How to use this calculator
- Enter or choose cohort size at the start.
- Enter or choose still active at the end of each period.
- The retention at the last period updates as you type. Use Copy result or Share to keep it.
Method
Retention_t = A_t ÷ C; Churn_t = (A_{t−1} − A_t) ÷ A_{t−1}; Average churn = 1 − (A_k ÷ C)^(1 ÷ k)
- C
- cohort size at the start (A₀)
- A_t
- members still active after period t
- k
- number of periods
Worked example
- Retention: 800, 680 and 612 of 1,000 = 80%, 68%, 61.2%.
- Period churn: 200 ÷ 1,000 = 20%, 120 ÷ 800 = 15%, 68 ÷ 680 = 10%.
- Average churn = 1 − 0.612^(1/3) ≈ 15.10% per period.
After three periods 61.2% of the cohort remains, an average churn of about 15.1% a period.
What the result means
Falling period churn, as here, means the cohort is settling into a loyal core. A curve that flattens is a strong sign of product fit; one that keeps falling at the same rate is not.
Assumptions and limits
- The cohort is closed: reactivated and new members are excluded.
- Periods are equal in length.
- Covers one cohort; compare several cohorts to see whether retention is improving.
- Revenue retention (expansion and downgrades) is a different measure.
Questions
Why use a compound average churn?
Churn applies to a shrinking base. The compound rate is the constant churn that would leave the same number active after k periods, so it reproduces the result exactly.
How is this different from a retention-rate calculation?
A single-period retention rate starts from the customer count at the start and subtracts new customers. A cohort follows the same people across many periods.