A Novel Approach To Validating UX Personas

The following are highlights of a case study presented by Zeina Khamis of Manulife and Juliann Ng of element54 at NEXT, the Insights Association’s first-ever virtual event.

Zeina began by saying, “If you want to know how the lion behaves, you don’t go to the zoo, you go to the jungle.” She wanted to study those who used her firm’s web portal.

Her key criteria were to validate disparate hypothesis by bringing enough sample to bear on the problem that the results would not be dismissed. The ultimate goal was to make the personas the foundations of their UX work and to be able to recruit by persona for future studies.

Ideation, empathy exercises, and other qualitative techniques had produced many different hypotheses from many different teams, with no alignment and a fear that they might be missing other potential criteria. Zeina said she wanted “a study with statistical significance; I have too many hypothesis I want to test and I have too many people attached to their hypotheses. Too many cooks in the kitchen!”

Julianne proposed doing a segmentation study with a cluster analysis, with one of the deliverables to be a typing tool to find these personas for future research. Zeina reminded her, “I want mini in budget but not mini in analysis!”

Remarkably, the team kept the survey really short, at 7 minutes, which they believe is essential for B2B research. Respondents did evaluate two long batteries of attributes.

The research produced three distinct UX personas. Surprisingly, these personas disproved the internal teams’ hypotheses and were not based on demographics or firmographics.

Instead the results different on psychographics and technology usage, on factors like how easily frustrated they became, whether they preferred navigation or search, whether they preferred text or graphics, and their core UX desire. The teams were very surprised that none of the hypotheses came out as differentiating. They had placed a lot of weight going in on the role of the company, decision making authority, age, gender and industry.

Element54 identified three distinct UX personas: Amelia (53%), Marie (24%), and Chantal (23%). (Each of the personas were named after women as women predominated among the professionals surveyed for this study.)

Amelia is super thorough but wants to be efficient; she has a high threshold for digital frustration (is not frustrated easily); thinks that more is better, and indeed is afraid of missing information; she wants as much detail as possible, and she is navigation dominant (rather than search dominate). From the perspective of content consumption, she will consume text, video, infographics – anything she is provided.

Marie is about supporting rather than automating an “old-school” process. She gets frustrated very easily and is not as well versed to worked online. She does not want extra bells and whistles, and she prefers navigation to search.

Chantal is the only persona to prefer search to navigation. She finds what she needs, then gets out.

Because none of the teams’ expectations were favored over any others (because all had their hypotheses disproved), and because of the extensive sample, “all the cooks became aligned,” Zeina said, and the personas “were adopted by management, not pushed back on.” The teams are using the personas in their work and are using the typing tool for recruiting personas based on who they design for. “It was a big win!”

Zeina concluded by saying, “If we can’t go to the jungle, and the zoo won’t give us the answer, perhaps a safari might!”

NewsTechniqueJeffrey Henning, PRC