A cohort groups accounts around a shared starting event. It helps separate a group's behaviour from the effect of continuously adding new customers.
Choose a consistent origin
For a fictional reporting product, first useful activity may be more meaningful than an account created months before onboarding. Keep the definition consistent. A three-day-old customer cannot yet contribute a complete monthly-retention observation.
| Parameter | State explicitly |
|---|---|
| Origin | Signup, activation or paid start |
| Unit | User or organisation |
| Return | Activity that qualifies as usage |
| Interval | Week or month since origin |
| Exclusions | Internal and test accounts |
Interpret differences in context
One cohort may have received assisted onboarding or arrived through a different channel. Note product and packaging changes. An observed difference does not establish its cause, and small groups can produce unstable percentages.
Should the latest month dominate the review?
No. Compare cohorts at equivalent ages and retain counts alongside percentages. Investigate customers who remain subscribed but reduce meaningful activity, then connect the observation with support evidence and interviews. The chart should generate a precise question to investigate rather than an automatic conclusion about product quality.
Separate seasonal workflows
A product used for quarterly planning may appear inactive between planning cycles. Define return around its actual usage rhythm and compare equivalent periods. Daily activity is not automatically the right retention criterion for every English-language B2B product.
From this guide to your project
Growth ↗
A product finds its market when acquisition, activation and retention work together.
Tell us about your ideaExamples are educational and do not describe client results. Content prepared with AI assistance; see our editorial policy. Our method ↗