Buyer personas from interviews, not assumptions: how to build one for an online store

The persona you have is probably a mirror

"Maria, 34, urban, values quality, active on Instagram." I have seen a version of this slide in dozens of e-commerce companies. It is not wrong, exactly. It is just useless for deciding what to write in an ad, which objection to answer on the product page, or which channel deserves next month's budget.

The reason is how it was built: a team in a room, agreeing on what they already believed about the customer. The output is a description of the team's assumptions with a name on it.

Eight to ten interviews change what you know

You do not need a large sample to change decisions. Eight to ten conversations with recent customers, 30 to 40 minutes each, are enough to surface the patterns that matter: what triggered the purchase, what alternatives they weighed, what almost made them leave, and how they describe the product in their own words.

Pick people who bought in the last 60 days so they remember the decision. Mix first-time and repeat buyers. If you sell in several countries, interview in each one; the trigger in Hungary is rarely the trigger in Romania.

What to ask, and what not to ask

Ask about the last purchase, not about preferences in general. People are accurate about what they did and unreliable about what they would do.

Useful questions: What was happening in your life when you started looking for this? What else did you consider? What made you hesitate? Where did you first hear about us? What did you expect that did not happen? How would you describe this product to a friend?

Avoid: Would you buy again? Do you like our brand? How important is price to you? These produce polite answers and no decisions.

Record the interviews with permission. The exact phrases customers use are your future ad copy.

From transcripts to segments

After eight interviews you will see two to four clusters based on the trigger and the decision path, not on age or income. In a supplements brand, the split was between people managing a specific health concern and people buying for general wellbeing. Same product, opposite messaging. Demographics did not separate them at all.

For each segment, write one page: the trigger, the alternatives considered, the top two objections, the channel where they first heard of you, and three verbatim quotes. That page is the persona.

Turning it into campaigns

Each segment gets its own ad angle, built on its trigger and its main objection. The landing page addresses that objection above the fold. Budget goes first to the channel the segment actually mentioned, not the one the agency prefers.

When a cosmetics client came with research already done by another firm, the strategy was written in two weeks, because the hard part was finished. When a client skipped research to save a month, the first quarter of ad spend became the research, at a much higher price.

When you do not need this

If your product is a commodity bought on price and availability, or if you already have solid research less than 18 months old, skip it. Update the existing work instead.

For everything else, this is what the consumer research service does: interviews, segments, personas, and the messaging map that comes out of them.