top of page
2 days ago
3 min read

How do I build ecommerce personas that change what customers see?

By Trayan Hristov · Founder, Thrivelia · 19 November 2026

Imagine two people buying exactly the same trousers. One has finally found a cut that works and wants another colour. The other needs them for a wedding on Saturday. Put both in “female fashion buyer” and you still have no idea what to show either of them. I’d build the persona around the decision each person is trying to make.

Start with the moment where you could be wrong

Pick a first product with enough orders, returns and customer questions to study. Read the product reviews alongside support tickets, exchange reasons and a handful of voluntary post-purchase answers. Include people who kept the product, people who returned it and people who never bought. A five-star review tells you what landed for a buyer. It can’t tell you why someone left the size selector.

I’d ask the team to mark each observation as stated, observed or inferred. “I need these for Saturday” is stated. Choosing express shipping is observed. Calling that person an occasion buyer is an inference. If the only signal is shipping speed, show an accurate delivery estimate and let the customer tell you more. An inference is a poor reason to enrol someone in a permanent occasionwear segment.

Write a persona that earns its place

For an illustrative apparel store I’d start with three working rows. Each one has a different next experience.

Customer job

Evidence strong enough to act on

What changes

Repeat a fit I already trust

Customer confirms the last pair fit and kept it

Show that cut and inseam in available colours.

Get an outfit ready for an event

Customer selects an event date or delivery deadline

Show arrival dates and pieces that complete that outfit. Retire the date after the event.

Get the first purchase sorted

Open exchange or unresolved fit question

Route to service and pause fit-based recommendations until resolved.

A person can move between those rows. If yesterday’s wedding guest comes back for everyday trousers, their current choice wins. If the size is out of stock, offer a stock alert or a genuinely comparable cut. A confident recommendation for an unavailable size spends the trust you just earned.

Decide which question is worth asking

Here’s the one I’d test on the order confirmation page: “What would make choosing your next pair easier?” Give a few useful options and a skip. Ask only if merchandising can act on the answer. Save the response with its date and product context. An explicit correction overrides a purchase inference. An exchange invalidates the assumption that the original size worked. Everyone who skips still gets useful care and fit information.

Have merchandising own the fit equivalences, customer service own the recovery rule, and CRM own when the branch starts and ends. Test against the existing journey among eligible first-time buyers. Read kept second orders per eligible buyer after your agreed window. Check returns and contribution alongside it. If confirmed fit changes attention but more buyers exchange the second pair, inspect the mapping before expanding it.

For the underlying method, see how we use progressive personalisation. For a public example where two buyers disagree about the very same pair, work through TALA’s flares.

Run the exercise

Use the Journey brief tab to fill in one persona with merchandising and service. Write down the experience a buyer who skips the question will receive. Download the editable workbook or browse the other retention workshops.

 
 
 

Comments


bottom of page