Definition
Cohort retention measures the share of players from a defined starting cohort (e.g., everyone whose first deposit fell in a given week or month) who remain active in later periods — day D+N, week W+N or month M+N. It is the standard framework for judging whether an operator keeps the players it pays to acquire.
Formula
iGaming retention comes in two distinct cohort types, and every retention number must declare which one it is:
User retention (M+N) = players with any activity in M+N ÷ cohort size at M+0
Deposit retention (M+N) = players with a successful deposit in M+N ÷ first-deposit cohort at M+0
User retention — cohort defined by registration or first activity; "retained" = any activity (login, bet). Deposit retention — cohort defined by first deposit; "retained" = made a successful deposit in the offset period.
The matrix form — rows as starting cohorts, columns as offsets (D+1…, W+1…, M+1…) — makes cohort-over-cohort trends visible: reading down a column shows whether product changes are improving each new generation of players.
How operators use it
| Offset | Use | Reporting requirement |
|---|---|---|
| D1 / D7, W1 | Fast feedback on early journey changes | Exact event and timezone definition |
| M+1 | First monthly maturity checkpoint | Cohort entry, retained event and eligible population |
| M+3 / M+6 | Validate longer-horizon retention and LTV assumptions | Mature cohort size, censoring and revenue basis |
Retention affects long-horizon LTV, but the size of that effect must be calculated from the operator's cohort model. Judge initiatives with a pre/post design and a control where possible; a change in blended activity can come from cohort age or acquisition mix rather than the intervention.
Never mix cohort types or activity definitions. A retention table that omits its starting event, retained event, offset logic and cohort size is uninterpretable, and any benchmark comparison built on mismatched definitions is noise.
Common misinterpretations
The cardinal error is comparing user retention against deposit retention as if they were one metric. A second is reading blended retention across mixed traffic. Third, percentages without cohort sizes and uncertainty mislead. Finally, judging a product change by same-month activity instead of subsequent cohort curves can confuse seasonality with impact.
Related metrics: Second Deposit Conversion · Churn Rate · Paying Active Users · LTV
Common questions
What is a good month-one retention rate for an online casino?
There is no defensible universal 40–55% range. M+1 changes with the cohort entry event, retained-activity definition, product, market tier, acquisition source and observation rules. Establish an internal baseline with cohort size and uncertainty, then compare cohorts using identical definitions.
What is the difference between user retention and deposit retention?
User retention can count any defined activity against a registration or activity cohort; deposit retention counts a successful deposit against a first-deposit cohort. Different numerators and starting populations make the percentages non-comparable. Neither is inherently always higher.
Which retention offsets should an operator track?
D1/D7 and W1 for fast feedback on onboarding and early lifecycle changes; M+1 as the headline retention KPI; M+3 and M+6 to validate 90- and 180-day LTV assumptions. Each horizon answers a different question — no single offset is sufficient.