and participation predict learning outcomes in different MOOCs. Australasian Journal of Educational Technology, 34(1). Cao, X., Masood, A., Luqman, A., & Ali, A. (2018). Excessive use of mobile social networking sites and poor academic performance: Antecedents and consequences from stressor-strain-outcome perspective. Computers in Human Behavior, 85, 163–174. 834 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Cascio, W. F., & Montealegre, R. (2016). How technology is changing work and organizations. Annual review of organizational psychology and organizational behavior, 3, 349-375. Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS quarterly, 319-340. Donelan, H., & Kear, K. (2018). Creating and col aborating: Students’ and tutors’ perceptions of an online group project. International Review of Research in Open and Distributed Learning, 19(2). El en, P. S., Bearden, W. O., & Sharma, S. (1991). Resistance to technological innovations: an examination of the role of self-efficacy and performance satisfaction. Journal of the academy of marketing science, 19, 297-307. Gaudioso, F., Turel, O., & Galimberti, C. (2017). The mediating roles of strain facets and coping strategies in translating techno-stressors into adverse job outcomes. Computers in Human Behavior, 69, 189-196. Goodhue, D. L. (1995). Understanding user evaluations of information systems. Management science, 41(12), 1827-1844. Goodhue, D. L., & Thompson, R. L. (1995). Task-technology fit and individual performance. MIS quarterly, 213-236. Hendriks, A. J., Hofstee, W. K., & De Raad, B. (1999). The five-factor personality inventory (FFPI). Personality and individual differences, 27(2), 307-325. Johnson, D. W., & Johnson, R. T. (1983). Social interdependence and perceived academic and personal support in the classroom. The Journal of Social Psychology, 120(1), 77-82. Kim, H. J., Lee, C. C., Yun, H., & Im, K. S. (2015). An examination of work exhaustion in the mobile enterprise environment. Technological Forecasting and Social Change, 100, 255–266. Laurillard, D. (1993). Balancing the media. Journal of educational Television, 19(2), 81-93. Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. Springer publishing company. Mehta, A., Morris, N. P., Swinnerton, B., & Homer, M. (2019). The influence of values on E-learning adoption. Computers & Education, 141, 103617. https://doi.org/10.1016/j.compedu.2019.103617 Paul, N., & Glassman, M. (2017). Relationship between internet self-efficacy and internet anxiety: A nuanced approach to understanding the connection. Australasian Journal of Educational Technology, 33(4). Ramsden, P., & Moses, I. (1992). Associations between research and teaching in Australian higher education. Higher Education, 273-295. Rodel , J. B., & Judge, T. A. (2009). Can “good” stressors spark “bad” behaviors? The mediating role of emotions in links of chal enge and hindrance stressors with citizenship and counterproductive behaviors. Journal of Applied Psychology, 94(6), 1438. Sami, L. K., & Pangannaiah, N. B. (2006). “Technostress” A literature survey on the effect of information technology on library users. Library review, 55(7), 429-439. Tuapawa, K. (2017). Interpreting experiences of students using educational online technologies to interact with teachers in blended tertiary environments: A phenomenological study. Australasian Journal of Educational Technology, 33(1). Vuori, V., Helander, N., & Okkonen, J. (2019). Digitalization in knowledge work: the dream of enhanced performance. Cognition, Technology & Work, 21(2), 237-252. DOCTORAL CONSORTIUM A PROPOSAL FOR A STUDY OF THE PROCESS ASPECT OF THE INTEGRATED LIFELONG TREATMENT OF HEALTHCARE TO PATIENTS ŽIVA RANT, TOMAŽ KERN University of Maribor, Faculty of Organizational Sciences, Kranj, Slovenia ziva.rant@student.um.si, tomaz.kern@um.si Healthcare is in crisis. On the one hand, the population is ageing, on the other hand, we are faced with a lack of resources: human, spatial and financial. That means that we have to make the health system more efficient. During a period of treatment, patients come into contact with a variety of health providers at different levels of the health system. That leads to the gaps between different instances of treatment. According to experiences in Keywords: healthcare, other fields, such as industry, a process-oriented view of health patient, treatment could be one of the components of a successful integrated solution. We wonder: How does viewing the medical treatment lifelong treatment, a patient receives throughout their life as a process affect the gaps business between different instances of treatment? process DOI https://doi.org/10.18690/um.fov.4.2023.55 ISBN 978-961-286-751-5 836 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Healthcare is in crisis. On the one hand, the population is ageing, which increases the number of patients requiring healthcare and necessitates ever greater amounts of medical treatment. While new treatment methods and new drugs help people live longer and enjoy better treatment outcomes, they also require ever greater financial investment. On the other hand, we are faced with a lack of resources: human (shortfal of doctors and other health workers), spatial and financial. We cannot exercise a significant impact on the requirements or the restrictions, which means that we have to make the health system more efficient. During a period of treatment, patients come into contact with a variety of health providers at different levels of the health system. Let us give a few examples. In Slovenia, 99.3% (source: NIPH, 2023) of children are born in a maternity unit. 99.4 % of births are attended by skil ed health personnel in the Members of the European Union ( Proportion (%) of Births Attended by Skil ed Health Personnel - European Health Information Gateway, 2022). This is their first encounter with health treatment and treatment. After discharge from hospital, the newborn is visited by a home care nurse, who provides care to the baby and encouragement to the mother. The baby undergoes systematic checks, with health professionals monitoring their development in a paediatric clinic, vaccinating them and providing a number of other services. When a patient falls ill in Slovenia, they usually visit a primary-level general practitioner (GP), who provides treatment. In more complex cases, the GP sends the patient to a specialist. In the event of an acute il ness or an injury, the patient attends an emergency centre, where they begin acute treatment. In more serious cases, acute treatment continues in hospital. After the patient’s condition improves, they move to non-acutFcroe treatment and an extended period of hospital treatment. After their discharge from Ž. Rant, T. Kern: A Proposal for a Study of the Process Aspect of the Integrated Lifelong Treatment of Healthcare to Patients 837 hospital, treatment continues at primary level. If the patient is referred for rehabilitation, this takes place when they are stil in hospital, and may continue at a health resort. After discharge from hospital, treatment, care and health promotion may also be provided by a home care nurse. For older patients, treatment and care may continue in a care home, with palliative care often being provided at medical institutions as wel . Patients can therefore be treated as fol ows in the course of their life: • at primary level: by a general practitioner, family doctor or paediatrician, • at secondary level: by specialists in specialist clinics and in hospitals (a patient may also move between hospital departments), • at tertiary level: at University Medical Centre, where the patient may move between clinics, • rehabilitation may be performed by hospitals, health resorts and private physiotherapists, • home care services, • home help, day centres for the elderly, care homes, hospices. According to experiences in other fields, such as industry (Hammer, 2015; Hammer & Champy, 1995, 2003; Keen, 1997; Keen & Knapp, 1995; Urh et al., 2022), a process-oriented view of health treatment could be one of the components of a successful solution. It could increase quality of treatment and reduce unnecessary healthcare treatment, the duplication of medical examinations and treatment time, thereby improving treatment outcomes, reducing hospitalisation rates and, not least, bringing down costs. The benefits of improving treatment outcomes and reducing hospitalisation and sick leave rates are manifold. Patient satisfaction and quality of life improve: they do not have to spend money on treatment, and can use this renewed period of health to create new value. The digital transformation of processes also adds a new dimension. 838 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 2 Problem definition The study wil look at patient treatment in al periods of life as a process. This process takes place in an environment we wil refer to as the ‘health system’. During a period of treatment, patients come into contact with a variety of health providers at different levels of the health system. Whether a patient’s treatment is successful in a given period frequently depends on previous medical treatment and the information available on that treatment. More than 99% of babies born in Europe are attended by skil ed health personnel. ( Proportion (%) of Births Attended by Skil ed Health Personnel - European Health Information Gateway, 2022), while patients are usual y treated by their GP or occasional y by specialists in other institutions. During a hospital stay, they are treated by yet another set of hospital staff. Treatment often involves nurses, physiotherapists, pharmacists or psychologists. The problems we can identify are the fol owing: P1 Organisational and informational gaps arise during medical treatment at different healthcare providers. (Bürkle et al., 2017; Amelung et al., 2021) P2 Attending health workers are not always apprised of the activity of previous treatments. (Amelung et al., 2021) P3 Attending health workers are not always apprised of the drugs that have been prescribed in the course of previous treatment or the drugs that the patient is currently taking. (Bürkle et al., 2017; Žerovnik et al., 2018). P4 Fragmented care tend to foster duplication and the inefficient use of resources, producing gaps in the care of patients with multimorbidities and reducing overal health sector capacity by pushing the best health care workers to focus on single diseases (WHO Global Strategy on People-Centred and Integrated Health Services Interim Report, 2015). P5 Whether a patient’s treatment is successful in a given period frequently depends on previous medical treatment and the information available on that treatment. Ž. Rant, T. Kern: A Proposal for a Study of the Process Aspect of the Integrated Lifelong Treatment of Healthcare to Patients 839 The main goal of the doctoral dissertation is to look at patient health treatment as a business process – that is, as the integrated lifelong treatment of patient healthcare. We would like to demonstrate that viewing treatment over the course of a patient’s life as a process has a significant positive impact on reducing the gaps between different instances of treatment and thereby improve the medical care given to a patient. An organisational process is the organisation (Rant, 2001; Rant, 2007) • of inputs – for example, people, equipment, energy, procedures, materials, information; • of a sequence of work activities; • of the coordination of actions between the people (and occasionally machines) involved in the activities required to produce • a certain end-result or output • that has value to the customer and to investors. If we look at the process of integrated lifelong treatment of healthcare to patient, we can regard the individual elements in the above definition as: • As inputs we regard o health workers: GPs, specialists, nurses, home care nurses, physiotherapists, care workers; o information – medical reports, discharge letters, recommendations, prescriptions, procedures, treatment results. • Work activities are instances of healthcare treatment at different healthcare levels – primary, in specialist clinics and in hospital. • The end-result is the outcome of treatment. • The customer is the patient – a newborn baby, a patient, an injury victim, a person taking part in preventive treatment. Based on previous research, we expect that it wil be possible to propose and implement a new "integrated organisational model for lifelong integration of patient care" using appropriate digital technologies. This means that it is expected that digital transformation of processes wil be required for the implementation of the model (Davenport & George, 2018; Raskino & Waller, 2015; Sajja, 2017; Van Veldhoven & Vanthienen, 2022). 840 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 2.1 Research question How does viewing the medical treatment a patient receives throughout their life as a process affect the gaps between different instances of treatment? 3 Methodology 3.1 Research approaches The following research approaches wil be employed: 3.1.1 Case Study Research Methodology (Yin, 2018; Kljajić Borštnar, 2021) This involves an in-depth study of a certain situation (integrated healthcare), which differs markedly from existing treatment practices. 3.1.2 Design Science Research If we develop a new conceptual organisational model, we will also make use of Design Science Research (Hevner, 2007; Hevner et al., 2004) 3.1.3 Systems Development Life Cycle (SDLC) To analyse the problem, we wil use the Systems Development Life Cycle model (Dennis et al., 2014; Valacich et al., 2017). This study will use the first two activities of this methodology: planning and analysis. We will use Directly observing users (Valacich et al., 2017) and Analyzing procedures and other documents (Valacich et al., 2017). Ž. Rant, T. Kern: A Proposal for a Study of the Process Aspect of the Integrated Lifelong Treatment of Healthcare to Patients 841 3.1.4 Methodology for resolving a problem and methodology for building systems (Gričar & Piskar, 1988) 3.2 Working procedure We wil conduct 1. a literature review 2. an analysis of the current situation 3. a definition of the problems 4. an analysis of the problem 5. a proposed solution 6. an evaluation 4 Preliminary/expected results 4.1 Literature review We are reviewing the literature in several broad fields: • Literature in the field of integrated lifelong treatment of healthcare to patients. • Literature in the field of business processes. • Literature in the field of system theory. • Literature in digital transformation. Based on the findings from the literature, we will elucidate the problems identified by the authors. We are reviewing the literature on the basis of the fol owing key words: seamless, cross-sector treatment, integrated care, coordinated care, seamless coordinated care, Re-Engineered Discharge (RED), seamless transition (for patient), person-centred care co-ordination, integral treatment, holistic care, continuous patient care. 842 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY We have found studies on specific transitions between activities (from hospital to home care, transitions between the treatment of children and adolescents or adults (Nadarajah et al., 2021)) and on specific areas (eating disorders, EDs) (Nadarajah et al., 2021; Wade, 2022), chronic heart failure (Yang et al., 2022), heart attack (Robyn Blackadar & Mishaela Houle, 2009), anorexia nervosa (Stocker et al., 2022), frail older people (D. Kodner, 2012). However, we did not find studies of integrated lifelong treatment of healthcare to patients. The business process literature focuses mainly on business processes in industry, not on business processes in healthcare (Dumas et al., 2018; Hammer, 2015; Hammer & Champy, 1995, 2003; Keen, 1997; Keen & Knapp, 1995; Urh et al., 2022). 4.2 Further results We will carry out an analysis of the existing situation. With the help of real-world data collected by the National Institute of Public Health, we wil attempt to provide evidence for the problems we have identified regarding organisational and informational gaps. It is not enough for data to exist. It must be used. (Stevens et al., 2022) define the flow of data through the ecosystem: collection, transfer, processing and interpretation. (Rant, 2010) establishes that the data process comprises: the receipt of data, data control, the preparation of a database, a standard report and analysis, and the dissemination of data and information. We have obtained the number of discharge letters reported to the central repository of patient data (CRPD). We have obtained information on the number of hospitalisations completed in hospitals. We wil also try to obtain information on the number of times doctors have viewed discharge letters in the CRPD. This will help us to demonstrate that attending health professionals are not always apprised of the course of previous treatments. We have obtained data on prescriptions issued, and wil attempt to obtain information on the number of times prescriptions have been viewed. This will help us to demonstrate that attending health workers are not always apprised of the drugs Ž. Rant, T. Kern: A Proposal for a Study of the Process Aspect of the Integrated Lifelong Treatment of Healthcare to Patients 843 that have been prescribed in the course of previous treatment or the drugs that the patient is currently taking. From both aspects we wil define the problems, analyse them and make a proposal for a solution, which we will also evaluate. 4.3 Scientific contribution Medical treatment must place the patient at the centre and connect medical professionals around him. We can see treatment of a patient as a business process and individual types of medical treatment as phases and activities within this process. In process terms, we can deal with the general practitioner as the owner of the process, in conjunction with the patient and his family. Here the GP operates on the ‘case manager’ or ‘care manager’ principle. They manage the treatment process as a whole, calling on specialists in other fields for more in-depth activities. Access to information on previous instances of medical treatment must be provided to those currently providing healthcare at any level. This can be provided by a single repository: the central electronic health record (EHR). In Slovenia this is the Central Registry of Patient Data (CRPD). We are looking for the intersection between business processes, healthcare and digital transformation (Figure 1). 844 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Figure 1: The intersection between business processes, healthcare and digital transformation 4.4 Objectives of the doctoral dissertation and the envisaged scientific research results The aim of the research is to look at healthcare treatment to a patient as a business process, i.e. as the integrated lifelong treatment of healthcare to a patient. We would like to demonstrate that viewing treatment over the course of a life as a process can have a positive impact on reducing the gaps between different instances of treatment and thereby improve the medical care given to a patient. The objectives of the doctoral dissertation are: • To explore the literature in the field of integrated lifelong treatment of healthcare to patients – examination of the theoretical premises. • To bring together the theoretical premises in this field. • To determine the stakeholders in integrated lifelong patient treatment. • To research the problems (chal enges and difficulties) that arise in healthcare treatment (analysis of the problem). • To lay the foundations for an integrated organisational model of integrated lifelong treatment of healthcare to patients. Ž. Rant, T. Kern: A Proposal for a Study of the Process Aspect of the Integrated Lifelong Treatment of Healthcare to Patients 845 4.5 Potential results of the study and the importance of those results The results of the study will contribute to an in-depth understanding of the process view of integrated lifelong treatment of healthcare to patients. The findings will therefore contribute to knowledge in this field, as they will build on previous studies in this field. The result wil also contain a presentation of the conceptual organisational model of integrated lifelong treatment of healthcare to patients based on theoretical findings and practical experiences. This wil enable us to describe healthcare treatment as a business process from birth to death. In addition to activities in terms of organisation, we will also look in detail at the information systems associated with those activities. This requires us to set the following requirements: 1. Access to data on al instances of healthcare treatment is required if the work is to be successful. 2. Access to information on previous instances of medical treatment must be provided to those currently providing healthcare at any level. 3. A single repository is required – central EHR. An example of this is the Slovenian Central Registry of Patient Data. 5 Future development In our research, we aim to show that looking at lifelong healthcare as a process has a significant positive impact on reducing the gaps between different treatments and, as a result, can improve patient care. The implementation of such a view is beyond the scope of our research and may be a chal enge in the future. It would also be interesting to longitudinal y monitor the results of the introduction of a process view of lifelong healthcare and to compare them in terms of the consequences of different treatments. 846 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY References Amelung, V., Stein, V., Suter, E., Goodwin, N., Nolte, E., Ran, ·, & Editors, B. (2021). Handbook Integrated Care Second Edition. Bürkle, T., Denecke, K., Lehmann, M., Zetz, E., & Holm, J. (2017). Integrated care processes designed for the future healthcare system. Studies in Health Technology and Informatics, 245, 20–24. https://doi.org/10.3233/978-1-61499-830-3-20 Davenport, T. H., & George, W. (2018). Why So Many High-Profile Digital Transformations Fail. Harvard Business Review, march 09, 2018. https://hbr.org/2018/03/why-so-many-high-profile-digital-transformations-fail Dennis, A., Wixom, B. H., & Roth, R. M. (2014). Systems analysis and design. John Wiley & Sons. http://www.saigontech.edu.vn/faculty/huynq/SAD/Systems_Analysis_Design_UML_5th ed.pdf Dumas, M., La Rosa, M., Mendling, J., & Reijers, H. A. (2018). Fundamentals of business process management: Second Edition. Fundamentals of Business Process Management: Second Edition, 1–527. https://doi.org/10.1007/978-3-662-56509-4/COVER Gričar, J., & Piskar, S. (1988). Sistemski inženiring : celostna sistemska metodologija za ustvarjalno reševanje problemov (B. Lipičnik, Ed.). ZOP - Zavod za organizacijo poslovanja. Hammer, M. (2015). What is business process management? Handbook on Business Process Management 1: Introduction, Methods, and Information Systems, 3–16. https://doi.org/10.1007/978-3-642-45100-3_1 Hammer, M., & Champy, J. (1995). Preurejanje podjetja : manifest revolucije v poslovanju. Gospodarski vestnik. Hammer, M., & Champy, J. (2003). Reengineering the Corporation: A Manifesto for Business Revolution (Col ins Business Essentials). 272. http://www.amazon.com/Reengineering-Corporation-Manifesto-Revolution-Essentials/dp/0060559535 Hevner, A. R. (2007). A Three Cycle View of Design Science Research. Scandinavian Journal of Information Systems, 19(2). Hevner, A. R., March, S. T., Park, J., & Ram, S. (2004). Design science in information systems research. MIS Quarterly: Management Information Systems, 28(1), 75–105. https://doi.org/10.2307/25148625 Keen, P. G. W. (1997). The Process Edge : Creating Value Where It Counts. Harward Business School Press. Keen, P. G. W., & Knapp, E. M. (1995). Every manager’s guide to business processes : a glossary of key terms & concepts for today’s business leader : a glossary of key terms & concepts for today’s business leader. Harvard. Kljajić Borštnar, M. (2021). Raziskovanje informacijskih sistemov. Kodner, D. (2012, February 7). Integrated care models for frail older people | The King’s Fund. https://www.kingsfund.org.uk/audio-video/dennis-kodner-integrated-care-models-frail-older-people Nadarajah, A., Dimitropoulos, G., Grant, C., Webb, C., & Couturier, J. (2021). Impending Transition From Pediatric to Adult Health Services: A Qualitative Study of the Experiences of Adolescents With Eating Disorders and Their Caregivers. Frontiers in Psychiatry, 12. https://doi.org/10.3389/FPSYT.2021.624942 Proportion (%) of births attended by skil ed health personnel - European Health Information Gateway. (2022). https://gateway.euro.who.int/en/indicators/hfa_598-7070-proportion-of-births- attended-by-skilled-health-personnel/ Rant, Ž. (2001). Kontinuirano učenje kot stalnica v procesni organizaciji : magistrska naloga. [Ž. Rant]. Rant, Ž. (2007). Procesi obstajajo tudi v zdravstvu = Processes exist also in health care. In V. Rajkovič (Ed.), Ustvarjalna organizacija (p. Str. 1592-1599). Moderna organizacija. Rant, Ž. (2010). Prenova procesa obdelave podatkov : študija primera. Organizacija, 43(6), str. A246-A254. Ž. Rant, T. Kern: A Proposal for a Study of the Process Aspect of the Integrated Lifelong Treatment of Healthcare to Patients 847 http://organizacija.fov.uni-mb.si/index.php/organizacija-si/article/viewFile/1218/988 Raskino, M., & Wal er, G. (2015). Digital to the Core: Remastering Leadership for Your Industry, Your Enterprise, and Yourself. Bibliomotion, Inc. Robyn Blackadar, & Mishaela Houle. (2009). The Alberta Cardiac Access Collaborative: Improving the Cardiac Patient Journey. Healthcare Quarterly, 13(Sp), 85–90. https://www.longwoods.com/product/21104 Sajja, P. S. (2017). Essence of Systems Analysis and Design: A Workbook Approach. Springer Verlag, Singapore. Stevens, G., Hantson, L., Larmuseau, M., & Verdonck, P. (2022). A human-centered, health data-driven ecosystem. Discover Health Systems 2022 1:1, 1(1), 1–12. https://doi.org/10.1007/S44250-022-00011-9 Stocker, A., Rosenthal, L., Mesquida, L., Raynaud, J. P., & Revet, A. (2022). Adult and child and adolescent psychiatrists’ experiences of transition in anorexia nervosa: a qualitative study. Journal of Eating Disorders, 10(1). https://doi.org/10.1186/S40337-022-00610-0 Urh, B., Kern, T., & Krhač Andrašec, E. (2022). Model ocenjevanja učinkov organizacijskih sprememb. 1011–1025. https://doi.org/10.18690/UM.FOV.3.2022.73 Valacich, J. S., George, J. F., Columbus, B., New, I., San, Y., Cape, F. A., Dubai, T., Madrid, L., Munich, M., Montréal, P., Delhi, T., São, M. C., Sydney, P., Kong, H., Singapore, S., & Tokyo, T. (2017). Modern Systems Analysis and Design 8th Edition. www.pearsoned.com/permissions/. Van Veldhoven, Z., & Vanthienen, J. (2022). Digital transformation as an interaction-driven perspective between business, society, and technology. Electronic Markets, 32(2), 629–644. https://doi.org/10.1007/s12525-021-00464-5 Wade, T. D. (2022). A systematic review: Solutions to problems caused by age transition between eating disorder services. European Eating Disorders Review. https://doi.org/10.1002/ERV.2945 WHO global strategy on people-centred and integrated health services Interim Report. (2015). www.who.int Yang, Y. F., Hoo, J. X., Tan, J. Y., & Lim, L. L. (2022). Multicomponent integrated care for patients with chronic heart failure: systematic review and meta-analysis. ESC Heart Failure. https://doi.org/10.1002/EHF2.14207 Yin, R. K. (2018). Case Study Research and applications, 6th edition. Paper Knowledge . Toward a Media History of Documents, 414. Žerovnik, Š., Locatelli, I., & Kos, M. (2018). Brezšivna skrb pri zdravljenju z zdravili v Sloveniji. Farmacevtski vestnik, 69(3). 848 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY DOCTORAL CONSORTIUM UNDERSTANDING PSYCHOLOGICAL CONTRACTS IN THE CONTEXT OF ORGANIZATIONAL CHANGE PROCESSES RELATED TO DIGITAL GOVERNMENT TRANSFORMATION WILLEMIJN A VAN HAEFTEN Hogeschool Utrecht University of applied sciences, Ultrecht, Netherland wil emijn.vanhaeften@hu.nl Digital transformation has revolutionized public sector organizations, bringing both opportunities and chal enges. This research proposal suggests an exploration of the impact of Digital transformation on public sector employees' psychological contracts and the chal enges faced by civil servants in response to this change. The study proposes a study consisting of four parts. The research methodology combines qualitative and quantitative approaches to capture the dynamic nature of psychological contracts and psychological contract breach in the public sector. The research aims to explore the distinct characteristics of civil servants' dynamic psychological contracts in the context of digital government transformation and how they experience psychological contract breaches during Keywords: continuous organizational change processes. The study psychological highlights the need for understanding employees’ tactics to contracts, manage psychological contract breaches and improve sustainable digital transformation, employability during digital transformation related change. The organizational implications of the study will contribute to the effective change, implementation of digital transformation in the public sector and public sector, enhances our understanding of employee perspectives on Bled managing psychological contracts. eConference DOI https://doi.org/10.18690/um.fov.4.2023.56 ISBN 978-961-286-751-5 850 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction The rapid pace of technological, economic, and social developments in the last decade has resulted in a digital transformation (DT) affecting activities and processes in the Dutch economy and society (EZK, 2021 p.6; Digitale Overheid, 2021). Digital technologies have become vital (Keijzer, 2021), not only for our society (Digitale Overheid, 2021), but also for the development and the execution of governmental processes and services (FreedomLab, 2021). For the Dutch Government, the ambition is to embrace innovations as set out in their plan: “Towards a safe, people-oriented, transparent, effective digital future”, (Digitale Overheid, 2021). While digitization offers opportunities to make public administration more effective, efficient, transparent and fair, it also increases the risk of rigid enforcement, with an excessive focus on measurable and control able matters (Freedomlab, 2021), as seen for example in the chal enges the Dutch Tax Authorities faces. Examples of other chal enges faced during digitalization are the air traffic control project (iCAS), the complexity of the Environmental Act software, and the new communication system C2000 of the Dutch police departments. More general y, the constantly evolving public sector and its employees are facing many new chal enges in response to digitization (Groeneveld, Kuipers & Van der Voet, 2022; Van der Wal, 2017a; 2017b). Traditionally, public values such as legality, equal treatment, and efficiency have been critical within public administration, with processes and formal hierarchy determining accountability and political legitimacy. However, with the accelerated change caused by digital transformation, values such as responsiveness, resilience, and innovation are becoming increasingly important in response to the VUCA environment (Groeneveld et al., 2022). Nowadays, civil servants are expected to balance the contradicting bureaucratic values and innovation values, (Figure 1), placing them in a difficult and ambiguous position especial y since political and societal tolerance for error and experimentation with taxpayers’ money is minimal (Van der Wal, 2017a). W. A van Haeften: Understanding Psychological Contracts in the Context of Organizational Change Processes Related to Digital Government Transformation 851 Figure 1: Position of civil servants, source: Van der Wal (2017a) Since most public sector employees are intrinsically motivated to work for the government (BKZ, OCW & CBS, 2022), the conflicting values caused by DT (FreedomsLab 2021) deeply affect civil servants and their psychological contracts (PCs). Although the importance of an individual’s ability to adapt to (the digital) changes is underscored, (BKZ, OCW & CBS, 2022) there is a need to understand how to align the motivation of people working in the public sector with these changes associated with DT and how to manage the impact. It is this need that has triggered this research. This research aims to contribute to the academic and societal debate on PCs in the context of change by exploring the distinct characteristics of civil servants’ dynamic PCs in the context of digital transformation in public sector organizations. More specifical y, the focus is on how civil servants’ experience PC breaches during continuous organizational change processes related to digital transformation and what repair tactics are used to prevent and repair PC breaches. Subsequently, the practical goal is to research concrete managerial and employees’ tactics to prevent and repair PC breaches in public sector organizations adopting new ways of working supported by digital technology. This can improve sustainable employability and enhance work engagement and performance. 852 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY We thus postulate the following research question: How do organizational change processes caused by digital government transformation impact psychological contracts of employees? To answer this question the following sub-research questions are formulated: SRQ1: What is known in existing literature from an employee’s perspective on characteristics of the public sector, digital transformation, and psychological contracts? SRQ2: How do civil servants perceive the dynamics in their psychological contract in the context of digital transformation? SRQ3: To what extent does digital transformation influence the psychological contracts of civil servants? SRQ4: What tactics can public sector employees employ to manage psychological contracts breaches during digital transformation? The structure of this research proposal is as follows: In section 2 the proposed methodology is described. In section 3, the theoretical background of the research is presented. The 4th section provides a description of the aimed results. 2 Methodology This chapter outlines the methodology proposed to investigate the impact of digital transformation on psychological contracts of public sector employees. Several authors have suggested that a combination of qualitative and quantitative research is required to capture how PC and PC breach develop in dynamic environments (Achnak & Hansen, 2019; Weinhardt et al., 2019 ; Wiechers et al., 2022). Therefore, this proposed research begins with a systematic literature review W. A van Haeften: Understanding Psychological Contracts in the Context of Organizational Change Processes Related to Digital Government Transformation 853 (study 1) to further explore the existing research, followed by a qualitative multiple-case study (study 2) and a quantitative daily diary study (study 3). The research is concluded by a design study focused on repair tactics of public employees (study 4), with the aim to support public managers in managing the dynamic PCs of civil servants in the context of DT. Figure 2 visualizes the research questions and research designs. Figure 2: Research design Study 1 The first study is a systematic literature study that utilizes the PRISMA-approach (Page et al., 2020), resulting in a transparent research process visual in a flowchart of the research process as wel as a literature synthesis as results. Inclusion criteria for the review wil be based on relevance to the research question, date of publication and quality of the research (Page et al., 2020). Aim is to create an assessable overview of existing literature from an employee’s perspective on characteristics of the public sector, digital transformation, and psychological contracts. 854 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Study 2 In a qualitative research design, the understanding of experiences requires open or semi-structured interviews with room to elaborate on concepts (Verhoeven, 2014). Study 2 adopts a qualitative design and comprises of in-depth, semi-structured interviews with the purpose of gaining a deeper understanding of how civil servants experience their PC dynamics in the context of DT (SRQ2). To capture the dynamic narrative of PC breaches and responding employees’ actions, the so-cal ed Critical Incident Technique will be applied. In this approach, the interview starts with a specific incident, which forms the starting point for the reconstruction of a chronological and meaningful narrative over time for the respondent (Langley & Meziani, 2020). This technique is well suited to capturing recollections of vivid PC breaches and the responses to these by both parties to the employment contract (Conway & Briner, 2005). Through collaboration with the iPoort and iPartnership networks as wel as via part-time students of Master of Informatics, suitable case studies (i.e. finished projects) wil be selected from the government’s digital transformation programs. Organizations and participants will be selected through a purposive sampling (Yin, 2017). The focus will be on public organizations of different sizes with civil servants closely situated to civilians. Study 3 The third study is a deductive quantitative study (Verhoeven, 2014) and aims to measure relationships between the key factors of DT and the PCs of civil servants in answering SRQ3. A temporal daily diary study will be conducted as it has the capability to study processes over time, asses day-to-day experiences and enable repeated measurements. This design increases the validity and reliability of the study’s outcomes and captures the dynamic nature of PCs. Additional y, its benefit over a cross-sectional questionnaire method is the possibility of examining immediate affective responses, limiting the risks of poor recal of events (Conway & Briner, 2005). To capture the nested within-person processes, the study wil be set up in two different public sector organizations undergoing a technology W. A van Haeften: Understanding Psychological Contracts in the Context of Organizational Change Processes Related to Digital Government Transformation 855 implementation process as part of a larger digitalization program. This provides the researchers the possibility to integrate a prospective view as a factor influencing the sense-making process after the PC breaches experienced by respondents. The daily survey is administered online and uses push messages to initiate response. Including daily mood ratings, validated measures of PC, organizational change, and digital transformation, as well as demographic questions. The data will be analyzed using multi-level mediation modelling, with the key variables examined within-person for their impact on employee’s PCs over the process of the technology implementation. Study 4 To answer SRQ4 a Participatory Action Design Research (PADR) is performed with a focus on possible tactics of public employees facing and recovering from PC breach in the context of DT. PADR is an especial y relevant research design when the research is conducted based on local needs of multiple public organizations (Venable et al., 2017). This study builds on the outcomes of all previous studies, and uses participative activities that are distinct in the PADR methodology, which allow to validate previous studies and lead to actionable organizational interventions. 3 Theoretical Background Preliminary research has been conducted in writing this research proposal. These results are described in this chapter. DT has a significant impact on society and is an important cause of today’s VUCA-environment. According to Van der Wal (2017b), this is especially true in public sector organizations, making the work of a 21st century public sector employee more challenging. Specifically, digital technologies profoundly reshape working arrangements in the public sector, and digitization may trigger a depersonalization of workplaces, impacting the social organizational climate (Dunleavy et al., 2005; Palumbo, 2021). Digital transformation requires a rethinking of employees' skil s, responsibilities, and competencies, as well as the overall structure of organizations and working values, culture in government (Tangi et al., 2021; Vial, 2019). Despite these chal enges, research on digital transformation in the public sector, also known 856 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY as ‘digital government transformation’, is still scarce, particularly from an employee perspective (Mergel et al., 2019; Tangi et al., 2021). An important perspective when studying organizational changes from an employee perspective is that of the psychological contract (PC) (Rousseau et al., 2018; Wiechers et al., 2022). One of the key concepts in the PC is reciprocity, which refers to the principle that if one party fulfil s their (implicit) obligations, the other party will respond in kind. This is in accordance with Adam’s (1965) need for equity theory in public administration (literature on the public sector). Psychological contract theory is based on Social Exchange Theory, which postulates that employees and employers engage in exchanges whereby each party to the exchange reciprocates the other’s contribution (Blau, 1964). Rousseau (1995) argued that perceived employer obligations motivate to engage in positive attitudes and work behaviors as expected reciprocity in the exchange relationship causes employees to attempt to restore balance if an imbalance is perceived. This reciprocity can operate at three levels of the exchange, at the same time: • Transactional contracts involve highly specific exchanges of narrow scope which take place over finite periods (e.g., provide labour for pay); • Relational contracts are broader, more amorphous, open-ended, and more subjectively understood by parties (e.g, extra work effort in exchange for promotion opportunities); • Ideological contracts, defined as an espousal to a cause, whereby employee contributions are dependent on the belief that their organization serves a higher purpose (e.g. , public sector organizations serving the public) (Blau, 1964; Rousseau, 1995). Guest (2004) found that organizational change is a predictor of PC breach, something Conway, Kiefer, Hartley and Briner (2014) confirmed for public organizations. A breach indicates a change of the mutual obligations, which affects the employee’s attitudes (Rousseau, 1995), and may be reciprocated by employees through disengagement (Zhao et al., 2007), reduced work effort, or even leaving the company (Bal et al., 2008; Conway & Briner, 2002). After a perceived breach there is a period of ‘sense-making’ in which the employee goes through a rationalization process, after which the employee can experience the negative feelings of a PC violation. Given that DT is a source of continuous, and often less predictable, W. A van Haeften: Understanding Psychological Contracts in the Context of Organizational Change Processes Related to Digital Government Transformation 857 change process (Vial, 2019), it can be argued that DT affects PC over time and may lead to breaches on a weekly or even daily interval (Griep & Van Tilborgh, 2018; Wiechers et al, 2022). Subsequently, this can lead to the above-mentioned employee attitudes which can result in unsuccessful implementation of digitalization. However, it is yet unclear how the dynamic nature of the PC and the sense-making process of an employee is affected by a continuous change. After a perceived PC breach, the importance of timely and adequate feedback of the organization for the PC’s recovery is essential (Rousseau et al., 2018; Van der Schaft et al., 2020). However, it is noteworthy that there is a lack of understanding about managerial tactics in this regard (De Ruiter et al., 2016). To enable adequate response, it is necessary to expand our understanding of appropriate tactics (behaviours) of public employees after a perceived breach. The consideration of public sector context is essential to understanding employee behavior (Coyle-Shapiro & Kessler 2003). It is therefore, necessary to take the characteristics of the public sector into account when studying PC dynamics, as the public sector differs fundamentally from the private sector (Meijer et al., 2022; Pandey & Wright, 2006). Even though the destabilizing impact of organizational change on employment relationships have been studied in both private and public sectors (Conway et al., 2014), literature on PC in the public sector is showing a considerable knowledge gap. Perry and Wise (1990) provide some insights as they found that even after a breach of perceived obligations, employees with a strong commitment to the public sector choose to remain engaged, because not to do so would be conflicting with their personal values. Moreover, Bunderson (2001) indicated that ideologies can shape PC and influence subsequent reactions to breach. This incorporation of ideology into the PC helps to explain why some employees may remain loyal to an organization after breaches in the transactional and ideological levels of the PC (Thompson & Bunderson, 2003). While public sector motivation (Perry & Wise, 1990) is commonly known in public administration, the notion of this motivation in the public sector is overlooked in the PC literature. 858 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 4 Expected results This research aims to contribute to both practical and scientific fields. The expected results are outlined in this chapter. The field of DT is currently emerging, while that of PCs is a wel -established academic field. However, there is limited research on the combination of PC and DT, particularly with respect to the public sector. This study aims to address PC in the public sector as the consideration of context is essential to the understanding of employees’ attitude and behaviors. Furthermore, it provides insight into contextual considerations in understanding how employees manage PCs. Moreover, the phenomenon of DT in relation to organizational change is distinct from past IT-related organizational changes and cannot be ful y explained using existing theoretical models (Markus & Rowe, 2018). To leverage and extend current knowledge, it is necessary to evaluate the compatibility between DT and established theoretical models in the field of organizational change (Hanelt et al., 2021). In response to the current VUCA environment with DT as driver for organizational changes, the need arises to understand the PC theory in the context of DT. This context of continuous change wil enhance our understanding of the management of dynamics of PC, for which scientific attention has been growing (Rousseau et al., 2018; Griep & Cooper, 2019). Additionally, the context of continuous change poses the possibility to take the effect of prospectively expecting more PC breaches into account as a variable in employees’ sense-making process, this will be innovative as this is an academic field that traditional y has been researched retrospective. Dunleavy et al. (2005) emphasize the importance of effectively guiding technology-centered changes in government. Different managerial skills are needed to govern change related to DT, as it often involves disruption of the status quo (Tangi et al., 2021). Even though DT could predict PC breaches, there is still a noteworthy lack of understanding of managerial actions suitable in managing breaches in PC (De Ruiter, 2017). This research addresses the knowledge gap on tactics used by employees (Henderson et al., 2020; Rousseau et al., 2018) by developing an artifact that takes the complexity of DT, the dynamics of PC and characteristics of the public sector into account. W. A van Haeften: Understanding Psychological Contracts in the Context of Organizational Change Processes Related to Digital Government Transformation 859 the practical contributions we aim to achieve are: 1. Improving the sustainable employability of civil servants by reducing the occurrence and impact of PC breaches; 2. developing managerial tactics tailored to the unique challenges of digital technology implementation; 3. Improving our understanding, from an employee perspective, how DT projects can be better organized to prevent unwanted outcomes as mentioned above. Ultimately, this research wil enhance our understanding of how public sector organizations and employees can manage PCs during times of change related to DT. References Achnak, S., & Hansen, S. D. (2019). 12. The role of time and timing in psychological contract research. In: Griep, Y & Cooper, C. (Eds.) Handbook of research on the psychological contract at work, (pp. 243-252), Edward Elgar Publishing.http://doi.org/10.4337/9781788115681 Adams, J. S. (1965). Inequity In Social Exchange. Advances in Experimental Social Psychology, 267– 299. https://doi.org/10.1016/s0065-2601(08)60108-2 Bal, P. M., De Lange, A. H., Jansen, P., & Van Der Velde, M. E. G. (2008). Psychological contract breach and job attitudes: A meta-analysis of age as a moderator. Journal of Vocational Behavior, 72(1), 143–158. https://doi.org/10.1016/j.jvb.2007.10.005 Blau, P. (1964). Exhange and Power in Social Life. Wiley. Bunderson, J. S. (2001). How work ideologies shape the psychological contracts of professional employees: doctors’ responses to perceived breach. Journal of Organizational Behavior, 22(7), 717–741. https://doi.org/10.1002/job.112 Conway, N., & Briner, R. B. (2002). A daily diary study of affective responses to psychological contract breach and exceeded promises. Journal of Organizational Behavior, 23(3), 287–302. https://doi.org/10.1002/job.139 Conway, N., & Briner, R. B. (2005). Understanding Psychological Contracts at Work:A Critical Evaluation of Theory and Research: A Critical Evaluation of Theory and Research. OUP Oxford. Conway, N., Kiefer, T., Hartley, J., & Briner, R. B. (2014). Doing More with Less? Employee Reactions to Psychological Contract Breach via Target Similarity or Spillover during Public Sector Organizational Change. British Journal of Management, 25(4), 737–754. https://doi.org/10.1111/1467-8551.12041 Coyle-Shapiro, J. (2003). The Employment Relationship in the U.K. Public Sector: A Psychological Contract Perspective. Journal of Public Administration Research and Theory, 13(2), 213–230. https://doi.org/10.1093/jopart/mug018 860 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY De Ruiter, M., Schalk, R., & Blomme, R. J. (2016). Manager Responses to Employee Dissent About Psychological Contract Breach. Management Communication Quarterly, 30(2), 188–217. https://doi.org/10.1177/0893318915623238 De Ruiter, M. (2017). Beyond repair? The role of supervisory leadership in the context of psychological contract Breach. Gildeprint Digitale Overheid. (2021). I-strategie Rijk 2021-2025: doorpakken op digitale transformatie. In digitaleoverheid.nl. https://www.digitaleoverheid.nl/document/i-strategie-rijk-2021-2025/ Dunleavy, P., Margetts, H., Bastow, S., & Tinkler, J. (2005). New public management Is dead--long live Digital-Era Governance. Journal of Public Administration Research and Theory, 16(3), 467– 494. https://doi.org/10.1093/jopart/mui057 FreedomLab. (2021). Toekomstverkenning Digitalisering 2030. In Rijksoverheid. https://open.overheid.nl/documenten/ronl-6381adb2-3806-4f21-8ffd-613784a3ecdd/pdf Henderson, K. E., Welsh, E. T., & O’Leary-Kelly, A. M. (2020). “Oops, I Did It” or “It Wasn’t Me:” An Examination of Psychological Contract Breach Repair Tactics. Journal of Business and Psychology, 35(3), 347–362. https://doi.org/10.1007/s10869-019-09624-z Griep, Y., & Cooper, C. (Eds.) (2019). Handbook of Research on the Psychological Contract at Work. Edward Elgar Publishing. Griep, Y., & Vantilborgh, T. (2018). Let’s get cynical about this! Recursive relationships between psychological contract breach and counterproductive work behaviour. Journal of Occupational and Organizational Psychology, 91(2), 421–429. https://doi.org/10.1111/joop.12201 Groeneveld, S., Kuipers, B., & Van der Voet, J. (2022). Anders kijken naar publiek leiderschap: Naar een duurzame bijdrage aan het oplossen van maatschappelijke problemen: Essay in opdracht van Bureau Algemene Bestuursdienst in het kader van traject Visie op Leiderschap. In Algemene Bestuursdienst. Leiden Leadership Centre (Universiteit Leiden). https://www.algemenebestuursdienst.nl/organisatie/documenten/publicatie/2022/01/27/a nders-kijken-naar-publiek-leiderschap Guest, D. (2004). Flexible employment contracts, the psychological contract and employee outcomes: an analysis and review of the evidence. International Journal of Management Reviews, 5–6(1), 1–19. https://doi.org/10.1111/j.1460-8545.2004.00094.x Hanelt, A., Bohnsack, R., Marz, D., & Marante, C. (2021). A Systematic Review of the Literature on Digital Transformation: Insights and Implications for Strategy and Organizational Change. Journal of Management Studies, 58(5), 1159–1197. https://doi.org/10.1111/joms.12639 Keijzer, M. C. G. (2021). Aanbiedingsbrief bij update Nederlandse Digitaliseringsstrategie 2021. In Rijksoverheid (Nr. 00000001003214369000). Geraadpleegd op 10 maart 2023, van https://open.overheid.nl/documenten/ronl-bb8a0829-c6ab-4781-8088-4fe8f2eda0c7/pdf Langley, A., & Meziani, N. (2020). Making interviews meaningful. The Journal of Applied Behavioral Science, 56(3), 370-391 Markus, M. L., & Rowe, F. (2018). Is IT Changing the World? Conceptions of Causality for Information Systems Theorizing. Management Information Systems Quarterly, 42(4), 1255–1280. https://doi.org/10.25300/misq/2018/12903 Meijer, A., Ingrams, A., & Zouridis, S. (2022). Public Management in an Information Age: Towards Strategic Public Information Management. Bloomsbury Publishing. Mergel, I., Edelmann, N., & Haug, N. (2019). Defining digital transformation: Results from expert interviews. Government Information Quarterly, 36(4), 101385. https://doi.org/10.1016/j.giq.2019.06.002 Ministerie van Binnenlandse Zaken en Koninkrijksrelaties [BKZ], Ministerie van Onderwijs, Cultuur en Wetenschap [OCW], & Centraal Bureau voor de Statistiek [CBS]. (2022). Kernrapport Werkonderzoek 2022: Uitkomsten van het Werkonderzoek 2022. In Rijksoverheid. Venster voor Medewerkers (ICTU). https://www.rijksoverheid.nl/documenten/publicaties/2022/11/22/kernrapport- werkonderzoek-2022-beschikbaar-inzicht-in-werkaantrekkelijkheid-bij-de-overheid Ministerie van Economische Zaken en Klimaat [EZK]. (2021). [Kamerstuk] Nederlandse Digitaliseringsstrategie 2021. In Rijksoverheid. Geraadpleegd op 2 maart 2023, van W. A van Haeften: Understanding Psychological Contracts in the Context of Organizational Change Processes Related to Digital Government Transformation 861 https://www.rijksoverheid.nl/documenten/kamerstukken/2021/04/26/ nederlandse- digitaliseringsstrategie-2021 Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T., Mulrow, C. D., Shamseer, L., Tetzlaff, J., Akl, E. A., Brennan, S. E., Chou, R., Glanvil e, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E., Mayo-Wilson, E., McDonald, S., Moher, D. (2021). The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ, n71. https://doi.org/10.1136/bmj.n71 Palumbo, R. (2021). Does digitizing involve desensitizing? Strategic insights into the side effects of workplace digitization. Public Management Review, 24(7), 975–1000. https://doi.org/10.1080/14719037.2021.1877796 Pandey, S., & Wright, B. E. (2006). Connecting the Dots in Public Management: Political Environment, Organizational Goal Ambiguity, and the Public Manager’s Role Ambiguity. Journal of Public Administration Research and Theory, 16(4), 511–532. https://doi.org/10.1093/jopart/muj006 Perry, J. L., & Wise, L. R. (1990). The Motivational Bases of Public Service. Public Administration Review, 50(3), 367. https://doi.org/10.2307/976618 Punch, K. F. (2016). Developing Effective Research Proposals. Sage. Rousseau, D. (1995). Psychological contracts in organizations: Understanding written and unwritten agreements. Sage. Rousseau, D. M., Hansen, S. E., & Tomprou, M. (2018). A dynamic phase model of psychological contract processes. Journal of Organizational Behavior, 39(9), 1081–1098. https://doi.org/10.1002/job.2284 Tangi, L., Janssen, M., Benedetti, M., & Noci, G. (2021). Digital government transformation: A structural equation modelling analysis of driving and impeding factors. International Journal of Information Management, 60, 102356. https://doi.org/10.1016/j.ijinfomgt.2021.102356 Thompson, J. A., & Bunderson, J. S. (2003). Violations of Principle: Ideological Currency in the Psychological Contract. Academy of Management Review, 28(4), 571–586. https://doi.org/10.5465/amr.2003.10899381 Van der Wal, Z. (2017a). De 21e-eeuwse overheidsmanager: Een reis door tijd, plaats en context. In Caop. CAOP. https://www.caop.nl/app/uploads/2019/05/OenA-2017-44-De_21e-eeuwse_overheidsmanager.pdf Venable, J. R., Pries-Heje, J., & Baskerville, R. L. (2017). Choosing a Design Science Research Methodology. The 28th Australasian Conference on Information Systems. Verhoeven, P. S. (2014). Wat is onderzoek?: praktijkboek methoden en technieken voor het hoger onderwijs. Van der Wal, Z. (2017b). The 21st Century Public Manager. Bloomsbury Publishing. Van Straalen, M. (2021). Verkenningsrapport HEO met regie naar verantwoordelijkheid. Vereniging Hogescholen. https://www.vereniginghogescholen.nl/system/knowledge_base/attachments/files/000/00 1/244/original/Verkenningsrapport_HEO_met_regie_naar_verantwoordelijkheid_-_VH_- _september_2021_-_def.pdf?1636551020 Vial, G. (2019). Understanding digital transformation: A review and a research agenda. Journal of Strategic Information Systems, 28(2), 118–144. https://doi.org/10.1016/j.jsis.2019.01.003 Weinhardt, J., Griep, Y., & Sosnowska, J. (2019). Toward a formal dynamic and computational modeling approach to better understand psychological contract dynamics. In: Griep, Y & Cooper, C. (Eds.) Handbook of Research on the Psychological Contract at Work (pp. 292-315), Edward Elgar Publishing. http://doi.org/10.4337/9781788115681 Wiechers, H., Coyle-Shapiro, J. A., Lub, X., & Have, S. T. (2022). The tremors of interconnected triggers over time: How psychological contract breach can erupt. Journal of Organizational Behavior, 43(7), 1172–1189. https://doi.org/10.1002/job.2645 Yin, R. K. (2017). Case Study Research and Applications: Design and Methods. SAGE Publications, Incorporated. 862 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Zhao, H., Wayne, S. J., Glibkowski, B. C., & Bravo, J. (2007 ). The impact of psychological contract breach on work-related outcomes: A meta-analysis. Personnel Psychology, 60(3), 647–680. https://doi.org/10.1111/j.1744-6570.2007.00087.x DOCTORAL CONSORTIUM PRESENTATION OF THE DISPOSITION DRAFT NEJA SAMAR BRENČIČ1,2 1 University of Maribor, Faculty of Organisational sciences, Kranj, Slovenia neja.samar@student.um.si 2 IZRIIS Institute, Ljubljana, Slovenia neja.samar-brencic@izriis.si The Covid19 crisis situation heralded long-term changes in the way health services are provided and accessed. The standards on which these services are based, or should be based, play an important role in ensuring their quality. Telehealth should be regarded as the people centred holistic service not merely as a technological solution. In the thesis I will illuminate how the methods, the procedures and the protocols which form the service of telehealth, are just as important as the technological solution or infrastructure itself. For such understanding, standards of telehealth need to be considered throughout the service development and implementation. In the process the Keywords: telehealth, highest service quality can be achieved if we leverage the support assessment, of properly developed assessment tools, based on specific standards, standards. This paper wil present the disposition draft of the DEXI model, thesis that will propose a model for evaluating services in every Bled phase of their evolution and implementation. eConference DOI https://doi.org/10.18690/um.fov.4.2023.57 ISBN 978-961-286-751-5 864 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Problem definition Standards of telehealth (TH) services are not yet set in national, EU and international strategies (WHO)1. Further on, the stages of implementation, if any, should be investigated and researched. Analysis and anticipated steps toward comprehensive solutions should be proposed. Standards and accreditation processes are a wel known and broadly used quality assurance in all fields of development, production and services. Yet due to the novelty of telehealth and related areas or sub-areas, the standards as well as accreditation programs are being developed alongside the increased need and sporadic implementation of services in daily professional and general population activities. The stil current COVID-19 pandemic and its effects have significantly increased interest in innovative health solutions, specifically also in the field of telehealth services. Key points of the problem (I want to address in my thesis) In the flood of information and new solutions, potential customers, suppliers and end users are wondering which of telehealth services are useful, necessary and tested. What is the level of their quality and suitability to a specific need? Older adults, patients and chronic patients are (potential) major users of telehealth services. Many are particularly vulnerable users, as they can need a greater degree of trust and guidance in choosing and deciding on a particular service. The issues are especial y relevant at this time as we tackle the ongoing issues of the pandemic - as part of which telehealth services are being increasingly considered. The crisis situation thus heralded long-term changes in the way health services are provided at 1 WHO Global strategy on digital health 2020-2025 reference on standards The strategic objective promotes standards for safety, security, privacy, interoperability, and the ethical use of data within and outside the health sector….interoperability standards…requirements for digital health in the design of norms and standards products…global minimum standards for electronic patient health records. etc.; https://www.who.int/docs/default-source/documents/gs4dhdaa2a9f352b0445bafbc79ca799dce4d.pdf N. Samar Brenčič: Presentation of the Disposition Draft 865 or accessed from a distance, namely through telehealth. The standards on which these services are based, or should be based, play an important role in ensuring their quality. There are only a few international and independent standards of services in this area. So, on one side standardisation institutions and user associations cal on stakeholders to co-operate and co-create in the research and development. On the other side, developers and users of telehealth services often do not have a clear understanding of the necessity of standards to ensure service quality. Furthermore, the meaning of the term ‘telehealth service’ itself can carry different orientations or emphases. The knowledge, on the side of the developers, of the complementary perspectives of the technological and service part of the solutions should be understood in order for the final product to be relevant, appropriate and to truly benefit the end user in a long-term sense. Telehealth should be regarded as a service or resource mediated by technology that people use to access or provide health and wel ness-related services, regardless of their location. Of course the technological infrastructure and service compositions wil be discussed in the thesis but the importance of understanding the service as a whole, will be explained in the context of the need for standardisation and accreditation. Furthermore, for telehealth service to be successful, it needs to gain the trust of physicians; health and social care and support providers; as wel as formal and informal carers. Only then will telehealth solutions truly serve a purpose that relates to the overal health of a person. In the thesis I wil il uminate how the methods, the procedures and the protocols which form the service of telehealth, are just as important as the technological solution or infrastructure itself. “Non-optimal” solutions are usually more likely to be related to acceptability, training, availability and, consequently, usability, than to the technological component itself or the possible “complexity” of the technical part. 866 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Technological solutions, which are only a part of the service of telehealth, are thus considered as likely to be successful y implemented only when certain service standards are achieved. The development and implementation of such standards as part of digital health strategies are still in their infancy. In the near future, more attention and focus wil be needed to establish operational procedures that can ensure appropriate quality standards for both the technological part and the services themselves. Only in this way wil the developers and service providers be able to respond appropriately to the increase and changes in needs, demand and user choices. It is recognised that many telehealth services, especial y those that relate to tele- and video-consultation with health staff are now being developed in an emergency. However, the standards within which services operate remain a matter of common interest, possibly heightened in importance for al the stakeholders involved. Telehealth is now no longer just an alternative form of health(care), so the development of standards, certifications and regulations must also be included in national priorities. This will ensure not only technological quality, but equally important quality of service. Following the necessity for standardisation is also the issue of accreditation. More and more medical service providers want to develop their expertise through the accreditation process or gain approval of the quality for their services. 1.1 Structure of the disposition of the doctoral dissertation 1. Problem definition 1.1. Significant research in the field of Telehealth Standardization and Accreditation Programs - International State of the art research and praxis 1.2. Research and praxis in Slovenia 2. Objectives of doctoral dissertation 2.1. Expected original contribution to science 3. Assumptions and potential limitations 4. Intended research methods 5. Intended chapters and subchapters N. Samar Brenčič: Presentation of the Disposition Draft 867 6. Basic literature 7. Analysis of the originality of the topic 7.1. Justification of the originality of the topic 7.2.1. Inquiries on Science citation Index SCI (WoS) 7.2.2. EBSCO (eBook Col ection (EBSCOhost) 7.2.3. EBSCOhost, Academic Search Complete (ERIC) 7.2.4. ProQuest (Interdisciplinary - Doctoral dissertations only) 8. Proposal of a potential mentor 9. Proposal for the scientific discipline of the competent department 10. Professional biography of the candidate 11. Biography of the candidate 11.1. 1.01 Original science article 11.2. 1.08 Published scientific conference paper 11.3. 1.12 Published scientific paper abstract at the conference 11.4. 2.01 Scientific monography 11.5. 2.12 Final report on research results 11.6. 2.14 Project documentation (design concept project, implementation project) APPENDIX 1 Criteria for the quality of Telehealth services 1.2 Significant research in the field of Telehealth Standardization and Accreditation Programs 1.2.1 International State of the art research and praxis As a part of the thesis research I found that the subject is so novel and even undefined to that level that under the same term there are different services in praxis. I decided to start organising the terms so that the demarcation and definition of terms might further serve the standardisation of processes and also contribute to the rise of the quality of services. For that purpose I prepared a paper for the FOV Portorož Conference that was held in March 2023 titled Demarcation and usage of the terms Telehealth, eHealth, Telemedicine and Digital health. I decided to examine the worldwide occurrence of each phrase using the Scopus database in pursuit to add to the definition of the concepts of these services. Other similar terms or sub-terms 868 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY will be a subject of further studies. A total of 95,884 documents contained one of the four terms in the title, abstract or keyword. Telemedicine was the most common term, with 64,149 documents referring to it, followed by e-health by 23,754 and then telehealth with 22,387 and Digital health with 8,599 documents. Articles were the most common type for the four key terms, fol owed by review articles and conference papers . Table 1: Number of documents with telehealth, e-health, telemedicine and digital health and their combinations in title, abstract and keyword for the period 1964 – 2022 Title or Title Abstract Keyword Abstract or Keyword Telehealth 6,048 10,065 17,769 22,387 eHealth 7,310 13,724 15,476 23,754 Telemedicine 12,883 21,735 57,278 64,149 Digital health 2,509 4,619 4,553 8,599 Telehealth or eHealth or 28,771 46,696 78,657 95,884 Telemedicine or Digital health source: Scopus In Table 1 we can see the number of documents referring to each term and their combinations in title, abstract and keyword. The majority of these articles (95%) were in English, (see Table 2) and 10 articles were also written in Slovenian language. In the search al journal categories, al languages and dates of the publication to the exception of the year 2023 for accurate statistics were included. Table 2: The top three languages of the articles retrieved using the four search terms telehealth, e-health telemedicine and digital health in the title, abstract or keyword no. % English 91,601 95 German 1,592 1,6 French 985 1,02 source: Scopus The distribution over time, by document type and by country was provided by Scopus functionalities. N. Samar Brenčič: Presentation of the Disposition Draft 869 Figure 1: Number of documents per year total The findings of the research exposed that the terms telehealth, e-health, telemedicine and digital health are frequently used reciprocal y or interchangeably. I noted that the diversity in the use of the four concepts indicates the vagueness of the concepts and the need for a more precise definition of services and the establishment of certain definitions. The conclusion regarding various usage of terms in the literature was that Telehealth, e-health, telemedicine, and digital health are concepts that have emerged due to advancements in technology in the last two decades and are stil being defined. Different levels of acceptance and use of the four terms points to the ambiguity in definition and in understanding specifics of each concept and the concrete service that is behind that particular concept. Telemedicine being the earliest and most popular term is followed by the term telehealth and e-health, based on the number of publications in the Scopus database. Least used also due to its novelty is the term Digital health. The number of publications is steadily increasing for all of the four terms as it was also foreseen by the authors in the past decade. While these terms are related, they have distinct differences that need to be understood. In order to reach such understanding, the protocols of the service that is behind each term needs to be defined. Only then the process of standardisation can begin as the proper terms wil be used by practitioners, developers and 870 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY financiers. The terms will be defined by the individual parts of the service that are characteristic of it and actually fall under a specific service title. Further research wil be needed in order to prepare a more precise demarcation of services. That might further serve the standardisation of processes and also contribute to the rise of the quality of services. 1.2.2 Research and praxis in Slovenia Slovenia is witnessing a progress in development of telehealth solutions on local and national level. Some overviews have been done in the past 3 years that give us some insight into the situation. In June 2022 I have prepared a brief assessment in regards to standards for Telehelath including Telemedicine, of al listed Telemedicine services. As a basis I used an article by Rant and Rudel (2020) and compared it with online search in June 2022 to verify the services are stil operabile. The results showed that there were four kinds of services available at the time of the assessment, namely (1) national, developed by the NATIONAL INSTITUTE OF PUBLIC HEALTH (National eHealth, Teleradiology, Teletransfusion, Tele-Stroke, Teleconsultations, etc.) (2) national /semi-privat, developed by TELEKOM SLOVENIA, (eHealth, eCare, e-WorkHealth, etc.) (3) national / research developed by The Medical University Ljubljana (Telefarm) and (4) national / private developed by private companies (MKS Cezar, T-Med Gluco, Gospodar Zdravja, etc.), total 21 services provided by 14 entities in Slovenia. Further research is planned for the thesis that wil include literature search and an updated list of all available Telehealth services (Telehealth, Telemedicine, Telecare etc.). 2 Objectives of doctoral dissertation Based on many years of research and practical work and case studies in the international environment, I realised that it is possible to provide better quality of TH services. N. Samar Brenčič: Presentation of the Disposition Draft 871 In the thesis I will be introducing a decision-making model and an appropriate decision-making tool, that will present a possibility for determining suitability of various Telehealth services according to TH standards and suitability according to criteria for specific services. As a part of the research work I wil prepare a unique model of assessment for Telehealth services that wil serve as a tool during the planning of the solution or during the development phase. It wil also serve for possible upgrading of the existing services. The intended outcome of the dissertation is a model for the evaluation of services (according to a certain standard that can be included as a condition) to increase the quality of TH and their standardisation in order to achieve the highest quality possible also for passing the accreditation processes. 2.1 Expected original contribution to science Such a model does not yet exist in the proposed form, according to the best of my research and knowledge and it represents an original innovation in the field of Telehealth. In the disposition of the doctoral dissertation I wil further present the research gaps, assumptions and potential limitations, intended chapters and subchapters, basic literature and the analysis of the originality of the topic. 3 Intended research methods The research method is in the process and wil need further input and fine tuning. Proposed method is the following: 1. Development of the criteria tree with stocks of value with sufficiently clear descriptions. 2. A set of criteria for the DEXI model. 872 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY For example: if the acceptance of telehealth services and how they are evaluated by users is a model example, then the proposal can be from here on for more demanding and advanced evaluations according to standards and pre-accreditation procedures that wil be used by developers. 3. Testing of the model I. (End-user group) Focus groups of 8 - 10 users wil test the services and answer the questionnaire in the laboratory environment. The focus group wil rate the usability of the app, services, etc. according to the DEXI model. I need to further decide if a) The focus group has available set of services for them to choose from or b) The focus group chooses the existing or hypothetical services c) Is the first entry their proposal - e.g. they choose from a set of services or do they enter services themselves. For example Samsung Health App, Telekom eOskrba, Smart watch xx, Monitoring patients Caesar, eHealth - some of these services, e.g. eRecipe Telecap, etc. 4. Testing the model II. (Expert group) We forward the written answers to experts - an expert group, who wil also evaluate based on these opinions. The testing can be organised as a focus group or individual testing. F.e. user group of 3 experts that talk to each other in a controlled environment and give a qualitative assessment or they propose assessment individually. These options will be considered also relating to the availability of experts. Each assessment wil be done only one time for one application. N. Samar Brenčič: Presentation of the Disposition Draft 873 Previously to the testing the questionnaire wil be proposed to the ethical commission for review. In the disposition I wil propose the (1) Planned course of research work and (2) Display of the basic elements of the assessment model. 4 Proposal of a potential mentor Proposed mentor for the thesis is the associate professor dr. Uroš Rajkovič, University of Mariboru, Faculty for organisational sciences, Kidričeva cesta 55a, Kranj. 4.1 Proposal of a potential co-mentor Proposed co-mentor is professor dr. Malcolm Fisk, De Montfort University, Faculty for Computing, Engineering and Media, School of Computer Science and Informatics, The Gateway, Leicester, LE1 9BH. 5 Proposal for the scientific discipline of the competent department A proposal for a scientific discipline: Organization and Management of Information Systems. Proposed department: Department of Informatics Acknowledgements Many thanks to my coleague dr. Drago Rudel and mentors prof. dr. Malcolm Fisk and prof. dr. Uroš Rajkovič for indispensable advice and guidance. References Fatehi F., Wootton R. (2012). Telemedicine, telehealth or e-health? A bibliometric analysis of the trends in the use of these terms. J Telemed Telecare. 18(8), 460-4. Samar Brencic N., Rudel D. (2021). Can a self-assessment tool help raise standards of eHealth and telemedicine services that are rapidly developing under COVID-19 emergency? JTTA journal Vol.16 suppl. Samar Brencic N., Rudel D., Fisk M. (2021). The importance of standards for quality Telemedicine and eHealth services in COVID-19 pandemics, JTTA journal Vol.16 suppl. 874 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Samar Brencic N., Rudel D. (2020). Standards of quality for telehealth services for older adults - their augmented significance in Covid-19 pandemics; 22th International multi-conference information society, Conference proceedings, 393 - 395. Samar Brencic N., Rudel D. (2020). Pomen standardov za kakovostne storitve zdravja na daljavo (v času COVID-19 pandemije), Digitalni mostovi v zdravstvu : e-Kongres MI'2020 : zbornik prispevkov in povzetkov SDMI, ur. Tomaž Marčun, Ema Dornik Dostop do e-publikacije: https://sdmi.si/files/strokovna_srecanja/zbornik%20MI2020.pdf Samar Brencic N. et al., Intuitive and intelligent solutions for elderly care, 2020. V Chaari, Lotfi ed. Proceedings of the 2nd International Conference on Digital Health technologies, Advances in Predictive, Preventive and Personalised Medicine Series 12, Springer International Publishing, DOI 10.1007/978-3-030-49815-3. Rudel, D., Fisk M. (2012). Telescope – telehealth services code of practice for Europe. Inform Med Slov; 17(1), 38-44. Rudel, D., Fisk M. (2011). Definitions of Terms in Telehealth. Infor Med Slov, 16(1), 28-46. http://ims.mf.uni-lj.si/archive/16%281%29/21.pdf (Accessible 2011-10-12). Rudel, D., Fisk M. (2018).Telehealth Quality Group EEIG, International Code of Practice for Telehealth Services, Ljubljana. Thiel T., Deimel L., Schmidtmann D., Piesche K., Hüsing T., Rennoch J., Stroetmann V., Stroetmann K. (2018). SmartHealthSystems: International Comparison of Digital Strategies’, Empirica, Bertlesmann Stiftung. DOCTORAL CONSORTIUM A QUANTITATIVE EXPERIMENT: INTER-TEAM RETROSPECTIVES IN SCALED AGILE PROGRAMS DANIEL TOEGL,1 TIM HUYGH,2 STEVEN DE HAES3 1 University of Antwerp, Antwerp, Austria daniel@toegl.at 2 Open University of The Netherlands, Department of Information Science, Heerlen, Netherland tim.huygh@ou.nl 3 University of Antwerp, Department of Management Information Systems, Antwerp, Belgium steven.dehaes@uantwerpen.be Despite the current popularity agile project management and scaling such frameworks, there is little academic research inquiring how to scale agile frameworks to be successful at the program level. Retrospectives, are understood to play an important role in attempts to successfully adjust scaled-up project frameworks. This research studies the particular impact of such meetings as a tool to scale a known success factor from agile projects and address current gaps highlighted in recent program management literature. To better understand the impact of retrospectives, a quantitative experiment, to inquire how program value is accomplished with inter-team retrospectives, is Keywords: conducted. With the proposed research design, we intend to test program management, if there is a measurable impact on program success and efficiency inter-team, by introducing inter-team retrospectives based on current intra-team, approaches from the literature. This will allow to better scaled agile, understand and further develop the retrospective meeting design software on an inter-team level within scale agile programs. development DOI https://doi.org/10.18690/um.fov.4.2023.58 ISBN 978-961-286-751-5 876 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Since the introduction of agile software development with Scrum (Schwaber and Sutherland, 2020), agile project management1 is receiving increasing popularity. With growing numbers of applied use cases across industries, interest of scaling such frameworks is increasing to benefit from its advantages on a larger scale. However, there is currently little academic research on agile program and portfolio management to be found in the peer-reviewed literature. In small agile projects, the development team coordinates work through frequent informal interaction among themselves and with customers. The Scrum framework has dedicated meetings for planning, review, and retrospectives. Many teams use visual boards, like in Kanban, to show who is working on what and the status of work tasks, which are designed for project size endeavors. Strode et al. (2012) explain coordination in such environments with agile teams and propose a model for coordination strategy and coordination effectiveness. For large-scale projects, there is less support in the academic literature as described by Dingsoyr et al. (2018b). Scrum prescribes regular meetings between Scrum teams ("Scrum of Scrums") in order to manage the interfaces between teams, covering the planning of upcoming activities and dependencies, not retrospectives. As we see from this current body of knowledge, focus is primarily on upfront planning activities at the beginning of each iteration (i.e. "Sprint Planning"). Agile however is consider as a series of cycles that aim and al ow for continuous improvement for which reason the concept of a retrospective was introduced on a project level. Inter-team retrospectives are currently not part of commonly referred to scaled agile frameworks (Toegl et al., 2023), despite the original characterization from Schwaber and Sutherland (2020, page 9): " The purpose of the Sprint Retrospective is to plan ways to increase quality and ef ectiveness". To close these identified gaps and extend the range of research methodologies for agile program management, a quantitative experiment is conducted to assess if an inter-team retrospective is a scalabale success/factor for agile program management. 1 Several definitions of project, program, portfolio and the respective management can be found in management literature. In this review, we are referring to the denitions stated by Ireland (2002). D. Toegl, T. Huygh, S. de Haes: A Quantitative Experiment: Inter-Team Retrospectives in Scaled Agile Programs 877 The result of this study would significantly support academics as wel as practitioners with further developing scaled agile frameworks, their key artifacts as wel as help organizations to become more successful in their agile transformation and Change initiatives. 2 Problem definition From a business perspective, agile projects have been very successful across organizations in particular for software development. To further apply this methodology on a larger scale, agile project frameworks need to be adjusted in order to be successful y scaled-up and bank in its benefits. While different Scaled Agile frameworks include different approaches on scheduled and unscheduled inter-team communication, the current literature investigates specific case studies as done by Paasivaara et al. (2012) or Dingsoyr et al. (2018b). Dingsoyr therefor suggest to further explore this relationship between intra-team and inter-team coordination in other approaches and settings other than his case study. From Toegl et al. (2023) we know that the academic literature covers success factors of scaled agile programs but does not investigate how to actual y scale success factors known from agile projects, such as retrospectives. Communication as wel as knowledge sharing and improvement to enable learning at inter-team levels and program levels show under-researched fields (Dingsoyr et al., 2018a). Understanding the impact of retrospectives as a tool of knowledge sharing and improvement as well as for inter-team communication as wel remain to be researched as described in the literature by Stettina and Schoemaker (2018) and Dingsoyr et al. (2018c). As highlighted by Dingsoyr et al. (2018c),the central chal enge in coordination is identifying the right form or artifacts, arenas, and degree of formalization in large endeavors, such as scaled agile programs, with high complexity - a typical indicator to apply an agile framework as stated by Salameh (2014). As the literature reveals, e.g. the systematic literature review by Dikert et al. (2016), agile program management is limited to experience reports and case studies of already completed programs and projects given the practicability of retrieving data. 878 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Others applied approaches are limited to exploratory studies such as Dingsoyr et al. (2018c). Table 2.1 summarizes the questions for this study: Table 2.1: Research questions ID Research question What is the effect of inter-team retrospectives on program/business output PS1 within an agile setting? What is the effet of inter-team retrospectives on program efficiency within PS2 an agile setting? What are advantages and disadvantages of the different formats of inter- PS3 team retrospectives? How do formats of inter-team retrospectives evolve over time and PS3.1 iterations? To find answers to these questions, close the identified gaps and extend the range of research methodologies for agile program management, a quantitative experiment is conducted to assess if an inter-team retrospective is a scalabale success/factor for agile program management. The result of this study would significantly support academics as wel as practitioners with further developing scaled agile frameworks, their key artifacts as wel as help organizations to become more successful in their agile transformation and Change initiatives. It is the contention that inter-team retrospectives in agile programs have a direct impact on program success and evaluated by efficiency, stakeholder satisfaction, and the success of meeting wider business goals, the relationship is also subject to other intervening, moderator impacts that must be considered as part of this study. We will follow the approach of Serrador and Pinto (2015) to achieve consistent results as shown by their work in this field. As a result, the proposed research model is shown in Fig. 2.1, highlighting not only the tested relationship between inter-team communication in agile programs and program success, but also the potential impact of various moderators on this direct effect. D. Toegl, T. Huygh, S. de Haes: A Quantitative Experiment: Inter-Team Retrospectives in Scaled Agile Programs 879 Figure 2.1: Research model The type of inter-team retrospective type needs to be considered as a moderator because the time invested in such meetings has a significant impact on time spent and budget given the potentially high number of involved stakeholders and hence influencing project success and efficiency and subsequently program success and efficiency. However, in this research primarily focus on the impact created by the presence of such retrospectives. Table 3.3 provides an overview of retrospective formats considered in this study. The same assumption can be made on learning effects on managing defects to reduce error. The central hypotheses (PS1, PS2) of this research are therefore about testing the effect of inter-team retrospectives as wel as starting the investigation which format of inter-team retrospectives (Table 3.3) shows most effect. Table 2.2 provides an overview of the hypotheses and moderating effects that shall be investigated as part of this research. Table 2.2: Hypotheses Hypotheses HP1 Inter-team retrospectives have a positive impact on proram success. HP2 Inter-team retrospectives have a positive impact on program efficiency. The type/format of inter-team retrospective moderates the effect of HP3 retrospectives on program success. The type/format of inter-team retrospective moderates the effect of HP4 retrospectives on program efficiency. 880 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY As the impact of the inter-team retrospective is expected to become more visible over time, meaning after a couple iterations, we want to assess the impact of the inter-team retrospective as a moderator. Following the approach of Treiblmaier and Putz (2020) in which a comparable multi-group field experiment was facilitated, we model inter-team retrospectives as a moderator between the program and project success as wel as project efficiency and therefore program success and efficiency and hypothesize for the statistical testing: Table 2.3: Nul hypotheses for testing program success Null hypotheses There is no effect on program success in an environment with inter-team H0 retrospectives. There is an effect on program success in an environment with inter-team H1 retrospectives (of format A or B). Table 2.4: Nul hypotheses for testing program efficiency Nul hypotheses There is no effect on efficiency in an environment with inter-team H0 retrospectives. There is an effect on efficiency in an environment with inter-team H1 retrospectives (of format A or B). 3 Methodology, design and focus We aim to research the impact of inter-team retrospectives within a simulated setting and compare it to the same setting without inter-team retrospectives. We consider program management success as the unit of analysis, following the approach of Shao et al. (2012). A post-positivism perspective with a deductive approach is used in this study in order to operationalize the program success and program context concepts - broadly following Shao et al. (2012). An experiment is used to col ect data which then al ows for generalizable results within the design's boundaries. The applied research design is therefore considered to be Experimental Research. D. Toegl, T. Huygh, S. de Haes: A Quantitative Experiment: Inter-Team Retrospectives in Scaled Agile Programs 881 Program business success shal be the output of the project team. Program efficiency shall be considered as reduction of error (defects). Stakeholder satisfaction to satisfy the expectations of project stakeholders is not considered to simplify the experiment. We gather qualitative data from the participants in regards to PS3 via their results from the retrospectives to better understand how the adjusted intra-team as wel as the inter-team retrospectives were perceived. To analyze data from these retrospectives, the success factors described by Hummel and Epp (2015) are applied for potential coding purposes. 3.1 Iterative research approach and Pre-studies Following an agile approach, two pre-studies were conducted with graduate students from the University of Applied Sciences Joanneum (Graz, Austria) in autumn 2021 and 2022 to confirm the feasibility of the operating the experiment. The goal of these pre-studies was to test assumptions made and to test the operationalization of the experiment in terms of its complexity. Findings of these pre-studies impacted the ratio of required Scrum Masters to teams (1:1), highlighted the need to play different games within a program to avoid competition (i.e. withholding information as wel as mistrust between teams). In regard to the number of iterations conducted during the experiments, we raised the number of iterations. 3.2 Procedure and data col ection Data is collected during the experiment, as described in Table 3.1 and Table 3.2, by the Scrum Master typical y every sprint or are calculated based on values provided from each project. 882 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 3.1: Measures on program level Dimension Calculation of scores Total program Sum of each team’s actual points delivered across all performance sprints. Sum of each team’s actual points delivered across all Total program defects sprints. Ratio of the total program performance to the total Total program efficiency program defects. Program sprint Sum of each team's actual points delivered in a sprint. performance Program sprint defects Sum of each team's defects in a sprint. Ratio of the program sprint performance to the Program sprint efficiency. program sprint defects. Table 3.2: Measures on project level Dimension Calculation of scores Sprint performance A team's actual points delivered in a sprint. Sprint defects Defects that occurred during a team’s sprint. Further calculations of each team’s total performance/efficiency wil be conducted. Inter-team retrospectives are considered as indicator variables (1 = took place; 0 = did not take place) for each program, team and sprint. In this experiment, we are simulating programs following the Scrum of Scrum framework (Paasivaara et al., 2012), which is chosen given its lowest level of complexity once the concept of Scrum is introduced. We have multiple Scrum teams with a Scrum Master each, who will be responsible to collect the data. Other data is calculated based on their output. 3.3 Structure of the experiment The experiment aims to simulate scale agile programs by playing agile games. To ensure comparable results, every program follows the same strategic goals. D. Toegl, T. Huygh, S. de Haes: A Quantitative Experiment: Inter-Team Retrospectives in Scaled Agile Programs 883 The Bal Flow Game was chosen as a feasible game for the Scrum teams (Fowler et al., 2001) in the experiment given the comparison by Przybylek and Olszewski (2016) in which it scored high in the considered factors, such as easy-to-understand. In addition, its scoring mechanism in every iteration of the game made it a suitable candidate for this research. The class-room activities described by Hurbungs and Nagowah (2019) Paper Planes (Heintz) and Pizza Game (Agile42) were chosen for the Scrum teams due to their suitability in a tertiary educational context as well as their objective to agile learning and teaching. In addition, we can apply the same scoring mechanism in every iteration of the games. For this experiment, we time-box a complete sprint/iteration to 15 minutes. During these 15 minutes, every team has to complete al Scrum ceremonies. The actual play time of each game is defined as 3 minutes. The Scrum Master is responsible to facilitate the game, stop the time and collect the results. Adoption of Scrum-of-Scrum questions We can adapt questions used by Paasivaara et al. (2012) in the Daily Scrum to the needs of the retrospective, along with (Dingsoyr et al., 2018a) on the category "other teams to facilitate "double loop" learning from the framework of Argyris (1976). Participants Participants are university students with no to little prior knowledge about agile project management. To further understand their knowledge, we perform a survey at the beginning of the experiment and dismiss students that have prior experience with the selected agile games. As part of the experiment, a set of roles is required that fol ow the Scrum (Fowler et al., 2001) and Scrum of Scrum frameworks (Paasivaara et al., 2012). In this set-up, Release Manager and Product Manager can be neglected. The role of the Product Owner is replaced with a rule set and guidelines for each sprint. This information is available to the Scrum Masters from the beginning and during each iteration wil be revealed to the project team. As part of this the Scrum Masters wil receive separate 884 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY information and is instructed to perform additional tasks, such as disrupting the process with adding three bal s during a sprint as part of the Bal Point Game. 3.4 Types/formats of inter-team retrospectives The different types or formats of inter-team retrospectives considered in this research are described in Table 3.3. Inter-team retrospectives are planned only for teams or representatives from the same program and not across multiple programs. Table 3.3: Considered formats of inter-team retrospectives Type Short description Inter-team retrospectives with by the team selected members of each A Scrum team. B Inter-team retrospectives between the Scrum Masters of each Scrum team. C Inter-team retrospectives with all Scrum teams and all their members. D No inter-team retrospective. We plan for team retrospectives after every sprint/iteration. Team retrospectives are planned once every team completed its team retrospective. Type C is only mentioned for completeness and are not considered suitable for any real-world program. It is therefore not further considered in this experiment. Sequence of sprints/iterations For this experiment, we plan to have nine sprints/iterations for each team in every program executed in paral el. Data is col ected before, during and after each sprint/iteration. 3.5 Data col ection approach Data is col ected to col ect each team's output and defects and therefore its contribution to the program. We expect, this al ows to understand the performance on a program as wel as project level. In general, the simplest relationships are examined firsthand then analysis continued using progressively more involved techniques. This includes a path-analysis of each scrum team over time. D. Toegl, T. Huygh, S. de Haes: A Quantitative Experiment: Inter-Team Retrospectives in Scaled Agile Programs 885 Other potential moderators such as Product Vision/Goals, Project Complexity and Team Experience are not further considered as these factors are harmonized in al our programs within the given experimental design. 3.6 Statistical evaluation The experiment provides a purposeful sample size with 30 inter-team retrospectives (10 for each program) with a total of 90 sprints (30 for each program), which is deemed sufficient for the statistical analysis with a regular t-test, as described by De Winter (2013). If required, the experiment can be scaled up include more teams in each program, however saturation is expected by 10 sprints. If we cannot confirm normal distribution with the gained data, the Mann-Whitney-U test can be used. The Jonckheere's trend test could additional y be used to detect differences in the teams across the multiple sprints. In particular, we would expect to see a trend when comparing the project teams with inter-team retrospectives to the comparison project teams without inter-team retrospective over time. 3.7 Data quality The data wil be col ected every sprint from the Scrum Masters and is limited to only very few input points. Photos of the retrospective boards with their sticky notes are taken in every sprint to collect the required data, which will be particularly relevant regarding PS3. Photos al ow to track and validate not only the raised points, but additional y al ow to count how often certain topics are raised within a retrospective. A retrospective board also al ows to visual y see where the majority of points are raised which al ows to easily identify if things went wel or not. As described by Schulze (2007), taking photos forces the participants to reect, while gaining insights on what is most significant. This simplicity is expected to provide sufficient meaningful data for every Scrum team in every iteration, compared to other media such as, e.g. video. To ensure data is properly col ected, the Scrum Masters wil receive an introductory training in advance. All Scrum Masters will be PSM1 certified. 886 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 3.8 Validity and reliability Following the structure and results of Brink (1993), the four threats (the researcher, the subjects participating, the situation and the methods of data col ection and analysis) are addressed. To reduce bias of the researcher by participating as wel as the risk of reducing bias with data collection, the Scrum Masters wil be responsible for data collection. They will receive an introductory training in advance, together with a retrospective guide and a fieldnote template that covers the required structure to cover and col ect data of the required key success factors. Such a preparation is considered especial y necessary to ensure the researcher is trusted by the Scrum Masters (Leininger and Reynolds, 1991). Data will be collected every sprint and is limited to only very few input points, to ensure the amount of work is minimized. Given the simplicity of the game and time constraints, the depth and variety of findings in retrospective meetings are expected to diverge little among teams when comparing to real-world large scale programs. 3.9 Ethics Approval of the Ethical board of the University of Antwerp is expected before conducting the experiment. The idea is to perform this primarily with adults, that are monetarily incentivized (e.g. EUR 1 / point) for each successfully delivered point. 4 Preliminary/Expected results In terms of reserach methodology, researching inter-team retrospective with an experiment expands the toolbox for conducting research in the eld of agile program management. D. Toegl, T. Huygh, S. de Haes: A Quantitative Experiment: Inter-Team Retrospectives in Scaled Agile Programs 887 Regarding the results of the experiment, the initial contention is that the simulated programs with inter-team retrospectives are more successful than those neglecting this ceremony. We expect that the nul hypothesis described in Table 2.4 and in Table 2.3 are rejected. This means, a possible outcome could be, that we can show a statistically significant impact of inter-team retrospectives in terms of the business outcomes (program success and program efficiency), while identifying a better performing retrospective format. We additional y expect to gain knowledge how intra-team and inter-team retrospectives evolve over time. We expect to further gain knowledge on who become ambassadors to represent teams at inter-team retrospectives and how the team members utilize inter-team retrospectives to their project's benefit. With this result, we believe found a success factor known from agile project management, that can be scaled to agile program management. These results will allow us to undergo real-world tests with this format with lower managerial risks as the proposed concept resulted from thorough academic research. This knowledge could impact scaled agile frameworks in a way, that inter-team retrospectives become widely adopted. Expected outcomes Type A and B retrospectives are expected to be the types with the most positive effect among the researched types. Additional y, as the iterations continue and changes in the set-up might be suggested by the intra-team as well as as the inter-team retrospectives, type A retrospectives might evolve into a type B retrospective, or alternatively, type B retrospectives might evolve into a type A retrospective. 5 Future development This research is part of a broader research question part of the author's PhD. Tying three research studies together, a possible combined outcome could be, that we contributed to the body of knowledge in multiple ways and lay out a way on how to improve success of scaled agile programs. 888 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Firstly, we identified an under-researched academic eld with high interest for practitioners and prepared an overview that supports further academic research more easily. The results of a systematic literature revealed that retrospectives are a known success factor in agile projects but academia barely researched on how to scale up this success factor to scaled agile programs. The here described experiment is planned as the second paper to solve the "managerial puzzle" described in the author's PhD project. Building on the results from this research paper, we want to gain in-depth knowledge on the perceived value of inter-team retrospectives, their connection to intra-team retrospectives and their evolution over time at a software service providing company, Parkside Interactive, in a third study. The best performing format of inter-team retrospective from the experiment wil be taken as a starting point for this case study. With these steps, we additionally close a number of previously identified research gaps and for the fist time analyze agile programs with a quantitative experiment. References Agile42. Pizza game. URL https://www.agile42.com/en/agile-teams/kanban-pizza-game. C. Argyris. Single-loop and double-loop models in research on decision making. Administrative science quarterly, pages 363-375, 1976. H. I. Brink. Validity and reliability in qualitative research. Curationis, 16(2):35-38, 1993. J. C. De Winter. Using the student's t-test with extremely smal sample sizes. Practical Assessment, Research, and Evaluation, 18(1):10, 2013. K. Dikert, M. Paasivaara, and C. Lassenius. Chal enges and success factors for large-scale agile transformations: A systematic literature review. Journal of Systems and Software, 119:87-108, 2016. T. Dingsoyr, M. Mikalsen, A. Solem, and K. Vestues. Learning in the large - an exploratory study of retrospectives in large-scale agile development. In International Conference on Agile Software Development, pages 191-198. Springer, Cham, 2018a. T. Dingsoyr, N. B. Moe, T. E. Faegri, and E. A. Seim. Exploring software development at the very large-scale: a revelatory case study and research agenda for agile method adaptation. Empirical Software Engineering, 23(1):490-520, 2018b. T. Dingsoyr, N. B. Moe, and E. A. Seim. Coordinating knowledge work in multiteam programs: Findings from a large-scale agile development program. Project Management Journal, 49(6): 64-77, 2018c. M. Fowler, J. Highsmith, et al. The agile manifesto. Software Development, 9(8):28-35, 2001. J. Heintz. Agile airplane game. URL http://gistlabs.com/2011/06/agile-airplane-game/. M. Hummel and A. Epp. Success factors of agile information systems development: A qualitative study. In 2015 48th Hawaii International Conference on System Sciences, pages 5045-5054. D. Toegl, T. Huygh, S. de Haes: A Quantitative Experiment: Inter-Team Retrospectives in Scaled Agile Programs 889 IEEE, 2015. V. Hurbungs and S. D. Nagowah. A practical approach to teaching agile methodologies and principles at tertiary level using student-centred activities. Agile and Lean Concepts for Teaching and Learning: Bringing Methodologies from Industry to the Classroom, pages 355-389, 2019. L. R. Ireland. PMI Lexicon of Project Management Terms. Project Management Institute, Inc., 2002. M. Leininger and C. Reynolds. Culture care diversity and universality theory. Nursing, 1991. M. Paasivaara, C. Lassenius, and V. T. Heikkila. Inter-team coordination in large-scale globally distributed scrum: Do scrum-of-scrums real y work? In Proceedings of the ACM-IEEE international symposium on Empirical software engineering and measurement, pages 235-238, 2012. A. Przybylek and M. K. Olszewski. Adopting col aborative games into open kanban. In 2016 Federated Conference on Computer Science and Information Systems (FedCSIS), pages 1539-1543. IEEE, 2016 H. Salameh. What, when, why, and how? a comparison between agile project management and traditional project management methods. International Journal of Business and Management Review, 2(5):52-74, 2014. S. Schulze. The usefulness of reflexive photography for qualitative research: A case study in higher education. South African journal of higher education, 21(5):536-553, 2007. 14 K. Schwaber and J. Sutherland. The scrum guide. Scrum Alliance, 2020. P. Serrador and J. K. Pinto. Does agile work? A quantitative analysis of agile project success. International Journal of Project Management, 33(5):1040-1051, 2015. J. Shao, R. Muel er, and J. R. Turner. Measuring program success. Project Management Journal, 43(1):37-49, 2012. C. J. Stettina and L. Schoemaker. Reporting in agile portfolio management: routines, metrics and artefacts to maintain an effective oversight. In International Conference on Agile Software Development, pages 199215. Springer, Cham, 2018. D. E. Strode, S. L. Huff, B. Hope, and S. Link. Coordination in co-located agile software development projects. Journal of Systems and Software, 85(6):1222-1238, 2012. D. Toegl, T. Huygh, and S. De Haes. Retrospectives in scaled agile: A systematic literature review. Working paper, 2023. H. Treiblmaier and L.-M. Putz. Gamification as a moderator for the impact of intrinsic motivation: Findings from a multigroup field experiment. Learning and Motivation, 71:101655, 2020. 890 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY DOCTORAL CONSORTIUM THE MULTI-CRITERIA DIGITAL AND SUSTAINABLE MATURITY ASSESSMENT MODEL KATJA MOHAR BASTAR University of Maribor, Faculty of Organizational Sciences, Kranj, Slovenia katja.mohar@student.um.si European Union (EU) recognized the strong need for digital and sustainable transition of enterprises and societies. Although digital technologies can contribute to achieving sustainable goals, many of enterprises still struggle with challenges of digitalization and digital transformation, especially SMEs, which due to limited resources often lag behind larger enterprises. In the paper, we focus on SMEs, since they represent majority of EU economy, and are strong contributor to the number jobs and GDP. To achieve goals for successful digital and sustainable transition, which are adopted by European Commission, it is important to use proper strategies. This is not possible without understanding the current state and predict the impact of given measures and develop new business models that can be derived from digital Keywords: transformation, sustainability, and circular economy. There are SMEs, several available tools for digital maturity assessment available regulation, and SMEs can estimate their position in digital journey. sustainability, digital However, to our knowledge, no tool, which would enable transformation, assessment of digital maturity and its impact on achieving twin sustainability goals, is currently available. Such a model could transition, maturity help SMEs to assess current state and plan for proper strategy to assessment achieve digital transformation towards sustainability goals. model DOI https://doi.org/10.18690/um.fov.4.2023.59 ISBN 978-961-286-751-5 892 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction Many crises in Europe and worldwide, caused by Covid-19, war in Ukraine, population growth and environmental exploitation, resulting shortage of natural sources and natural disasters, have led the European Commission to prepare the strategy for twin green and transition. Both strategies- EU digital strategy and green deal, are heading to the same goal- to make digital strategy work for people, while helping to achieve a target of a climate-neutral Europe by 2050 (European Commission, 2020c). Digital regulatory framework is combined under EU digital strategy and consists of several regulations and directives (European Commission, 2020c). Twin transition is one of the most important paradigms in the current time. Irreversible changes and further damage on the planet Earth can only be avoided with greater responsibility from each individuum, adjustments in business with the changes in business models and adopted legislative (Vidmar, 2021). EU framework for the twin transition aims to achieve sustainability, combat climate change and environmental degradation with harnessing digital technologies for sustainability and prosperity, and to empower citizens and business (Muench et al., 2022). Green transition, addressing environmental performance (planet) is only one of the elements in the triple bottom line of sustainability, besides economic performance (profitability) and social performance (people), (Dao et al., 2011) which al need to be addressed in the forthcoming economy. The transformation from pure economic success in linear economy to shared value which is including not only individual (economic) success, but the success of the society is the basis for the sustainable or circular economy (Brenner, 2018). The area is currently regulated by the Directive as regards disclosure of non-financial and diversity information by certain large undertakings and groups (NFRD directive) from 2014, which only affects large enterprises from 2018 and anticipates reporting in the annual reports on five dimensions and four sub-dimensions in pdf format with low reliability of reporting (European Commission, 2014). K. Mohar Bastar: The Multi-Criteria Digital and Sustainable Maturity Assessment Model 893 Under the EU Green Deal (European Commission, 2019) new directives and new standards are under preparation, also obliging SMEs for preparation of standardized reports, including financial taxonomy. In the paper, we wil focus on SMEs, which represent 99 % of European economy. Unfortunately they are lagging in adopting new development principles both in the digital transformation and consequently sustainability (Gorgels et al., 2022). 2 Problem definition The problem addressed in our research is focused to the twin transition of SMEs. SMEs represent 99% of the European economy; they provide 100 million of jobs and contribute 50% of Europe’s GDP. (Gorgels et al., 2022). At the same time, only 25% of EU SMEs work on green products or services and only 17% have successfully integrated digital technology. SMEs contribute about 65% of overall employment in the economy, but they are also responsible for around 60% of al greenhouse gas emissions by enterprises(European Commission, 2020a). New regulatory framework is setting rules for SMEs for mandatory reporting and taking actions towards the twin transition. The digital and sustainable maturity assessment model can help them with easier adopting to the regulations but also help them design new strategies for future development of their enterprises. We can observe that many enterprises are facing chal enges with establishing digital capabilities, competences, and changes of the organizational culture (Pucihar, 2020). With the emerged need for sustainable transformation of the world around us, and the simultaneous ongoing digital transformation, it is necessary for enterprises to find parallels in development of their business models. Digitalization is giving us the possibility to re-create the environment we live in, even though the population is rising, especially in the cities. In the cities, the impact of digital technologies is already recognized through the sharing economy business models (Hildebrandt et al., 2018). Dealing with the twin transition, digital transformation, and the innovation of digital business models already guided entrepreneurs towards pro-growth mentality and openness to changes (Ferreira et al., 2022). Recent research shows that sustainability and digitalization can integrate principles and strategies(Vidmar, 2021). The research investigated which information technology, organizational and business environment factors influence changes in business models, with the focus on its 894 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY digitalization. The research also focused on the impact of business model changes on the three components of sustainable performance: economic, environmental, and social. European Commission can influence European market by setting rules, that is why the twin transition(Muench et al., 2022) with the digital strategy(European Commission, 2020c) and European green deal (European Commission, 2019) targeted the same outcome – to create a climate neutral Europe by 2050 with including various factors of everyday life(Muench et al., 2022). EU Green deal includes different goals and therefore reporting standards about achieving those goals. Main goals are dedicated to environment, Europe to be the first climate-neutral continent by 2050, with at least 55% less net greenhouse gas emissions by 2030, compared to 1990 levels and planting additional 3 billion of trees in EU. EU green deal wil represent a series of benefits, fresh air, clean water, healthy soil, and biodiversity, renovated, energy efficient buildings, healthy and affordable food, more public transport, cleaner energy and cutting edge clean technological innovation, longer lasting products that can be repaired, recycled, and reused, future-proof jobs and skil s training for the transition with global y competitive and resilient industry(European Commission, 2019). Goals and expected results are affecting the economy, which wil have to transform to comply with the demands and to ensure the sustainable products. At the same time, the directives CSRD (European Commission, 2023), EU Taxonomy (European Commission, 2020b), ESRS standards(EFRAG, 2021) and CSDD directive(European Commission, 2022), will significantly influence the operations of the enterprises. Some of the key characteristics of SMEs are limited financial and human resources, operation in economic or geographic niches and in uncertain markets and policy environments, which al represent chal enges for the sustainable transition (Gorgels et al., 2022). SMEs also meets difficulties in influencing wider business environment and they are often owned and managed by the same person, on whom depend on all the values and beliefs(Kljajić et al., 2021a). SMEs were forced to proceed with the digital transformation in the last years, because of their competition, but also Covid19 pandemic push. Both transformations, digital and sustainable, need to start from the current state of the individual enterprise, which we can also state as maturity. From there, enterprises can plan strategies and further development. K. Mohar Bastar: The Multi-Criteria Digital and Sustainable Maturity Assessment Model 895 The maturity can be defined as “the state of being complete, perfect or ready” (Lasrado et al., 2015), the assessing the maturity of digitalization reflects the degree of digital transformation in the company (Alsufyani & Gill, 2021), which represents long and uninterrupted process, covering technology, culture, company strategies, staff, and end user needs (Nasiri et al., 2022). Our research emerges from the limited abilities of SMEs to design their own strategies for twin transition. In Slovenia, national digital maturity assessment model was developed in 2020 and it is under constant improvements. This assessment model enables individual SME to assess its current state in the digital transformation journey and compares themselves with its competition. In addition, the tool also provides automatical y generated report with recommendations for further development (Kljajić Borštnar & Pucihar, 2021). National digital maturity assessment model is a result of many studies and expert group involvement, and it gains from outcomes of already existing digital maturity models in other countries. The aim of our research is to extend the existing tool with sustainability criteria and adjusting existing digital and organizational readiness assessment criteria will provide multi-criteria digital and sustainable maturity assessment tool. SMEs will use the measured score for orientation on their position on the market as wel for the planning of future investments to gain compliance and benefits from twin transition to create future strategies connected to their needs for digitalization and green transition goals. The national level of collecting data enables policy makers to customize proper instruments to help SMEs, based on aggregated and analysed data. 3 Methodology As the umbrel a methodology for the proposed research, we follow the design science research methodology (DSR) , with the main goal the development of organization- informatics artefact to solve the relevant business needs, based on the existing knowledge (Hevner et al., 2004). The proposed solution – artefact- will be multi-criteria model for the assessment of the digital and sustainability maturity. Based on the scientific and expert literature review, existing models, policy literature and research, we wil conduct the conceptual research model and determine the initial set of criteria for the assessment of digitalization and sustainability maturity. In the next step we wil determine value ranges and validate the criteria set through 896 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY the semi- structured interviews with experts on the digitalization and sustainability fields. Semi-structured interview provides the additional questions, which were not planned, but they can significantly contribute to the model development and understanding of the problem (Kallio et al., 2016). The model we will develop will be the multi attribute decision model, which are considered as useful decision-making tools in the complex situations with evaluation process with including al the relevant factors, which can affect the decision (Kljajič et al., 2021b). We will develop multi-criteria assessment model and will test it on the 6-10 SMEs to confirm its sensibility for differences among enterprises. In the development phase of multi-attribute decision making model, we will use the DEXi program – a software which is freely available. DEXi method requires the decomposition of the decision problem into smal er problems with lower complexity and easier to solve (Bohanec, 2020). In the last step we wil evaluate the model and analyse the data from SMEs, with which we wil confirm the suitability of the further usage of the model in practice. One of the goals is test of the model in practice to achieve opinions and recommendations for the improvements of the model. 3.1 Literature review Initial y we wanted to establish the time when the topics of digitalization and sustainability became of more importance, regarding the number of published sources. Through the knowledge bases Web of Science, Scopus and ProQuest Dissertations and Thesis we’ve searched for the key words “digitalization” and “sustainability”, col ected the results, and presented the median through the Figures 1 and 2, presenting the number of articles with included keyword digitalization (Figure 1) and sustainability (Figure 2). Because we assume that some of the articles are the same, published in different knowledge base, we did not calculate the sum, but took the median. K. Mohar Bastar: The Multi-Criteria Digital and Sustainable Maturity Assessment Model 897 We assume that the number of articles as wel represent the importance of the topics. In the Figure 1 and Figure 2 we can see that both topics became relevant in early 19th Century, but raising relevance in the last 10-15 years, while now they represent one of the most important topics. 8000 "Digitalization" 6000 4000 2000 0 1922 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 2015 2018 2021 No of results Figure 1: Number of articles on the keyword “digitalization” "Sustainability" 60000 40000 20000 0 1970 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 2015 2018 2021 No of results Figure 2: Number of articles on the keyword “sustainability” In the next step, we used three knowledge bases: Scopus, Web of Science and ProQuest Dissertations and Thesis, to search by the key words listed below. We were adding the concepts, applied to the research problem, to narrow the selection. In this process, we concluded that for the research problem – the design of multi-criteria model for digital and sustainable maturity assessment in SMEs - there is no articles in knowledge bases, we searched. In the Scopus and Web of Science there is 898 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY only one article, which is dealing with twin transition model for SME as well digital and sustainable maturity, but it is strictly focused on the manufacturing enterprises. The key words, which were used in the search, are: • Digitalization • Sustainability • Digitalization AND sustainability • Digital AND maturity AND assessment • Sustainable AND maturity AND assessment • Digital AND maturity AND maturity AND assessment AND SME • Digital AND Sustainable AND maturity AND Assessment AND SME AND model • Digital AND Sustainable AND maturity AND Assessment AND SME AND multi-criteria AND model • Twin transition AND sustainability AND digital AND model AND SME K. Mohar Bastar: The Multi-Criteria Digital and Sustainable Maturity Assessment Model 899 Table 1: Inquires in the ProQuest Dissertations and Theses Global database Database Key word (the combination Number of results in the of key words) database ProQuest Dissertations Digitalization 8919 and Theses Global Sustainability 227448 Digitalization AND 2727 sustainability Digital AND maturity 576 AND assessment Sustainability AND 409 maturity AND assessment Digital AND sustainability 157 AND maturity AND assessment Digital AND Sustainable 564 AND maturity AND Assessment AND SME Digital AND Sustainable 564 AND maturity AND Assessment AND SME AND model Digital AND Sustainable 54 AND maturity AND Assessment AND SME AND model AND multi- criteria "Twin transition" AND 0 sustainability AND "digital maturity assessment" AND model AND SME "digital maturity 0 assessment" AND "sustainability" AND "multi criteria model" 900 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 2: Inquiries in the Web of Science database Database Key word (the combination Number of results in the of key words) database Web of Science Digitalization 21.112 Sustainability 375.239 Digitalization AND 1712 sustainability Digital AND maturity AND 449 assessment Sustainability AND maturity 403 AND assessment Digital AND Sustainable 23 AND maturity AND Assessment Digital AND Sustainable 3 AND maturity AND Assessment AND SME Digital AND Sustainable 2 AND maturity AND Assessment AND SME AND model Digital AND Sustainable 0 AND maturity AND Assessment AND SME AND multi-criteria AND model Twin transition AND 1 sustainability AND digital AND model AND SME K. Mohar Bastar: The Multi-Criteria Digital and Sustainable Maturity Assessment Model 901 Table 3: Inquiries in the Scopus database Database Key word (the combination Number of results in the of key words) database Scopus Digitalization 32136 Sustainability 348195 Digitalization AND 1749 sustainability Digital AND maturity AND 730 assessment Sustainability AND maturity 407 AND assessment Digital AND Sustainability 45 AND maturity AND Assessment Digital AND Sustainable 6 AND maturity AND Assessment AND SME Digital AND Sustainable 5 AND maturity AND Assessment AND SME AND model Digital AND Sustainable 0 AND maturity AND Assessment AND SME AND multi-criteria AND model Twin transition AND 1 sustainability AND digital AND model AND SME Besides the sources, which we wil study and sort in the fol owing process, we have found approximately 50 other sources, including directives, regulations, guidelines, and recommendations, used to regulate the field from the European Commission. Literature review will give us the basis for preparation of the list of criteria, which wil be the base for multi-criteria model for digital and sustainable maturity assessment of SME in Slovenia. 902 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 4 Expected results Main expected result is the developed and validated multi-criteria model for assessment of digital and sustainability maturity of SMEs. With our model, SMEs wil be able to assess their current situation on the market, the position towards the competitors and strategize the future steps for successful digital and sustainable transition. At the same time and on the larger, aggregated scale, the model wil enable the decision makers to design proper politics and programs to support the development, resilience and market respond of SMEs. The model must distinguish between different levels of both, digitalization and sustainability stage separately, and on the other hand measure impact of digitalization on achieving sustainability goals. The assessment sensibility can be reached with the setting of proper rules of the decision-making model - tool. 5 Future developments Further systematic literature review is the first step in the designing of the multi-criteria model. In the preparation of systematic literature review, we are about to research the existing literature review on related areas. We are not going to timely limit the overview of the related literature, since both fields are developing for many years, even though the intersection of both fields just recently began. That is why we need an interdisciplinary approach to understand and combine both fields. Main areas of review wil be digital maturity assessment, sustainability maturity assessment and both assessments combined, especial y into the model. Through the literature review, we wil search for theoretical papers and already used methodologies and principles and the basis and knowledge to identify the criteria for the model. Once, when criteria wil be defined, we will validate them through the interviews with experts on the fields of sustainability and digitalization. We are planning 6 interviews, 3 with the sustainability experts and 3 with experts in the field of digitalization. When the criteria are determined and validated through the interviews, we wil combine them into the multi-criteria decision support model, and once again validate it with the expert focus group. K. Mohar Bastar: The Multi-Criteria Digital and Sustainable Maturity Assessment Model 903 After the model is validated through the expert focus group, we will test and validate it again with 15 SMEs. References Alsufyani, N., & Gil , A. Q. (2021). A Review of Digital Maturity Models from Adaptive Enterprise Architecture Perspective: Digital by Design. Proceedings - 2021 IEEE 23rd Conference on Business Informatics, CBI 2021 - Main Papers, 1, 121–130. https://doi.org/10.1109/CBI52690.2021.00023 Bohanec, M. (2020). IJS delovno poročilo DEXi: Program for Multi-Attribute Decision Making User’s Manual. Brenner, B. (2018). Transformative sustainable business models in the light of the digital imperative-a global business economics perspective. In Sustainability (Switzerland) (Vol. 10, Issue 12). MDPI. https://doi.org/10.3390/su10124428 Dao, V., Langella, I., & Carbo, J. (2011). From green to sustainability: Information Technology and an integrated sustainability framework. Journal of Strategic Information Systems, 20(1), 63–79. https://doi.org/10.1016/j.jsis.2011.01.002 EFRAG. (2021, April). ESRS standards. https://www.efrag.org/lab6?AspxAutoDetectCookieSupport=1 European Commission. (2014). DIRECTIVE 2014/95/EU OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL - o f 22 October 2014 - amending Directive 2013/34/EU as regards disclosure of non-financial and diversity information by certain large undertakings and groups. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32014L0095 European Commission. (2019). A European Green Deal. https://commission.europa.eu/strategy-and-policy/priorities-2019-2024/european-green-deal_en European Commission. (2020a). Unleashing the ful potential of European SMEs. https://doi.org/10.2775/218854 European Commission. (2020b). EU taxonomy for sustainable activities. https://finance.ec.europa.eu/sustainable-finance/tools-and-standards/eu-taxonomy-sustainable-activities_en European Commission. (2020c, February 19). A Europe fit for the digital age. https://commission.europa.eu/strategy-and-policy/priorities-2019-2024/europe-fit-digital-age_en European Commission. (2022). Directive (EU) 2022/2464 of the European Parliament and of the Council of 14 December 2022 amending Regulation (EU) no 537/2014, Directive 2004/109/EC and Directive 2013/34/EU, as regards corporate sustainability reporting. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32022L2464 European Commission. (2023, January). Corporate sustainability reporting. https://finance.ec.europa.eu/capital-markets-union-and-financial-markets/company-reporting-and-auditing/company-reporting/corporate-sustainability-reporting_en Gorgels, Stefan., Priem, Maximilian., Blagoeva, Tsvetelina., Martinelle, A., Milanesi, Giulio., European Commission. Directorate-General for Internal Market, I., DIW econ., & PwC. (n.d.). Annual report on European SMEs 2021/2022 : SMEs and environmental sustainability : background document. Hevner, March, Park, & Ram. (2004). Design Science in Information Systems Research. MIS Quarterly, 28(1), 75. https://doi.org/10.2307/25148625 Hildebrandt, B., Hanelt, A., & Firk, S. (2018). Sharing Yet Caring: Mitigating Moral Hazard in Access-Based Consumption through IS-Enabled Value Co-Capturing with Consumers. Business and 904 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Information Systems Engineering, 60(3), 227–241. https://doi.org/10.1007/s12599-018-0532-6 Kallio, H., Pietilä, A.-M., Johnson, M., & Kangasniemi, M. (2016). Systematic methodological review: developing a framework for a qualitative semi-structured interview guide. Journal of Advanced Nursing, 72(12), 2954–2965. https://doi.org/10.1111/jan.13031 Kljajić, M., Borštnar, K., & Pucihar, A. (2021a). electronics Article. https://doi.org/10.3390/electronics Kljajić, M., Borštnar, K., & Pucihar, A. (2021b). electronics Article. https://doi.org/10.3390/electronics Kljajić Borštnar, M., & Pucihar, A. (2021). Multi-Attribute Assessment of Digital Maturity of SMEs. Electronics, 10(885). https://doi.org/10.3390/electronics Lasrado, L., Vatrapu, R. K., Andersen, K. N., Lasrado, L. A., & Vatrapu, R. (2015). MATURITY MODELS DEVELOPMENT IN IS RESEARCH: A LITERATURE REVIEW Democratic and technological innovation: An inquiry into the relations of power, technology and democracy in Greenlandic e-democracy View project Collaborative Representations View project MATURITY MODELS DEVELOPMENT IN IS RESEARCH: A LITERATURE REVIEW. https://doi.org/10.13140/RG.2.1.3046.3209 Muench, Stefan., Stoermer, Eckhard., Jensen, Kathrine., Asikainen, Tommi., Salvi, Maurizio., Scapolo, Fabiana., & European Commission. Joint Research Centre. (2022). Towards a green & digital future : key requirements for successful twin transitions in the European Union. Nasiri, M., Saunila, M., & Ukko, J. (2022). Digital orientation, digital maturity, and digital intensity: determinants of financial success in digital transformation settings. International Journal of Operations and Production Management, 42(13), 274–298. https://doi.org/10.1108/IJOPM-09-2021-0616 Pucihar, A. (2020). The digital transformation journey: content analysis of Electronic Markets articles and Bled eConference proceedings from 2012 to 2019. Electronic Markets, 30(1), 29–37. https://doi.org/10.1007/s12525-020-00406-7 Vidmar, D. (2021). VPLIV INFORMACIJSKIH TEHNOLOGIJ NA TRAJNOSTNO USPEŠNOST ORGANIZACIJ. DOCTORAL CONSORTIUM VALIDATION OF DATA MATURITY CRITERIA FOR SMALL AND MEDIUM-SIZED ENTERPRISES BLAŽ GAŠPERLIN University of Maribor, Faculty of Organizational Sciences, Kranj, Slovenia blaz.gasperlin1@um.si Data is a fundamental source for any business to operate and to develop. Inefficient data management can lead to lack or a flood of data and consequently non-optimal business decisions. This is particularly evident in small and medium-sized enterprises Keywords: (SMEs), which often lag behind due to limited resources data, (financial, human, time, knowledge). To support SMEs in data understanding how to manage and utilize data effectively we maturity, data propose a data maturity assessment multiple-cirteria model. maturity Important criteria were identified from the literature and have criteria, been validated through semi-structured interviews with seven validation, SMEs, Slovenian SMEs. The results suggest some new criteria to Bled describe the data maturity, relevant to Slovenian SMEs. eConference DOI https://doi.org/10.18690/um.fov.4.2023.60 ISBN 978-961-286-751-5 906 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY 1 Introduction In today's society, there is an increasing focus on the transition from traditional to digital business, dictated by rapid and constant changes in the business environment and in society in general. The driving force of these changes is digital transformation, which leads organizations to a new way of doing business, changes in business processes, development of new products and services, and new business models (Kraus et al., 2021). An important part of digital transformation is also the data that the organization creates and captures throughout the business process. Data is the foundation of any information system, and its effective control and management is becoming increasingly important due to its exponential growth. According to the European Commission, organizations that invest in data-driven innovation experience 5% to 10% faster growth (European Commission, 2022). Access to, control and management of data is thus becoming an increasingly important strategic resource and a necessity for the further organizations' development. While large enterprises usual y have a clear overview of their data and wel -organized data management, smal and medium-sized enterprises (SMEs) usually lag behind in this endeavor, as they often have limited human, financial and knowledge resources. To help enterprises assess their level of maturity in data management, a number of data maturity assessment models have been developed and proposed. Current data maturity assessment models are not tailored to SMEs, since they are too extensive or incomplete in the number of criteria proposed, too complex in their structure, or focused on large enterprises. Therefore, there is a need to develop a new data maturity assessment model, that wil be tailored to SMEs and will include an appropriate set of criteria that SMEs can use to systematical y and comprehensively assess data maturity. In this paper, we present the results of data maturity criteria validation based on seven Slovenian SMEs. To obtain the results, we fol owed a design science research approach (Hevner, 2007) and conducted semi-structured interviews (Adams, 2015). Based on the interviews conducted with the SMEs, we will develop a final list of criteria and data maturity assessment scales that wil represent the basis for development of a data maturity multi-criteria assessment model for SMEs. B. Gašperlin: Validation of Data Maturity Criteria for Smal and Medium-Sized Enterprises 907 2 Theoretical findings The previous literature on data maturity models is disperse and focuses on different fields, from data maturity models for the public sector and Open Government (Çaldağ & Gökalp, 2022; Okuyucu & Yavuz, 2020; Rahmatika et al., 2019), to Big Data (Coleman et al., 2016; Comuzzi & Patel, 2016) and data analytics (Mach-Król, 2022), data-driven logistics (Muehlbauer et al., 2022), and data-driven decision making (Nijzink, 2020). Further, we present a few papers that are often cited in the field of data maturity. (Comuzzi & Patel, 2016) developed a maturity model that would help organizations to generate and appropriate value from Big Data. The proposed model assess the data maturity through five main domains: Strategic alignment (Strategy); Organisation (People and Culture); Governance; Data (management and data analytics) and Information technology (IT infrastructure). The limitation of the proposed model is that it does not provide specific suggestions on what steps an organization should take to improve Big Data maturity and the capabilities mentioned above. The model also lacks some of the criteria for a comprehensive assessment of data maturity, such as data quality, data security, use of data to support decision making, etc. (Coleman et al., 2016) proposed a Big Data maturity model specifical y for SMEs to advance in data maturity and also proposed recommendations. The model assess the data maturity through seven domains: business strategy; data management; people and analytical skills; technological infrastructure; enterprise adoption (engagement in data-centric management) and data governance. Although the proposed model is intended for SMEs, it focuses on data analytics, which is only one of the elements needed for a comprehensive data maturity assessment. (Peña et al., 2018) proposed a data maturity model for the SMEs in a healthcare sector, using the ELECTRE methodology, based on the ISO 15504. The assessment of data maturity is based on the four data maturity domains: people (human resources), processes, technology and data. The structure of the model is more difficult for SMEs to understand, since it is based on the set of mathematical expressions and operations. It is also not clear how SMEs have organized data storage, whether there is organized any training in the data management field and what is the level of awareness and mindset of 908 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY employees related to data management, which represent an important part of organizational culture and a basis for a comprehensive data maturity assessment. 3 Problem definition The fundamental problem is that the majority of SMEs do not exploit the potential of data for decision-making, which translates in a lower responsiveness to the dynamic demands of the environment and lower competitiveness. Although various software solutions are available to organizations (enterprise software solutions (ERP and CRM), human resource management (HRM) solutions, document systems) that allow them to manage the data generated during the execution of business processes, the implementation of such solutions is stil low (SURS, 2021). The implementation and use of business software solutions is the basis for being able to capture, manage, share and store data. This ensures the entire lifecycle of data management, otherwise a comprehensive approach to data management is not possible. Since SMEs usually lack resources (financial, human, time, skills), they need a comprehensive, systematic and easy-to-use tool to help them assess the state of their data and understand the next steps to data maturity. For a comprehensive assessment of an organization's data maturity, several criteria must be considered: Use of digital technologies and software solutions, data quality, data security, organizational culture, human resources, strategy, data lifecycle, use of data for decision making and others. Therefore, we need to consider data maturity as a multi-criteria problem. 4 Methodology As a research approach, we fol owed a design science research (DSR) (Hevner et al., 2004) (Figure 1). The final solution will be a developed IT artifact – in our case, a data maturity assessment model for SMEs. Design science research refers to the iterative sequence of expert activities, to produce an innovative product (artifact). B. Gašperlin: Validation of Data Maturity Criteria for Smal and Medium-Sized Enterprises 909 Figure 1: Methodology process - adapted from (Hevner, 2007) Based on the literature review, we defined a preliminary list of criteria to be used to assess the data maturity of SMEs (rigor cycle). After formulating a preliminary list of criteria, we prepared the interview questions. We used a semi-structured interview, as this allowed us to include additional questions that were initially not planned. Based on the literature reviewed that addressed the data maturity assessment of the SMEs and the prepared preliminary list of criteria, we defined the interview questions and conducted the semi-structured interviews with a total of seven SMEs. Before conducting the interviews, we obtained information about each SME from the SME website, such as the their business area, number of employees, and whether they had a social media presence, which helped us to define the final list of interview questions for the criteria validation. After interviewing the SMEs, we prepared a transcript of each interview and analyzed the results. The findings obtained from the interviews represent the results of the validation of the data maturity criteria and the feedback to our existing knowledge and literature (relevance cycle). 5 Preliminary/Expected results and future developments 5.1 Preliminary results In the following, we present the results of data maturity criteria validation of the 7 interviewed SMEs (SMEs A to G), of which one is micro-sized (SME A), three are 910 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY small (SMEs E, F, and G) and three are medium-sized enterprises (SMEs B, C, and D). The interviewed SMEs operate in different business sectors – telecommunication (SMEs A and B), manufacturing (SMEs C, E, F and G) and healthcare sector (SME D). The results of the criteria validation are presented in Tables 1 through 7. The (X*) in Tables (1-7) indicates that the SMEs are either using only partial solutions (i.e. Internet of Things only) or the implementation of solution (i.e. e-business implementation) is stil in the planning phase. Validation of criteria presented in Table 1 shows that criteria related to the implementation of Microsoft solutions (e.g. Office 365) and e-signing proved irrelevant for assessing the data maturity of SMEs. Since all interviewed SMEs use the Microsoft Office software, we can not identify any significant differences that would indicate how data mature each of the interviewed SME is based on the use of the Microsoft solutions. In this context this criterion is irrelevant and will be removed for data maturity assessment. Similarly, the criterion of e-signing implementation is irrelevant as it can be integrated as part of e-business and also does not help us to differentiate the data maturity level of the SMEs. B. Gašperlin: Validation of Data Maturity Criteria for Smal and Medium-Sized Enterprises 911 Table 1: Use of digital technologies and software solutions by SMEs SMEs Technologies & software solutions A B C D E F G Accounting programs X X X X X X X Human resource (HR) management X X X X X X software solutions Enterprise software solutions (ERP, CRM) X X X* X X* X* Microsoft Office – i.e. Office 365, Office X* X X X* X X X 2019 Document system X X X X X Sales and marketing programs X X X X Basic programs for data analysis - Excel X X X X X Programs for data analysis - internal program (own solution) X X Advanced data analytic programs – i.e. Microsoft Power BI, Tableau, SAS BI, Qlik X X Sense Social media (i.e. Facebook, LinkedIn, X X X X X X X Instagram …) Advanced technologies (Artificial intel igence, high performance computing X* X* X* X* (HPC), digital twin, Internet of Things, robotics) Implementation of E-business X* X X X* X* X X* Implementation of E-signing X X X X X The interview results related to data lifecycle criteria (Table 2) show that SMEs are not yet fully exploiting the potential and value of the data they have. One of the most relevant criterion to assess the data maturity of SMEs is data innovation, but the results show that the main use of data is still focused on supporting the day-today business operations. Innovation based on data or data-driven innovation (Babu et al., 2021) refers to the new value creation based on data, such as conducting market analyses for particular products, adaptation to the market and to customers, 912 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY development, upgrade and introduction of new products and services, improving the competitive nature of organization. Therefore, data innovation is most important criterion to include for data maturity assessment. During the SME interviews, we found that SMEs prefer to store their data within the company, and are less wil ing to use cloud services to store their data. Nevertheless, the criterion of data storage is important in assessing SMEs' data maturity, as it can show different maturity levels of SMEs, from those that still store their data on paper or only on their computers, to those that are more mature and also use cloud solutions. The validation also revealed that an SME that has a document system may also have electronic archiving. Despite the fact that electronic archiving is part of the document system, it is important to consider it as an independent criterion. The distinction according to the level of archiving (paper archiving, digital archiving, certified digital archiving) shows how mature an SME is in this context and is therefore a relevant criterion for evaluating the data maturity of SMEs. B. Gašperlin: Validation of Data Maturity Criteria for Smal and Medium-Sized Enterprises 913 Table 2: Data lifecycle by SMEs SMEs Data lifecycle A B C D E F G Internal data capture X X X X X X X External data capture (publicly available X* X X X* X* X X* data) Data Capture - Social Media X Data storage - company servers X X X X X X Data storage - NAS server X X X X Data storage - cloud X X X Data analysis - Excel X X X X X Data analysis - internal program X X Data analysis - advanced tools (e.g. PowerBI) X X Data analysis - Social networks (Facebook, X LinkedIn, Instagram) Data use - support of business and daily X X X X X X X activities Data use - development of new products and X X services Data use - market trends predictions X X Data use - strategic level X X X E-archiving X X X* X The reporting of data is another criterion to consider as it differentiates the SMEs from those who only report on data that is mandatory by the legislation, such as report on packaging and financial data (taxes, financial records) and those who alredy report data related to sustainability. The case of SME F has shown a need to include an additional criterion related to regulation and the sustainability aspect (i.e. reporting on CO2 consumption), that we need to consider for a comprehensive data maturity assessment of SMEs. 914 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 3: Data security implementation by SMEs SMEs A B C D E F G Care for data security - X each individual Care for data security - own IT department Care for data security - X X X X person (administrator) Care for data security - external contractors X X X* X Data (outsourcing) security Security mechanisms - backup copies of data X X X X X* and documents Data access - granted appropriate rights to X X X X access data Information security X X X X policy Regarding data security (Table 3), none of the SMEs have an employee or IT security expert specifical y responsible for data security, and the majority of SMEs prefer external contractors (outsourcing). The results of the criteria validation showed that the first criterion related to care for data security (care for data security - each individual) seem to be not relevant, as only one SME (SME G) stated that it has no one to take care of data security and that is the responsibility of each individual. But from the view of assessing the data maturity of SMEs, it can show the differences in the organization and appointment of the roles (i.e. each individual, data security administrator, IT department, data steward). B. Gašperlin: Validation of Data Maturity Criteria for Smal and Medium-Sized Enterprises 915 Table 4: Data quality strategy and data quality review results by SMEs SMEs A B C D E F G Establishment of a data quality X X X strategy Data A formal y written strategy for X X quality data quality Data quality review X* X X X Data quality assessment X X X Data quality is one of the most important criteria to ensure accurate decisions based on validated data. Although the validation results of the data quality criteria in Table 4 appear to be less relevant to SMEs, al criteria must be included to obtain a comprehensive data maturity assessment. Table 5: Establishement of data management strategy and investments in data management by SMEs A B C D E F G Established data management Data X strategy management strategy A formally written data X management strategy Information infrastructure X X X X X X X investments Investments Data management X X investments The results in Table 5 show that the criteria related to data management strategy and investment in data management do not seem to be so relevant when assessing SMEs' data maturity, as only one SME (E) indicated that it already has this data aspect in place and attaches high importance to it. Only one other SME (SME G) indicated that they always use part of their investments specifical y for improving data management. Despite the fact that the majority of SMEs only invests in information infrastructure, it is important to also consider the investments in data managent and to start implementing a data management strategy, either as part of the business 916 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY strategy as this is a fundamental basis for the transition to a data-driven enterprise as indicated by (Davenport & DalleMule, 2017). Table 6: Organizational culture and human resources by SMEs A B C D E F G Awareness of efficient data X* X* X* X* X X X* management Organizational culture Spreading the culture of efficient X* X* X X data management among employees Education and training in the field X X* X X of data management Encouragment of employees for better X X X X X X Human data management resources (HR) Open X X X X X X X communication External contractors - involved X X* X X X* External contractors - not involved X X The results of Table 6 show that 4 out of 7 interviewed SMEs organize training and education for their employees so that they can better manage their data and acquire additional skills. This indicates that organizing training and educadion on data management is an important criterion to consider, even though interviewed SMEs currently educate their employees mainly on how to use Excel. The criterion of open communication is highly relevant, as the results in Table 6 show that open communication is present in al 7 interviewed SMEs, and for this reason must be considered when assessing SMEs' data maturity. B. Gašperlin: Validation of Data Maturity Criteria for Smal and Medium-Sized Enterprises 917 From the results in Table 7, SMEs are beginning to consider the use of data at a strategic level as wel , but they agree that intuition and experience must also be considered. In the case of SME C, they are particularly committed to basing their decisions on data. The director of SME C stated that 70% of strategic decisions are based on data and the remaining 30% are based on soft factors, reflecting the environment and employee attitudes. The situation is similar in SME F, where the percentage of data-driven decisions is 50%. Other SMEs are not yet ready to adopt this mindset. When we asked SMEs whether they have adequate data when making strategic decisions, the answers were mixed. Three of seven SMEs interviewed (C, D and F) stated that the data are adequate. In the case of SME A and B the adequacy of data for strategic decisionsis is not so relevant. At the moment they are more focused to provide the data to support the operating of the business and then start to build from that point further. Based on this observations the criterion related to the adequacy (relevance) of data for decision-making is important to include for assessing the data maturity of SMEs. 918 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Table 7: Establishment of data-driven decision making on strategic level A B C D E F G Intuition-driven decision making Intuition-driven + data-driven decision X* X X X X X X* making Data-driven decision making Data sources for Decision decision-making - X X* X X X* making CRM, ERP (strategic Data sources for decisions) decision-making- X X X X X X (from Excel) Data sources for decision-making (Geographic X information systems (GIS) Adequacy of data in NR NR X X X X X* decision-making NR – Not relevant 5.2 Expected results and future developments The semi-structured interviews we conducted with the seven SMEs revealed opportunities for the inclusion of new criteria, such as those related to regulation and the sustainability aspect (i.e. reporting on CO2 consumption, packaging) and the need to remove some of the currently proposed criteria (i.e. e-signing, implementation of Microsoft Office solutions). Since only the results of 7 interviewed SMEs are presented, we expect to add additional criteria or remove some of the currently proposed criteria in future research, as shown by the results of the criteria validation in this paper. In future research, we wil focus on formulating the final list of data maturity criteria for SMEs and finalising the definition of the measurement scales, which will help us develop the first version of B. Gašperlin: Validation of Data Maturity Criteria for Smal and Medium-Sized Enterprises 919 a multi-criteria data maturity model for SMEs. We expect that we wil need to develop the model in several iterations before the final version. Acknowledgements The research is financially supported by the Slovenian Research Agency of the Republic of Slovenia (ARRS): program no. P5-0018 "Decision support systems in digital business" and by the program for Young researchers, no. 54752-0586-21. References Adams, W. (2015). Conducting Semi-Structured Interviews. In J. Wholey, H. Hatry, & K. Newcomer (Eds.), Handbook of Practical Program Evaluation (4th ed.). Jossey-Bass. https://doi.org/10.1002/9781119171386.ch19 Babu, M. M., Rahman, M., Alam, A., & Dey, B. L. (2021). Exploring big data-driven innovation in the manufacturing sector: evidence from UK firms. Annals of Operations Research, 1–28. https://doi.org/10.1007/S10479-021-04077-1 Çaldağ, M. T., & Gökalp, E. (2022). The maturity of open government data maturity: a multivocal literature review. Aslib Journal of Information Management, 74(6), 1007–1030. https://doi.org/10.1108/AJIM-11-2021-0354 Coleman, S., Göb, R., Manco, G., Pievatolo, A., Tort-Martorel , X., & Reis, M. S. (2016). How Can SMEs Benefit from Big Data? Challenges and a Path Forward. Quality and Reliability Engineering International, 32(6), 2151–2164. https://doi.org/10.1002/QRE.2008 Comuzzi, M., & Patel, A. (2016). How organisations leverage Big Data: a maturity modl. Industrial Management & Data Systems, 116(8), 1468–1492. https://doi.org/10.1108/IMDS-12-2015-0495 Davenport, T. H., & Dal eMule, L. (2017). The 2 Types of Data Strategies Every Company Needs. Harvard Business Review. https://hbr.org/2017/05/whats-your-data-strategy European Commission. (2022). Data Act. https://digital-strategy.ec.europa.eu/en/library/data-act-factsheet Hevner, A. R., March, S. T., Park, J., & Ram, S. (2004). Design Science in Information Systems Research. MIS Quarterly, 28(1), 75–105. https://doi.org/10.2307/25148625 Kraus, S., Jones, P., Kailer, N., Weinmann, A., Chaparro-Banegas, N., & Roig-Tierno, N. (2021). Digital Transformation: An Overview of the Current State of the Art of Research. SAGE Open, 11(3), 21582440211047576. https://doi.org/10.1177/21582440211047576 Mach-Król, M. (2022). Conceptual Framework for Implementing Temporal Big Data Analytics in Companies. In Applied Sciences (Vol. 12, Issue 23). https://doi.org/10.3390/app122312265 Muehlbauer, K., Wuennenberg, M., Meissner, S., & Fottner, J. (2022). Data driven logistics-oriented value stream mapping 4.0: A guideline for practitioners. IFAC-PapersOnLine, 55(16), 364– 369. https://doi.org/https://doi.org/10.1016/j.ifacol.2022.09.051 Nijzink, H. J. (2020). Data-driven Decision-making Maturity. http://essay.utwente.nl/85376/ Okuyucu, A., & Yavuz, N. (2020). Big data maturity models for the public sector: a review of state and organizational level models. Transforming Government: People, Process and Policy, 14(4), 681–699. https://doi.org/10.1108/TG-09-2019-0085 Peña, A., Bonet, I., Lochmul er, C., Tabares, M. S., Piedrahita, C. C., Sánchez, C. C., Giraldo Marín, L. M., Góngora, M., & Chiclana, F. (2018). A fuzzy ELECTRE structure methodology to assess big data maturity in healthcare SMEs. Soft Computing, 23(20), 10537–10550. https://doi.org/10.1007/S00500-018-3625-8 920 36TH BLED ECONFERENCE - DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY Rahmatika, M., Krismawati, D., Rahmawati, S. D., Arief, A., Sensuse, D. I., & Dzulfikar, M. F. (2019). An Open Government Data Maturity Model : A Case Study in BPS-Statistics Indonesia. 2019 7th International Conference on Information and Communication Technology (ICoICT), 1– 7. https://doi.org/10.1109/ICoICT.2019.8835352 SURS. (2021). Elektronska izmenjava informacij znotraj podjetij, po velikosti podjetij glede na število zaposlenih in samozaposlenih, Slovenija, letno. https://pxweb.stat.si/SiStatData/pxweb/sl/Data/-/2965311S.px/table/tableViewLayout2/ 36TH BLED ECONFERENCE DIGITAL ECONOMY AND SOCIETY: THE BALANCING ACT FOR DIGITAL INNOVATION IN TIMES OF INSTABILITY ANDREJA PUCIHAR (ET AL.) University of Maribor, Faculty of Organizational Sciences, Kranj, Slovenia andreja.pucihar@um.si The Bled eConference, organised by the University of Maribor, Faculty of Organizational Sciences, has been shaping electronic interactions since 1988. The theme of the 36th conference is "Digital Economy and Society: The Balancing Act for Digital Innovation in Times of Instability". In times of instability, which include political, economic, resource, health, and environmental chal enges on the one hand, and technological disruption on the other, it is critical to ensure that digital innovation continues to lead to the right and sustainable solutions that are tailored to the needs of al people, enterprises and society. It is very important to keep in mind the protection of our planet, including fauna and flora. These efforts include adopting appropriate regulatory frameworks, fostering digital literacy and skills development, promoting inclusive access to digital technologies, and addressing the ethical, social and environmental implications of digital transformation. The papers in this conference Keywords: proceedings address digital transformation of enterprises, digital artificial intelligence and data science solutions, decision economy, digital analytics for business and societal chal enges, new, digital and society, data driven business models, digital consumer, digital education, digital digital health, digital ethics, restructured work and solutions for transformation, digital smart and sustainable cities. We continue to provide an open innovation, forum for academia, including students, industry, and policy instability, makers where everyone can contribute to creating a better world. balancing DOI https://doi.org/10.18690/um.fov.4.2023 ISBN 978-961-286-751-5 Document Outline 1 Introduction 2 Theoretical Background 3 Research Design: Taxonomy Development Process 3.1 Determination of meta-characteristics and ending conditions 3.2 First cycle: Literature research and classification 3.3 Second cycle: Meta-model and taxonomy revision 4 Taxonomy 5 Taxonomy applied to the SLR Platform 6 Discussion, Limitations, and Future Work 1 Introduction 2 Theory and concepts 3 Methods and Materials 4 Results & Cross-Case Comparisons 5 Discussion 1 Introduction 2 Background 3 Method 4 Results 5 Discussions, Limitations & Conclusions Funding Acknowledgments References 1 Introduction 2 Literature Review 2.1 Digitalisation’s Role in Changing Business Models 2.2 Business Model Innovation and Digitalisation in Ports 3 Methodology 3.1 Research Design 3.2 Selection of Sites 3.3 Data Collection and Analysis 4 Findings 4.1 Drivers for Business Model Change in RoPax Ports 4.2 Changes in the Business Models of RoPax Ports 4.3 Digitalisation Efforts at RoPax Ports 5 Discussion 6 Conclusion References 1 Introduction 2 Review of the Literature 2.1 Digitalization, Data Analytic Capability, and Industry 4.0 2.2 Theoretical Foundations 2.3 Data Analytic Capability Development 3 Methodology 4 Findings 5 The Data Analytic Capability Wheel as a Metaphor 5 Discussion 7 Limitations and Further Research 1 Introduction 2 Literature review and hypothesis 3 Methodology 3.1 The measures 3.2 Survey administration, sample and data collection 4 Data analysis and findings 4.1 Reliability and validity 4.2 Structural model findings 5 Discussion 5.1 Theoretical contributions 5.2 Managerial implications 6 Conclusions Appendix 1 Introduction 2 Literature Review 2.1 Digitalisation and its Effect on Business Models and Ecosystems 2.2 Typologies for Digital Solutions in Industrial Sectors 3 Methodology 3.1 Research Design 3.2 Data Collection and Analysis 4 Findings 4.1 Challenges Faced by RoPax Ports 4.2 Typology Development 4.3 Analysis of Digital Solutions for RoPax Ports 5 Discussion and Conclusions 1 Introduction 2 Analytical Approach 2.1 Material for Analysis 2.2 Analysis Scheme 2.3 Analytical Procedure 3 Malicious Insider Threat Types 3.1 Disgruntled Employee / Disgruntled Leaver 3.2 Data Transfer to Competition 3.3 Industry Espionage 3.4 State Espionage 3.5 Taking Advantage of Privileges for Personal Gain 3.6 Unauthorized Inspection of Personal Data 3.7 Intellectual Property Sale 3.8 Whistleblowers 3.9 Politically Motivated Sabotage 3.10 Extortion 3.11 Illegal Use of IT Infrastructure 4 Conclusion 1 Introduction 2 Background and Related Work 3 Research Method 3.1 Data Collection 3.2 Data Analysis 4 Results 5 Discussion and future research 6 Conclusion 1 Introduction 2 Literature Review 3 Methodology 3.1 Content analysis and data collection 3.2 Data Analysis 4 Discussion and Result 4.1 Falling trends 4.2 Rising trends 4.3 Differing trends of programming languages (R, Python, SQL) 4.4 Background requirement/specialization 4.5 Grouping of professions 5 Conclusions Appendix A: Literature Review of Data-related Positions 1 Introduction 2 Literature Review 3 Methodology 4 Findings 4.1 Customer-Related QR Code Usage Barriers 4.2 Company-Related QR Code Usage Barriers 5 Discussion 6 Limitations and Future Research Appendix A: Examples of QR code tags Appendix B: Details about the study’s interview participants Appendix C: Setting of the experiment in the brick-and-mortar store 1 Introduction 2 Methods 3 Findings 4 Discussion 4.1 Theoretical implications 4.2 Reflections on teaching 1 Introduction 2 Theoretical Grounding 2.1 Learning Strategy, Exam Preparation, and Time Management 2.2 Motivation 2.3 Social Learning 3 Method 4 Artifact Description 4.1 Motivational Goal Setting (DP1) 4.2 Enabling Social Interactions and Feedback (DP2) 4.3 Focus and Control (DP3) 4.4 Fostering Resilience (DP4) 4.5 Context- and Learner Adaptation (DP5) 5 Evaluation 6 Discussion 7 Conclusion References 1 The History of Leadership and Group Dynamics Research 2 Rank Dynamic Mechanisms and Strategies in Recent Works 3 Hypothesis and Objective of This Work 4 Data Collection and Analysis 4.1 Descriptive Analysis and Assessment of Model Quality 4.2 Exploratory Factor Analysis (EFA) 4.3 Calculation of Higher Order EFA 4.4 Confirmatory Factor Analysis (CFA) 6 Summary and Outlook 1 Introduction 4.2 Discussion 1 Introduction 1 Introduction 2 Theoretical framework 2.1 Blended and Active Learning 2.2 Community of Inquiry 2.3 Adaptive Learning Technologies 3 Methodology 3.1 Context 3.2 Research Design 4 Results 4.1 Interviews with teachers 4.1.1 Ambiguous definitions 4.1.2 Motivations for deployment of the ALT 4.1.3 Teachers recognize added value, but also make critical comments 4.1.4 Teaching presence dominant 4.1.5 Teacher dashboard could be used more effectively 4.1.6 Teachers did not sufficiently share or discuss 4.2 Student trace data analysis 5 Conclusions and Discussion Appendix A - Further Explanation ALT Appendix B - Trace Data Explanation 1 Introduction 2 Research Background 2.1 Meta-platform offerings 2.2 Data sovereignty dimensions and indicators 2.3 Control mechanisms: Smart contracts and certification 3 Research approach 4 Results 5 Discussions and conclusions 1 Introduction 2 Theoretical Framework 3 Methods and Materials 3.1 Sample and Data Collection 3.2 Measures 4 Analysis and Results 5 Conclusion and Discussion 1 Introduction 2 Theoretical Background 2.1 Omnichannel CX in B&M Store Environments 2.2 QR Codes 3 Methodology 4 Findings and Discussion 5 Conclusion, Limitations, and Future Research Appendix 1: Information on the participants 1 Introduction 2 Responsible online purchasing and its obstacles 3 Data collection and analysis 4 Findings 5 Discussion 1 Introduction 2 Theoretical background 3 Research Method 4 Results 5 Discussion 6 Conclusion 1 Introduction 2 Theoretical background and conceptual model 3 Research Method 4 Results 4.1 Evaluation of Measurement Model 4.1.1 Reflective Constructs 4.1.2 Formative Constructs 4.2 Evaluation of Structural Model 5 Recommendations and Discussion 6 Conclusion 1 Introduction 2 Business Model Innovation Objectives and Industry-Academia Collaboration: Literature review 3 Methodology 4 Results 5 Conclusions 1 Introduction 2 Research Objectives 3 Literature Review 3.1 Financial responsibility 3.2 BankID 3.3 Digital Payment Methods 4 Method 5 Results 6 Discussion 7 Conclusion 1 Introduction 2 Theoretical background 2.1 Sport and wellness technology digital coaching 2.2 Exercise Self-efficacy 3 Methodology 3.1 The Digital Coach Used in the Study 3.2 Research approach, data collection and analysis 4 Findings 4.1 Quantitative findings related to exercise self-efficacy and attitude 4.2 Qualitative findings 4.2.1 COVID-19 influencing exercise behavior 4.2.2 Adaptation and usage of the digital coach 4.2.3 Motivational influence of the digital coach 4.2.4 Ideal digital coach 5 Discussion 6 Limitations and suggestions for future research Acknowledgements 1 Introduction 2 Review of Related Work 3 Relevant Theories 3.1 The Theory of Task Technology Fit (TTF) 3.2 Design Science Research Methodology (DSRM) 4 Methodology 4.1 Identification of problem and motivation 4.2 Definition of the objectives for a solution 4.3 Design and Demonstration 4.4 The Artifact 4.5 Assessment 4.6 Partnering with a Commercialization Partner 4.7 Commercialization Partner’s Assessment 4.8 Recommendations for Commercialization 5 Results 6 Discussion 7 Conclusions 1 Introduction 2 Research Background 2.1 Game Story 2.2 Learning Outcome 3 Serious Game for Learning Presentation Skills 4 Fictional or Realistic Game Story: Study on Learning Outcome 4.1 Study Design 4.2 Study Results 4.3 Discussion of Results 5 Conclusion and Future Research 1 Introduction 2 Research background 3 Method 4 Analysis and results 5 Conclusions and implications 1 Introduction 2 Related Work 2.1 Physical Activity and ADHD 2.3 Simulator Based Interventions 3.2 Apparatus 3.3 Movement Assessment 3.4 Interviews 6 Conclusion 1 Introduction 2 Personalized and personal 3 Research Method 4 Results 5 Discussion 6 Conclusions 1 Introduction 2 Background 3 Research method 4 Data Collection & Analysis 5 Results 5.1 Discovery phase 5.2 Conformance Checking phase 5.3 Improvement phase 5.4 Non-functional requirements 6 Discussion 7 Conclusion 1 Introduction 2 Theoretical Background 2.1 Participation and motivation 2.2 Conversational Agents and Virtual Companions 3 Methodology/Design Science Research 3.1 Problem Space 3.2 Solution Space 4 Designing a Participation Companion 4.1 Deriving User Stories for Participation Companions 4.2 Designing of the artifact 5 Explorative evaluation 6 Discussion 7 Conclusion 1 Introduction 2 Methodolodgy 2.1 Simulation model 2.2 Centralized market 2.2.1 Costs and profits distribution rule 2.2.2 Internal pricing mechanism 2.3 Decentralized market 2.3.1 Random bidding strategies 2.3.2 Learning algorithm based bidding strategies 3 Case study description 4 Simulation results 4.1 Case I: 40% PV penetration 4.2 Case II: 60% PV penetration 5 Conclusions 1 Introduction 2 Background 3 Research Design 4 Design Requirements 5 Design Principles and Features 5.1 DP1: Structuring the Process by Decision Points 5.2 DP2: Support Cyclic Divergent and Convergent Thinking 5.3 DP3: Enable Organisational Learning 6 Discussion 7 Conclusion References 1 Introduction 3 Findings 5 Reflection and directions for further research 1 Introduction 2 New competencies in the innovation management framework 3 Methodology 3.1 Methodological design and presentation cases of study and results 4 Results and Discussions 5 Conclusions Appendix 1 Introduction 2 Research Model 3 Methodology 4 Results 4.1 Construct Reliability and Validity 4.2 Indicator Reliability and Validity 4.3 Model Estimates 5 Discussion and Conclusion 6 Limitations and Future Research Appendix A: Indicator-Level Statistics Appendix B: Effects of the Control Variables 1 Introduction 2 Research Model 3 Methodology 4 Results 4.1 Construct Reliability and Validity 4.2 Indicator Reliability and Validity 4.3 Model Fit and Model Estimates 5 Discussion and Conclusion 6 Limitations and Future Research Appendix A: Item Wordings Appendix B: Indicator-Level Statistics Appendix C: Effects of the Control Variables 1 Introduction 2 Ecosystems in business and information systems literature 3 Information systems and biological ecosystems 3.1 Information systems as socio-technical systems and core concepts 3.2 Biological ecosystems and core concepts 4 Information systems as ecosystems 5 Conclusion 1 Introduction 2 Prior Work 2.1 The Pragmatic Metatheoretic Model 2.2 Authentication Theory 3 Assertions Relating to (Id)Entity 3.1 Introduction (1) An Identity Assertion (2) An Entity Assertion (3) A Simple Property Assertion (4) A Complex Property Assertion (5) A Principal-Agent Assertion (6) (Id)Entity Match Assertions 3.2 The Authentication Process 3.3 Evidence in Support of the Authentication Process 3.4 Data Quality's Role in the Authentication Process 4 Implications (1) The Effectiveness of Identity Management (2) The Effectiveness of Other Business Processes (3) The Economics of IS Design (4) Stakeholder Interests 5 Conclusions 1 Introduction 2 Lessons from the Forest Industry 3 Analytics Modelling Guides Problem Solving 4 An Agenda for Decision Analytics 5 Summary and Some Conclusions 1 Introduction 2 Motivation & Gap 3 Theoretical Background 3.1 Technostress: A Fragmented and Evolving Literature 3.2 IT Identity 4 Preliminary Research Design and Research Model 5 Conclusion and Expected Contributions 1 Introduction 2 Digital Business Models and Privacy Concerns 2.1 Focused Literature Review: Digital Business Models and Personal Boundaries on Privacy 3 Initial Findings and Future Research Direction Appendix 1 Introduction 2 Theoretical background 3 Research method 4 Preliminary results 5 Conclusion and further research 1 Introduction 2 Methodology 3 Results 4 Discussion and Outlook The research was funded by the European Union as part of the Erasmus+ "Collaborative development of Al capabilities in SMEs" project (Grant number #2O22.T. ATO1.KA22O.H ED.OOOO89256). References 1 Introduction 2 Related literature 3 Methodology 4 Preliminary results 5 Conclusion 1 Introduction 2 Method Followed for the Scoping Review 3 Results 4 Discussion 5 Conclusions Aknowledgements 1 Introduction 2 Theoretical background 3 Research method 4 Preliminary results 5 Conclusion and discussion Appendix Figure 1: Adapted PRISMA 2020 flow diagram 1 Introduction 3 Digitalisation of vessels and traffic Summary of Research in Progress 1 Introduction 2 Job Demands-Resources (JD-R) Model 3 Generative AI and Virtual Assistants 4 Methodology 5 Conclusion 1 Introduction 2 Background 2.1 Introducing technological and digital devices 3 Towards an adaptive tool for implementation activities 4 Conclusion, limitations and further research Appendix 1: implementation framework for technological devices in operating rooms 1 Introduction 2 Search engine concept in the IoT context 3 Component-based approach The idea to combine IoT with software components and services is not very new. The group of Zhiming Liu (2010) has been working on software-hardware components fusion for a long time before the rise of IoT. In another reference, Ruppen et al. (2015) ... In this section, we continue the architectural approach described in (Gula, Flakova, 2017; Giedrimas, Backys, 2022; Saari, Nurminen, Rantanen, 2022) and will analyze the most popular component-based technologies: CORBA, EJB, OSGi, and .NET. 3.1 CORBA CORBA (Common Object Request Broker Architecture) is a middleware technology that provides a way for distributed applications to communicate with each other. CORBA was an industrial standard for the past two decades. They are still several legacy syst... 3.2 EJB and OSGi 3.3 .NET 3.4 Web services 3.6 Frameworks After the analysis, we came to the following conclusions: 1 Introduction 2 Literature Review and Theoretical Framework 2.1 Task Technology Fit (TTF) 2.1.1 Task Requirements 2.1.2 Technological Characteristics 2.1.3 Individual Characteristics 2.2 Technostress 2.3 Satisfaction with online learning and expected performance. 2.4 Hypotheses. 3 Method 3.1 Measures. 3.2 Analyses and Results. 1 Introduction 2 Problem definition 2.1 Research question 3 Methodology 3.1 Research approaches 3.1.1 Case Study Research Methodology 3.1.2 Design Science Research 3.1.3 Systems Development Life Cycle (SDLC) 3.1.4 Methodology for resolving a problem and methodology for building systems 3.2 Working procedure 4 Preliminary/expected results 4.1 Literature review 4.2 Further results 4.3 Scientific contribution 4.4 Objectives of the doctoral dissertation and the envisaged scientific research results 4.5 Potential results of the study and the importance of those results 5 Future development 1 Introduction 2 Methodology 3 Theoretical Background 4 Expected results 1 Introduction 1 Introduction 2 Problem definition 3 Methodology 3.1 Exploratory phase 3.2 Confirmatory phase 3.3 Additional considerations 3.4 Conceptual model 4 Expected results 5 Future development 6 Current issue 1 Introduction 2 Problem definition 3 Methodology, design and focus 3.1 Iterative research approach and Pre-studies 3.2 Procedure and data collection 3.3 Structure of the experiment 3.4 Types/formats of inter-team retrospectives 3.5 Data collection approach 3.6 Statistical evaluation 3.7 Data quality 3.8 Validity and reliability 3.9 Ethics 4 Preliminary/Expected results 5 Future development 1 Introduction 2 Problem definition 3 Methodology 3.1 Literature review 4 Expected results 5 Future developments 1 Introduction 2 Theoretical findings 3 Problem definition 4 Methodology 5 Preliminary/Expected results and future developments 5.1 Preliminary results