Creating More Personas Improves Representation of Demographically Diverse Populations: Implications Towards Interactive Persona Systems

Presentation at NordiCHI'22, by Dr. Joni Salminen
Systems mentioned:
Automatic Persona Generation (https://persona.qcri.org)
Survey2Persona (https://s2p.qcri.org)

Transcript

1. Creating More Personas Improves Representation of Demographically Diverse Populations: Implications Towards Interactive Persona Systems Joni Salminen University of Vaasa Lene Nielsen IT University of Copenhagen Soon-gyo Jung Qatar Computing Research Institute, Hamad Bin Khalifa University Bernard J. Jansen Qatar Computing Research Institute, Hamad Bin Khalifa University Presented at NordiCHI’22 Aarhus, Denmark
2. How many personas to create? Nobody knows. • Traditionally, only a few created because: • Manual methods pose a technical limit • End-user abilities to handle many personas pose a cognitive limit …however, (a) with data-driven persona development, the technical limit is removed by hyperparameter selection (b) with interactive persona systems, the cognitive limit could be removed by interactive features (sorting, filtering, searching)
3. Why even worry about the number? Because… 1. For personas to be data-driven, their number should also be data-driven. 2. More personas can enable more diverse view of the audience, including presentation for outliers/marginal groups that would otherwise be left unnoticed. 3. Theory-wise, we should know what the optimal number of personas is (and how to determine it), so that we can provide valid recommendations for persona creators in industry.
4. The past literature favors a small number
5. Personas represent distinct user types. However, while online user data can be demographically and behaviorally heterogeneous, most studies generate less than ten personas, regardless of how heterogeneous the data is.
6. Methodology and key findings • To address the question “How many personas to generate?”, we apply data-driven persona creation in a dataset with 250 million YouTube views from a global news and media organization. • We focus on a statistically optimal number of personas, namely, how the distribution of demographic persona attributes deviates from the baseline user data. • Altering the number of generated personas, ranging from 5 to 160 personas per set, we find that more personas cover more age groups and countries, thus improving the statistical correspondence with the raw user data, and increasing the representation of demographic diversity by including more fringe user segments. • While the user representation continuously improved with more personas, the relative diversity gain was maximal with 40 personas, implying that, using our data, one ought to create more than 4 times more personas than generally advocated.
7. Results
8. More personas than traditionally created are preferred from a diversity perspective.
9. Age diversification effect
10. Practical implications The results imply that organizations with heterogeneous online audiences benefit from many personas in terms of more inclusive user representation. We further demonstrate how an interactive persona system can help stakeholders navigate many personas with possibly smaller cognitive effort.
11. Features for coping with more personas
12. Thanks for your attention! Dr. Joni Salminen jonisalm@uwasa.fi ps. Reach out if you’re interested in collaborating!