Four bike shops in a row, and not one of them mentioned a woman’s frame.
I’d walk in, tell the guy behind the counter I wanted a road bike, and get steered toward the same unisex frame in a smaller size, like sizing down solved a geometry problem. After enough test rides that just didn’t feel right, a fifth shop finally handed me a women’s frame. It was like a light switch clicked. I found a bike that fit like a glove, and I walked out wondering why it took five tries to get there.
That question followed me straight into my UX/UI coursework, right into the unit on user personas. The premise is simple enough: define a target user, assign them demographics, design for that person. But the road bike industry already had a target user, and she wasn’t me. My Sprocket case study covers this in more detail, but the research is blunt about who the “average” road cyclist is assumed to be: a Caucasian male, average age 36. Four shops built their whole sales pitch around that persona, and none of them had a script for anyone outside it.
Since 2020, most companies have built out real DE&I functions, and it’s changed how I read a user persona. Assigning a fixed demographic to a target user can be a useful shortcut, but it can just as easily become a quiet filter for who gets excluded from the design process entirely.
Where the standard persona breaks
A demographic persona describes a market, not a need. “36-year-old male, college-educated, suburban” tells a marketing team who to advertise to. It tells a design team almost nothing about what that person is trying to accomplish or where they get stuck trying to do it. The bike shops I visited weren’t lacking women’s frames. They had them in the back. What they lacked was a persona that included me as someone worth selling to on the first visit.
That’s the failure mode of creating personas around who a company assumes its users are instead of what those users need. The persona becomes a filter applied at the sales counter, at the product roadmap meeting, at every decision point downstream, and every one of those decisions quietly narrows who the product is actually built for.
What inclusive design does instead
Yeti didn’t retire the persona. It widened the aperture. When Yeti launched, it built around a narrow outdoor-enthusiast persona: camo, desert tan, “tactical” everything, marketed hard to that one demographic. I watched that lineup shift over a few years into a wide range of colors and product lines built around different use cases instead of one assumed buyer.
The shift paid off in ways a demographic persona would have missed entirely. Chefs use Yeti coolers to run pop-up kitchens. Farmers haul produce to market in them. Offices hand out branded Yeti tumblers as standard swag. None of those users fit the original outdoor-enthusiast persona, and Yeti didn’t need to predict every one of them. It needed a persona built around a behavior, staying cold outdoors, instead of a demographic.
That’s the same shift I think about when I write up accessibility considerations as a UX designer. A persona built around a task or a friction point scales to people the original research never talked to. A persona built around age, race, and gender doesn’t scale. It just excludes precisely.
Where personas still earn their keep
A persona is a hypothesis, not a headcount. I’m not arguing to cut personas from the design process. A good persona is still one of the fastest ways to align a team around who they’re solving for and what that person is trying to do. The problem was never the persona. It was treating a demographic guess as finished research instead of a claim that still needs testing.
That testing is where UX interviews do the real work. Every assumption baked into a persona, what this person values, where they get stuck, what they’ll tolerate, should get checked against what real people say in an actual interview. If the interview data contradicts the persona, the persona is wrong, not the participant.
Reflective Coda
Build personas around what someone is trying to do and what’s stopping them, not around who they statistically tend to be. Write the persona, then go test it against real interview data before you design a single screen against it. If a persona would exclude someone with a legitimate need for the product, that persona was drawn too narrow in the first place. Fix the persona, not the exception.
I ride a bike that fits my frame now, built by a company that finally asked the right question. It took five shops to find one that saw me as a rider first and a demographic second. Good design shouldn’t take five tries.
