< Back Towards a Pragmatic Approach for studying Normative Recommender Systems Binst, B., Vandenbroucke, H., Li, D., Puskas, I., Van der Elst, P., & Smets, A. (Accepted/In press). Towards a Pragmatic Approach for studying Normative Recommender Systems: Exploring Power Dynamics in Digital Platform Markets. In Proceedings of the the First Workshop on Normative Design and Evaluation of Recommender Systems . Recommender systems (RSs) have become an integral part of digital platforms, helping users to navigate through the abundance of online content and services. This paper argues that the study of normative RSs should go beyond the mere conceptualization of normativity and examine the conditions and contextual factors that might foster normative RSs. RSs often operate within multisided markets, involving multiple stakeholders who influence and are influenced by the RS. Ideally, developers of RSs aim for value alignment among diverse stakeholders, fostering a win-win scenario. However, in practice, system providers hold more power over the design, often optimizing for their own objectives and neglecting the needs and values of other stakeholders. These imbalanced power dynamics in digital platform markets pose significant challenges to the design. This paper critically examines two key economic dynamics in digital media markets that challenge the development of normative RSs. The first hurdle describes the susceptibility of digital platform markets to lock-in and monopolization due to network effects, which skews the power balance in favor system. A second hurdle critically examines the pervasive \"attention economy\" and corresponding engagement-centric logic in the design of RSs that might diverge from explicit or long-term users' preferences despite the success in increasing engagement. To illustrate the importance of these dynamics in real-world RSs, we present how these hurdles manifest in the domains of video-on-demand and news media. In conclusion, this paper advocates for a more pragmatic approach to studying normative RSs by considering the two hurdles affecting the RS development in contemporary digital platform markets. Furthermore, it highlights the fruitfulness of this pragmatic approach by proposing policy recommendations to tackle normative challenges in RSs’ design. Recommender Systems Previous Read the article Next

< Back Intended, afforded, and experienced serendipity: overcoming the paradox of artificial serendipity Smets, A. Intended, afforded, and experienced serendipity: overcoming the paradox of artificial serendipity. Ethics Inf Technol 27 , 33 (2025). https://doi.org/10.1007/s10676-025-09841-6 Designing for serendipity in information technologies presents significant challenges for both scholars and practitioners. This paper presents a theoretical model of serendipity that aims to address this challenge by providing a structured framework for understanding and designing for serendipity. The model delineates between intended, afforded, and experienced serendipity, recognizing the role of design intents and the subjective nature of experiencing serendipity. Central to the model is the recognition that there is no single definition nor a unique operationalization of serendipity, emphasizing the need for a nuanced approach to its conceptualization and design. By delineating between the intentions of designers, the characteristics of the system, and the experiences of end-users, the model offers a pathway to resolve the paradox of artificial serendipity and provides actionable guidelines to design for serendipity in information technologies. However, it also emphasizes the importance of establishing ‘guardrails’ to guide the design process and mitigate potential negative unintended consequences. The model aims to lay ground to advance both research and the practice of designing for serendipity, leading to more ethical and effective design practices. Previous Read the article Next

< Back Close Contact for Context: Qualitative Methodological Considerations for Assessing Prominence and Discoverability Iordache, C., Leiva, M. T. G., & Raats, T. (2025). Close Contact for Context: Qualitative Methodological Considerations for Assessing Prominence and Discoverability. Media Industries, 12(1), 111-132. https://journals.publishing.umich.edu/mij/article/id/7625/ This article advocates for a “close contact for context” approach as a crucial step before, or in combination with, automated data collection and comparative analyses to assess the prominence and discoverability of European works in video-on-demand (VoD) catalogs. Grounded in media industries research, the method emphasizes the importance of close observation within the dynamic interface environments of VoD services. To this end, we present a set of reflections based on an exploratory study conducted on four US-based subscription VoD services in Spain and Belgium and propose three key sets of methodological considerations to be integrated in future assessments of prominence and discoverability tools and practices. First, a broader analysis of market characteristics and their interrelations is essential for conducting comparative research and understanding the strategies of transnational players within each market. Second, VoD services must be treated as individual cases shaped by specific industry practices and business models. Third, we recommend “getting personal” with the object of study through detailed analyses of interface architectures and consumption affordances, and propose the concept of a title’s discovery arc to analyze its spatiotemporal placement. This approach underscores the value of contextual, qualitative insights to complement quantitative evaluations. Previous Read the article Next

< Back Frustrations with Access to Preferred Content: User Experiences with Bilibili’s Recommender Systems Reflected on Weibo Li, D., Ranaivoson, H., & Ballon, P. (2026). Frustrations with Access to Preferred Content: User Experiences with Bilibili’s Recommender Systems Reflected on Weibo. International Journal of Human–Computer Interaction , 1–34. https://doi.org/10.1080/10447318.2026.2669039 Understanding user experiences with recommender systems (RS) on audiovisual platforms is critical for user-centric design and long-term service sustainability. However, existing research often overlooks how users articulate their experiences, particularly within specific platform contexts. This research examines 38,545 Weibo posts to identify patterns in users’ expressed experiences with the RS of Bilibili (YouTube’s Chinese counterpart). Combining natural language processing and thematic analysis, we examine sentiment trends, key concerns, and underlying patterns. While RS are generally assumed to match user preferences, we found persistently expressed user frustrations with the lack of truly relevant and favored recommendations—those that align with long-term interests, highlight quality, consider context and platform-specific preferences, and respect user feedback. These frustrations are related to the platform’s business considerations and user-platform power asymmetries. The findings underscore the platform-specific nature of user experience and contribute to discussions on adaptive RS that satisfy user needs and support sustainable business models. Recommender Systems Previous Read the article Next

< Back Heritiana Ranaivoson Research Professor heritiana.renaud.ranaivoson@vub.be While digital technologies have been here for quite some time already, I am still amazed by the changes they keep triggering in the cultural and media industries and how they allow (or constrain) these industries to innovate. I am particularly interested in the impact these changes have on the diversity of the content these industries produce and recommend. And eventually the policies that are needed in this new environment. I analyse these changes using my background in industrial economics, business modelling and media policy, with a peculiar interest in combining quantitative and qualitative research. I obtained my PhD in Economics (Université Paris 1, Panthéon-Sorbonne) in 2008 on cultural diversity in the recording industry. I also studied Economics and Management at Ecole Normale Supérieure de Cachan (France). Since then, and notably since I joined imec-SMIT-VUB in 2010, I have led several research projects at international and national levels for the European Commission ( H2020 , study contracts ), the Unesco , Google , etc. I have published extensively in the fields of cultural and media diversity , media innovation , media policy, audiovisual platforms , and the economic impact of digital technologies on creative industries . Visit my research profile

< Back Public service algorithms: Balancing the scales between public mission and market pressures at the BBC and VRT. Iordache, C., Martin, D., & Johnson, C. (2025). Public service algorithms: Balancing the scales between public mission and market pressures at the BBC and VRT. MedieKultur: Journal of Media and Communication Research, 40(78), 59-80. https://doi.org/10.7146/mk.v40i78.147182 The public service media (PSM) shift to digital-first strategies has generated a rethinking of priorities in content production and distribution. This not only involves the integration of algorithms for the curation of their video-on-demand portals, but also a far-reaching reform at the organisational level. As personalisation through recommender systems is increasingly popularised by commercial streaming services, PSM are faced with a balancing act between market pressures and fulfilling their public mission. This article contributes to this discussion by investigating how public service algorithms are developed in practice in the cases of the BBC (UK) and VRT (Flanders-Belgium), and how their implementation is guided by the organisations’ remits. Through document analysis and semi-structured interviews conducted in 2024 with 16 PSM representatives, we discuss the ways in which market, policy, and organisational contexts inform the use of algorithms by the two organisations and suggest the need for a re-theorisation of public service algorithms. Public Service Media Previous Read the article Next

< Back Not One News Recommender To Fit Them All: How Different Recommender Strategies Serve Various User Segments Vandenbroucke, H., Maes, U., Michiels, L., & Smets, A. (2025). Not One News Recommender To Fit Them All: How Different Recommender Strategies Serve Various User Segments. In 19th ACM Conference on Recommender Systems Proceedings . ACM. Many news recommender systems (NRS) adopt a one-recommender-for-all approach, overlooking that users engage with news in fundamentally different ways. In this work, we identify user clusters based on various engagement metrics that go beyond clicks by employing cluster analysis on two real-world datasets: EB-NeRD and Adressa. Next to that, we evaluate the performance of common rec-ommender strategies: popularity, collaborative filtering (EASE and ItemKNN), and a content-based model across these user clusters, which exhibit varying reading behaviors and information needs. Our findings show that different recommender strategies are effective to varying degrees depending on the user cluster. This study contributes to NRS research by providing a grounded clustering of users derived from real-world datasets and emphasizes the importance of user-centered evaluations for understanding how NRS strategies serve audiences with varying levels of news engagement. Recommender Systems Previous Read the article Next

< Back Pieter Van der Elst PhD Researcher pieter.van.der.elst@vub.be Pieter Van der Elst joined imec-SMIT as a PhD candidate in 2023. Researching on the broader topic of algorithmic recommenders in the Flemish media industry, his thesis focuses on how the public service broadcaster VRT positions itself within a contemporary media context. His academic interests lie mainly in media economics and datafication of media. Before his start at imec-SMIT he briefly worked as a consultant for one of the leading social bureaus for artists (SBK’s) in the Flemish cultural industry, providing services exclusively to VRT. He obtained a Bachelor’s degree in Journalism and subsequently a Master’s degree in Communication Studies. Visit my research profile

< Back Understanding Algorithmic gatekeepers to promote Epistemic welfare ALGEPI Technological and economic developments have led to the availability of an overwhelming quantity of digital content. Therefore, it has become crucial, in particular for media content providers, to incorporate algorithmic gatekeepers, which filter, rank and recommend content. Are these algorithmic gatekeepers undermining media’s contribution to epistemic welfare? In ALGEPI, we start from the novel concept of epistemic welfare, defined as the individuals’ right to know and be exposed to trustworthy, independent and diverse information while respecting individual rights to their own data. By connecting legal, political, technological and sociocultural perspectives, we will develop a conceptual framework for epistemic welfare. This framework will allow us to understand the effects of algorithmic gatekeepers on epistemic welfare. The new interdisciplinary research program will build on multi-method empirical research applied to algorithmic gatekeepers. Thanks to the theoretical and empirical findings, we will be able to develop normative instruments to align algorithmic gatekeeping with epistemic welfare. Taking the novel concept of epistemic welfare as a starting point allows all project partners to embark on a joint, integrated exploration of the concept. This may lead to a paradigmatic shift in the conceptualization of the impact of algorithmic gatekeepers in media sectors, proposing to expand the notions of media pluralism and consumer welfare to epistemic welfare. Multistakeholder challenges and opportunities of news recommender systems in newsrooms: a comparative systematic literature review 2026 Read You Might Also (Not) Like … : Newsroom Wariness to News Personalisation? 2026 Read Researchers on this project Aina Errando PhD Researcher Adelaida Afilipoaie Postdoctoral Researcher Heritiana Ranaivoson Research Professor Consortium partners Namur Digital Institute HUMAN-IST ReSIC Center de recherche Namur Digital Institute 1/6 Tags Artificial Intelligence, Platforms, Regulation, Media Policy Previous Next

< Back People, personalisation, prominence: A framework for analysing the PSM shift to digital portals and interrogating universality across contexts Iordache, C. et al. (2024). People, personalisation, prominence: A framework for analysing the PSM shift to digital portals and interrogating universality across contexts. International Journal of Cultural Studies. https://doi.org/10.1177/13678779241296556 In the context of enhanced platformisation, Public Service Media (PSM) are once again forced to rethink the ways in which they achieve core public values. To this end, PSM have been prioritising the development of their own video-on-demand portals. To contribute to ongoing research, we propose a theoretical framework that can be applied by future PSM work, based on the operationalisation of platformisation in PSM policy documents and strategy. We identify the shared priorities across ten media organisations in seven media markets: Belgium—Flanders and Wallonia-Brussels, Canada, Denmark, Italy, Poland, and the UK. The study is based on the qualitative analysis of 61 documents, outlining the PSM remit and how they report and present themselves to governments, collaborators, and audiences, contextualised by ongoing national and regional debates. Findings confirm that the framework of people, personalisation, and prominence can serve as a useful theoretical basis for understanding and interrogating universality across contexts. Public Service Media Previous Read the article Next

< Back Challenges and requirements for implementing due prominence: aligning stakeholder interests in Flanders Van der Elst, P., Afilipoaie, A., & Raats, T. (2025). Challenges and requirements for implementing due prominence: aligning stakeholder interests in Flanders. Frontiers in Communication, 10. https://doi.org/10.3389/fcomm.2025.1569132 Since the introduction of Article 7a in the Audiovisual Media Services Directive (AVMSD), which enables Member States to impose measures that ensure prominence of general interest services on user interfaces, discussions within media policy on the need of these measures have increased markedly. In Member States such as Germany, France, Italy and Ireland, commitments were already imposed on players offering user interfaces to audiovisual—and in some cases also audio—media services and content of general interest. This article presents the findings of a study examining the desirability and feasibility of implementing due prominence measures in Flanders (Belgium). The research adopts a media policy analytical approach and examines the (i) positions of the different stakeholders on the desirability and feasibility of a revised prominence regime, (ii) the extent to which interests across different stakeholders can be aligned and (iii) the considerations governments need to take into account when developing due prominence measures in Member States. The research is based on a stakeholder consultation conducted with broadcasters, television distributors, streaming services and hardware manufacturers operating in the Flemish market. The research is complemented with interviews with representatives of different market actors. The study identifies considerable oppositional logics between broadcasters and hardware manufacturers and raises challenges in defining the scope of ‘general interest’, the addressees of prominence measures, contextual factors such as markets size and path-dependency in media policy, and the commensurability between different existing legal frameworks at the national and EU level. The authors argue for clearer, more explicit guidelines at the EU level regarding the interpretation of Article 7a. Stakeholders Previous Read the article Next

< Back It’s (Not) All About That CTR: a Multi-Stakeholder Perspective on News Recommender Metrics Vandenbroucke, H., & Smets, A. (2024). It’s (Not) All About That CTR: a Multi-Stakeholder Perspective on News Recommender Metrics. In Proceedings of the 18th ACM Conference on Recommender Systems. Recommender systems are increasingly used by news media organizations. Existing literature examines various aspects of news recommender systems (NRS) from a computational, user-centric, or normative perspective. Yet research advocates studying the complexities of real-world applications around NRS. Recently, a multi-stakeholder approach to NRS has been adopted, allowing to understand different stakeholder perspectives on NRS development and evaluation within the news organization. However, little research has been done on the different key performance indicators (KPIs) and metrics considered valuable by different stakeholders. Based on 11 interviews with professionals from two commercial news publishers, this paper demonstrates that stakeholders prioritize distinct KPIs and metrics related to the reach-engagement-conversion-retention funnel. The evaluation of NRS performance is often limited to short-term metrics like CTR, overlooking the multiplicity of stakeholders involved. Our findings reveal how different purposes, KPIs, and metrics are valued from the journalistic, commercial, and tech logic. In doing so, this paper contributes to the multi-stakeholder approach to NRS, advancing our understanding of the real-world complexity of NRS development and evaluation. Recommender Systems, Newsmedia Previous Read the article Next