Building a persona that actually helps you design
Florent Dabernat
April 18, 2026
2 min read
In this article
A home care booking app seeing usage drop after sign-up needs to understand who its users really are, beyond internal assumptions. Well-built personas exist exactly for that.
Collecting the right data
Quantitative collection relies on surveys, analytics and intake questionnaires. Qualitative collection relies on interviews, in-situation observation and shadowing. A point often overlooked: segmentation is best done by behaviors and motivations, rather than purely demographic criteria like age or occupation, which rarely predict actual usage behavior.
The structure of a persona sheet
A useful persona sheet includes a fictional photo, a first name, a signature quote that sums up their mindset, their goals, their frustrations, their digital skill level, a typical usage scenario and the everyday context in which they use the product.
The common mistake is stopping at the photo and the first name: it's the usage scenario and the context that actually make a persona actionable for design decisions, not decorative biographical details.
On the home care app case, twelve interviews made it possible to segment by behavior and land on two personas and one antipersona, leading to a mobile-first prioritization and built-in messaging.
Primary, secondary, antipersona
We distinguish the primary persona, who captures most of the product's value, the secondary persona, who influences the solution without being its central target, and the antipersona, who embodies an off-target profile that shouldn't be optimized for as a priority.
The antipersona is often the most underused of the three: it lets a team quickly settle prioritization debates by explicitly asking "are we designing for them instead of our real target?"
Concrete example: a primary persona "Léa," 29, self-employed nurse, with needs around reliability and time savings, using the app on mobile during her breaks. This level of precision, grounded in a real usage scenario, directly steers concrete design choices, unlike a generic sheet.
Frequently asked questions
-
How many personas should you create for a project?
- Two to three primary personas is usually enough, with an antipersona if the risk of mistargeting is high. Beyond that, personas become hard for the team to keep in mind collectively.
-
Does a persona have to be based only on real data?
- Ideally yes, but simulated data can serve as a starting point as long as it's later validated or adjusted through real interviews before major decisions rest on it.
-
What's an antipersona actually used for in practice?
- It acts as a guardrail during prioritization trade-offs: it helps the team spot when a proposed feature actually serves an off-target profile rather than the real priority users.