Define the customer and the order
Choose how customers are identified and which orders count. Decide how cancellations, exchanges and fully returned orders are treated. If the available source cannot reliably distinguish new buyers, state that limitation before producing a retention rate.
Align the observation window across cohorts. Comparing a newly acquired group with one that has had a full season to return can create a misleading impression of deterioration.
Investigate what happens in between
Review the experience after the first purchase: delivery, product satisfaction signals, lifecycle messages and relevant assortment. The reason for not returning is not automatically a missing discount or reminder.
Look at the initial purchase context. A buyer attracted by a specific promotion or occasion may have a different next need from another group. Treat these differences as investigation paths grounded in available evidence.
Prepare one relevant next step
A proposal might adjust message timing, improve useful post-purchase information or test a category-relevant follow-up. State the audience, evidence and intended outcome in a brief the team can review.
Evaluate over a window suitable for the buying cycle. Separate observed repeat purchase, campaign-attributed orders and any experimentally estimated additional effect. Those are different forms of evidence.
Avoiding a premature conclusion
A new September cohort has fewer repeat orders than a June cohort. Before changing a flow, the team compares the groups at the same age and checks promotional and product context. Only then does it decide whether the journey needs a new proposal.
Turn the guide into a useful review
- Define a first-time customer and a valid order.
- Use equal observation time.
- Review acquisition and product context.
- Distinguish attributed orders from additional effect.
Bring the question.
Keep the context.
Map the relevant sources and specialist workflow for your team.
Explore your stack
