Glossary - Customers and retention
Buyer persona
A buyer persona is a named, evidence-based profile of one group of customers who buy for similar reasons: who they are, what they need, what they buy and how they decide. Ecommerce teams use personas to brief creative, choose audiences and plan the range, and the best ones are built from order data and ranked by the margin each group earns.
| Variable | Definition |
|---|---|
| Share of customers | Customers assigned to the persona ÷ all customers in the period. |
| Share of total margin | The persona's summed margin LTV ÷ margin across all customers. Margin, not revenue, so discount-heavy groups are not overvalued. |
| Index | 100 means the group earns its headcount share. Above 100 over-indexes on value; below 100 under-indexes. |
A persona is a description, so there is no single formula for it. What can be measured is whether it holds up: that the group exists in your data at a meaningful size, that its behaviour differs from the other groups, and that it is worth something. The margin index is the simplest check of the third.
Buyer persona examples
Three illustrative personas for a skincare brand, built the data-first way: behaviour from orders, then the reason behind it.
| Persona | Who they are and why they buy | What they buy and how to reach them |
|---|---|---|
| The routine replenisher | Has a set routine and reorders when a product runs out. Values consistency and not having to think about it. | Full-size moisturiser and cleanser, every 8 to 10 weeks. Timely reorder reminders and subscribe options; rarely needs a discount. |
| The gift buyer | Buys for someone else, mostly around holidays. Wants something that looks considered and arrives on time. | Gift sets and bundles, higher AOV, one order a year. Gift guides, clear delivery cut-offs, wrapping. |
| The discount sampler | Tries many brands and moves on. Responds to deals, not the product story. | Travel sizes and sale items, first order with a code. Cheap to acquire, rarely returns; do not build acquisition around them. |
Worked example
The largest persona by headcount is the least valuable, and a fifth of customers carry more than half the margin. A persona exercise built from surveys and headcount would brief creative at the sampler, because that is who shows up most. Built on margin, the same exercise says find more replenishers, which is exactly the job of finding more of your best customers. This concentration is what value concentration measures across the whole base.
What makes a good buyer persona?
A good persona is specific enough to change a decision. It is grounded in what customers did, not only what they said in an interview. It is large enough to target and distinct enough that a different message or product would suit it. It carries a value, so you know which persona to grow and which to serve without chasing. And it is refreshed, because the customer mix drifts as channels, range and prices change.
A persona that reads "women 25 to 45 who love self-care" fails every one of those tests. It could describe almost any brand, and it gives a creative team nothing to act on.
Buyer persona vs related concepts
| Concept | What it tells you | How it differs |
|---|---|---|
| RFM segmentation | Customers grouped by recency, frequency and value | Pure behaviour, no motive. Personas add the why and make the group briefable. |
| Cohort analysis | How groups acquired at the same time behave later | Groups by start date; a persona groups by need and behaviour across dates. |
| Ideal customer profile | The type of customer worth pursuing most | A target; personas describe all the real groups you have, good and bad. |
| Customer lifetime value | What a customer is worth over time | The number that ranks personas; the persona says who sits behind it. |
Common mistakes
- Writing personas from imagination. A persona invented in a workshop describes who the team hopes buys. Start from orders and let the groups emerge.
- Ranking personas by revenue. Discount-heavy groups can lead on revenue and trail on margin. Rank by margin LTV.
- Stopping at demographics. Age and location rarely explain why someone buys. Motive, occasion and product path do.
- Making too many. Ten personas become ten briefs nobody uses. Three to five that genuinely differ are enough for most brands.
- Never connecting them to spend. A persona that never becomes an audience, a brief or a product decision is decoration. Tie each one to new-customer CAC and the channel that finds it.
FAQ
A name, the need or occasion that drives purchase, what they buy first and next, typical order value and frequency, the channels where they arrive, what makes them hesitate, and their value in margin. Demographics help targeting but should never be the whole persona.
Group customers by behaviour in your order data: first product, repeat pattern, discount use and value. Size each group and rank it by margin. Then add the motive with a short survey or a handful of interviews per group, and write it up as a brief the team can act on.
Usually three to five. Enough to reflect genuinely different reasons for buying, few enough that each gets its own creative, audience and offer. If two personas would get the same brief, merge them.
A segment is a rule that sorts customers, such as bought twice in 90 days. A persona is the human story behind a segment: who they are and why they buy. Good personas are built on segments, so they can be counted, targeted and measured.
Updated September 2026