Strategy & economics
Cohort analysis
Also known as: cohort retention · retention cohort
Grouping customers by a shared start event (sign-up month, first purchase) and tracking their behaviour over time to expose retention, churn, and LTV dynamics aggregate metrics hide.
Cohort analysis groups customers by a shared entry event — usually sign-up month or first-purchase date — and tracks a chosen metric across each cohort over time. Unlike aggregate retention numbers, cohort retention exposes when customers churn (month 1 is usually the worst), whether retention is improving or degrading across cohorts (each new cohort's m3 retention should beat last cohort's m3), and where lifecycle programs deliver real lift. The canonical cohort chart is a triangle: rows are cohorts (by start month), columns are months since start, cells are percent still active. Aggregate retention numbers average across all cohorts and hide declining product-market fit. Cohorts reveal it. Every serious lifecycle program reports against cohort retention, not aggregate.