18+ --:--:-- UTC

Cohort Analysis

By Vadym Zelinskyi

A method that groups players by a shared start point — usually sign-up or first-deposit month — and tracks each group's behaviour over time.

Definition

Instead of looking at all players at once, cohort analysis follows each intake group (a 'cohort') across its lifetime, so metrics like retention, deposits and revenue can be compared week-by-week or month-by-month from the same starting line. This separates the effect of when a player joined from broader trends, revealing whether product or marketing changes actually improved the experience for newer players. A classic output is a retention curve or triangle showing what percentage of each cohort is still active N months later. Operators also use cohorts to estimate lifetime value and payback period before a full lifecycle has played out.

Worked example

The January first-deposit cohort might retain 40% of players at month one and 18% at month six; comparing it with the April cohort shows whether a redesigned welcome journey improved early retention.

Why it matters

For a learner, it teaches you to compare like-with-like over time rather than averaging everything together. For a professional, it is the backbone of retention and LTV forecasting and of judging whether an intervention actually worked.

See it — predict, then watch

Compare two first-deposit cohorts — the lift only shows when you track them separately.

Cohort · comparing start-groupsJanApr
Retention by month for two first-deposit cohorts

A +8pp better welcome journey lifts month-6 retention from 9% to 12% — the difference only shows when you track each cohort separately instead of one blended average.

More interactive models in the Lab.

Further reading

Authoritative references for business & metrics — regulators, standards bodies and primary sources. Independent, not affiliated.