I recently stumbled across an article about personas written by Andrea Wiggins late last year in Boxes and Arrows. Wiggins does a nice job talking about how personas can help the design and development process, and some approaches for creating a good persona set. But what really gave me pause was the title: “Building a Data-Backed Persona.” Data-backed? Wait a minute is there any other kind?Well, yes, there are; we call them provisional personas at Cooper. These user models are based on internal assumptions and second-hand data rather than first-hand data gathered through ethnographic research techniques (most notably via in-person interviews with and observations of individual users). But in my worldview, provisional personas should always be the exception, not the rule. By definition, a “persona” implies that the archetypal user behavior patterns it describes are based on first-hand, observed behavior conducted in the field.
Wiggins’ article title was a good reminder to me that as personas have grown in popularity over the years, some fundamental principles have gotten lost in the translation. Basing design decisions on perceived user attitudes and behaviors—especially for products that target specialized roles and workflows—leads to second-guessing during the design process. Connecting the dots between research findings and persona descriptions reminds everyone about what personas are: shorthand for the key findings discovered during a real investigation.
Personas not based on real data aren’t really personas, just like Hostess twinkies aren’t really food. Sure they’ll fill you up, but in the long run those empty calories are a poor foundation to a healthy diet.