Choose the cohort and activity

Group customers by a meaningful starting event, such as first activation, and define what counts as later activity. Different customer workflows may have different natural frequencies. An occasional event-use customer should not automatically be judged against someone expected to use a service every week.

Preserve the denominator

Keep the original cohort size visible when calculating later participation. Explain exclusions and data gaps. Changing the denominator as customers disappear can make retention look stronger than it is and prevents readers from understanding the real scale of continued engagement.

Compare like-for-like groups

Look at customer segment, acquisition route and product conditions alongside the retention pattern. A new onboarding process may coincide with a different audience. Treat the pattern as evidence to investigate and test, rather than automatically attributing every change to the most recent product update. Keep reactivation visible rather than assuming customer behaviour follows a simple permanent decline. A person may miss one observation period and return later. Reporting the chosen activity definition clearly helps readers understand what the cohort table can and cannot establish.

Imagine this

Illustrative operating example

A hypothetical cohort begins with 100 activated customers and 40 meet the defined activity event in month three. That is 40% month-three activity against the original cohort. It does not mean the other 60 customers have permanently left.

Make it practical.

Use this checklist as you read. Checks are not stored.

Published by Tap Tap Go. Product availability and eligibility vary by programme and market. About our stories.