What we research: Here are the main topical areas (TAs) the Persona Team is pursuing:

• TA1: Personas and Social Impact (Social Good) — The key issue here is, “How can personas be of help in societal problems?”. These are sometimes called “socio-technical” or “wicked” problems and they usually involve technological and social components, so their solutions need elements from both sides. Socio-technical problems can range from politics (e.g., group polarization, fake news) to society and health (e.g., addiction, sickness). A particular example is sustainability: solving the sustainability crisis requires BOTH appropriate technology AND understanding of human behavior (personas could play a role in the latter. The “novelty” of this approach comes from the fact that user personas were traditionally developed for commercial purposes, to understand users and customers of products and services. But this line of research expands the application of personas into non-commercial, public good domains.
  • Related concepts in computational sciences: AI for good, data for good.
  • Related research: Guan, K. W., Salminen, J., Jung, S.-G., & Jansen, B. J. (2023). Leveraging Personas for Social Impact: A Review of Their Applications to Social Good in Design. International Journal of Human–Computer Interaction, 0(0), 1–16. https://doi.org/10.1080/10447318.2023.2247568
• TA2: Explainable Data-driven Personas — The key issue here is, “How can data-driven personas be explained to stakeholders in an effective way?”. Data-driven personas are often created using algorithms and systems with many steps that involved varying degrees of complexity. The purpose of explaining these steps is to increase persona users’ trust in and knowledge of data-driven personas AND the boundaries and risks of data-driven personas, so that the persona users’ (e.g., company decision makers) would not have too high or low expectations about personas and what they actually are and can accomplish.
  • Related concepts in computational sciences: Explainable AI (XAI), fairness, accountability, and transparency (FAccT).
  • Related research: Salminen, J., Santos, J. M., Jung, S., Eslami, M., & Jansen, B. J. (2019). Persona Transparency: Analyzing the Impact of Explanations on Perceptions of Data-Driven Personas. International Journal of Human–Computer Interaction, 0(0), 1–13. https://doi.org/10.1080/10447318.2019.1688946
• TA3: New User Segmentation Methods — The key issue here is, “How can customer and user segmentation algorithms be renewed and improved?”. The background is that the algorithms used for people segmentation (specifically clustering and data dimensionality reduction) were not originally developed for people segmentation, but for statistical simplification of data that serves typically either machine learning or statistical analysis purposes (in both, the purpose is to “find the signal” in noisy data). While there is some overlap between these purposes and purposes of people segmentation, a key distinction is that these traditional segmentation algorithms often simplify user representation unnecessarily much, meaning that they generate a handful of segments even from a very, very diverse people data. In the process, outlier behaviors and marginalized user segments are often lost. So, the idea is that we can develop new techniques that not only are more robust for the loss of diversity in user representation, but also have better explainability than the complex and sometimes intractable clustering and dimensionality reduction algorithms.
  • Related concepts in computational sciences: user segmentation, algorithm development.
  • Related research: Salminen, J., Mustak, M., Sufyan, M., & Jansen, B. J. (2023). How can algorithms help in segmenting users and customers? A systematic review and research agenda for algorithmic customer segmentation. Journal of Marketing Analytics. https://doi.org/10.1057/s41270-023-00235-5
• TA4: Immersive High-Realism (Deepfake and Generative AI) Personas — The key issue here is, “What does AI technology, including deepfake, Generative AI, and Metaverse (VR) imply for persona development and application?”. These technologies enable MUCH different personas than initially conceived in the late 1990s when personas were introduced as static profiles. While dynamism and interaction can be additional factors that support persona users’ experience and information gain from the personas, they may also evoke new interaction challenges in terms of perception, usability, and optimal design. We need to thus understand the trade-offs that the new interaction technologies pose to the art and science of user representation.
  • Related concepts in computational sciences: deepfake development, generative AI, LLMs.
  • Related research: Kaate, I., Salminen, J., Santos, J. M., Jung, S.-G., Almerekhi, H., & Jansen, B. J. (2024). “There Is something Rotten in Denmark”: Investigating the Deepfake persona perceptions and their Implications for human-centered AI. Computers in Human Behavior: Artificial Humans, 2(1), 100031. https://doi.org/10.1016/j.chbah.2023.100031
• TA5: Integrating Personas into (Semi-)Autonomous Decision Making — The key issue here is, “How can data-driven personas add value in information systems?”. Decisions about people are increasingly made in informations systems; these decisions are increasingly automated but they can also be semi-automated in which a human decision makers makes the “final call”, or completely manual, in which a human decision makers examines data or information about people and then decides. Information systems where personas could be of value include, for example, online advertising platforms, CRM systems, web analytics tools, and so on. However, it is not immediately clear how personas should be integrated into various user interfaces (UIs) AND decision-making processes taking place within organizations. Also, personas might play a role in helping to evaluate or predict the impact of decisions made in these systems on the people that are affected by said decisions. To build such functionalities, much research and thinking are needed.
  • Related concepts in computational sciences: decision support systems, information systems, online platforms, user interfaces.
  • Related research: Jansen, B. J., Jung, S., & Salminen, J. (2020). From flat file to interface: Synthesis of personas and analytics for enhanced user understanding. Proceedings of the Association for Information Science and Technology, 57(1). https://doi.org/10.1002/pra2.215
• TA6: Interactive Persona Systems — The key issue here is, “How do various stakeholders interact with personas in real-time to accomplish their professional goals?”. On the other hand, “How can we develop interactive persona systems to support those goals?”. The former question deals with understanding persona user behavior from a scientific perspective. The latter question deals with the practical side of developing better persona systems that help people in their jobs. This involves understanding the range of tasks that professional users engage in on a daily basis and “how people learn about other people”. An impactful perspective here is that when we can analytically measure user behavior in interactive persona systems, the patterns and observations we make can count toward scientific theory of persona user behavior. So, the work is not only empirical but requires conceptual effort as well.
  • Related streams: human-computer interaction, user analytics.
  • Related research: Jung, S.-G., Salminen, J., Aldous, K. K., & Jansen, B. J. (2025). PersonaCraft: Leveraging language models for data-driven persona development. International Journal of Human-Computer Studies, 197, 103445. https://doi.org/10.1016/j.ijhcs.2025.103445

Q: “Sounds interesting, but how do you actually carry out this research?”. We support interdisciplinary research and most typically do mixed-method studies combining quantitative and qualitative analyses. Our preferred choice is experimental research. We have done (and continue doing) in-person laboratory studies, onsite user studies, remote user studies, collecting data such as system logs, gaze interaction (eye tracking), and user perceptions and attitudes.

Q: “I like it and would like to know more about joining the team!”. If you are interested in joining the team as a PhD student, there is a Pre-PhD Task that can help you determine if “this is for you”. You may also contact Prof. Joni Salminen (jonisalm (at) uwasa (dot) fi) for more information.

NOTE: In addition to persona research, we also do research on educational technology, especially the use of AI in education (AIED). This research leverages Cipherbot, an educational AI chatbot.

Keywords: Data-driven personas, Human-computer interaction, Quantitative UX, educational AI