Customer Retention Rate: Cohort Formula, Examples, and Churn Relationship
Shusaku Yosa
Customer retention rate measures the share of an opening customer cohort that still meets your retention condition at the end of a period. Calculate retained opening customers ÷ opening customers × 100. Adding new customers does not, by itself, improve this measure. Define the customer unit and retention condition before calculating the percentage.
Separate customer, usage, and revenue retention
An active contract, a meaningful product action, and retained revenue answer different questions. A customer can keep paying without using a product, or keep using it on a cheaper plan. Stripe's customer retention overview distinguishes retaining existing customers from acquiring new ones.
Measure | Retention condition | Denominator and limitation |
|---|---|---|
Customer or contract retention | An opening customer remains under a valid contract | Unique opening customers; do not mix customers with subscriptions |
Product usage retention | A user performs a defined return action after entry | Eligible users in a starting cohort; activity is not contract status |
Revenue retention | Revenue from the opening cohort is preserved | Opening revenue from that same cohort; define expansion and contraction |
Revenue retention can rise even when customer count falls, because remaining customers expand. As explained in Stripe's net dollar retention guide, new-customer revenue is separated from changes within the existing base. A customer-count percentage and a revenue percentage therefore need different interpretations.
Freeze the opening cohort
- Choose the observation period and endpoints, such as valid contracts at the start and end of a month.
- Save the opening customer IDs. Decide whether one company with several contracts counts once or several times.
- Check those same IDs against the closing retention condition.
- Classify new acquisitions, reactivations, pauses, and transfers separately rather than adding them to the opening denominator.
Document whether trials qualify, when a customer becomes paid, and how scheduled cancellations or overdue payments affect status. A changed rule can create an apparent improvement without any change in customers. If the opening cohort is empty, the percentage is undefined.
A fictional example: why subtracting acquisitions can fail
Suppose a fictional business starts a month with 100 customers. Of these, 90 remain at month-end. It acquires 20 new customers, but only 15 of them remain at the endpoint. There are no reactivations or transfers, so the closing total is 105.
Group | Initial count | Remaining at close | Treatment |
|---|---|---|---|
Opening customers | 100 | 90 | Denominator 100; numerator 90 |
New acquisitions | 20 | 15 | Exclude from opening-cohort retention |
All closing customers | 105 at close | 90 + 15 | Reconcile the two groups |
The correct retention rate is 90 ÷ 100 × 100 = 90%. Subtracting all 20 acquisitions from the closing 105 produces 85%. That shortcut subtracts five new customers who have already left. To use an aggregate shortcut, you must identify new customers still present at close and reconcile other movements. Tracking the opening IDs directly makes these assumptions easier to check.
A different example—150 of 200 customers acquired in a particular month still active at the same elapsed age—gives 75%. That is acquisition-cohort retention, with a different starting point. See the guide to user retention cohorts and observation windows when the question concerns return activity.
When does retention equal 100% minus churn?
The relationship holds when the same closed opening cohort is divided completely into retained and not-retained states at the same endpoint. In this example, 90 retained and 10 not retained give 90% and 10%.
A count of customers who cancelled at any time during the month may differ: someone can cancel and reactivate before close. Likewise, not returning on Day 7 is not the same as cancelling a subscription. Check the cohort, period, and status rule before subtracting one reported percentage from another.
Turn a change into a testable question
- Split comparable cohorts by acquisition month, plan, or onboarding route.
- Investigate low usage, discounts, contract restrictions, payment failure, and cancellation reasons separately.
- Specify the intervention, its cost, and the elapsed period over which retention and gross profit will be reviewed.
High retention alone does not prove satisfaction or higher profit. Fixed contracts and discounts can retain customers too. Avoid universal acquisition-versus-retention cost ratios or promised profit uplifts. Connect retention evidence to acquisition investment using LTV, CAC, payback, and cash-flow analysis.