Customer cohort analysis for fashion retailers

Define who belongs to the cohort

Write down the first-purchase definition, valid order rules and customer identity used to form the group. Decide how guest checkout, cancellations and refunded first orders are treated in your analysis. A cohort definition must stay stable enough that another team member can reproduce the membership.

Shopify’s customer cohort report supports reviewing acquisition and retention with different metrics, intervals and first-order filters. Confirm the selected report configuration before exporting a result. A built-in report is a starting point for the business question, not a substitute for understanding its definition.

Source: Shopify: customer cohort reports ↗

Compare customers at the same age

A cohort acquired six months ago has had more time to purchase again than customers acquired last month. Compare a defined interval after the starting event and include only cohorts old enough to complete that interval. Leave immature observations visibly incomplete rather than entering a zero.

Then inspect the commercial context. A promotion-heavy cohort and a full-price cohort may differ in category, discounting and expected buying rhythm. Use those differences to frame the next investigation. Do not attribute a change automatically to the lifecycle programme when acquisition and assortment changed too.

Use a small set of complementary measures

Start with the question about repeat purchase, then add a clearly defined revenue view and the appropriate return context. Distinguish the share of customers who buy again from the number of orders they place. A few frequent buyers can move an average without representing most of the cohort.

For each measure, keep the denominator, time window and handling of adjustments visible. Review meaningful segments only when the sample and purpose support them. Avoid slicing the report into ever smaller groups until an attractive story appears; record the primary comparison in advance.

Prepare an action for a specific customer moment

If a comparable cohort has a weaker second-purchase pattern, investigate what customers bought, what became available next and which journey they experienced. This can support a brief for better product education, a relevant category introduction or a revised timing hypothesis.

Separate that analytical grouping from permission to contact customers. The lifecycle team checks eligibility and exclusions in its sending system. Evaluate the resulting change with an appropriate design rather than assuming that any later cohort improvement was caused by the new message.

Turn the guide into a useful review

1. Fix customer identity and valid first-order rules.
2. Use the same completed observation interval.
3. Read repeat purchase with product, revenue and return context.
4. Connect the finding to a specific journey hypothesis.