Set the measurement boundary first
Cohort and data prerequisites
- Name the product, plan or customer group included in the review. (not complete)
- Record the opening date, closing date and billing cadence. (not complete)
- Define which cancellation, non-renewal, payment-failure and reactivation events count. (not complete)
- Choose whether new customers are excluded from the opening cohort or reported separately. (not complete)
- Reconcile opening customers and recurring revenue to the same billing records used for closing values. (not complete)
| Measure | Opening base | Loss included | Decision use |
|---|---|---|---|
| Customer churn | Customers active at the cohort start | Customers meeting the defined churn event | Review retention volume and replacement demand |
| Recurring-revenue churn | Recurring revenue from the opening cohort | Recurring revenue lost under the stated rule | Review contribution and plan-value exposure |
Review churn in a fixed sequence
- Freeze the cohort definition and period before looking at the result.
- Reconcile opening customers and revenue, then classify each loss, reactivation and timing adjustment once.
- Calculate customer and revenue churn using their own opening bases and units.
- Compare only with a period or cohort prepared under the same rules.
- Group observed reasons without treating missing or free-text evidence as a known cause.
- Choose one bounded action, owner, start date and success measure, then review contribution and service cost as well as churn.
Customer churn rate = customers lost from the opening cohort / opening customers
- customers lost
- Opening-cohort customers that meet the documented churn-event rule during the period (customers) — reconciled business record
- opening customers
- Customers active in the defined cohort at the start of the period (customers) — reconciled business record
Use the Subscription Churn Impact Planner for a comparable scenario and contribution interpretation. Do not silently move new customers into the denominator.
Worked cohort reconciliation
This neutral user scenario demonstrates the arithmetic; its values are assumptions, not a market benchmark.
A business starts a monthly cohort with 200 customers and 20,000 currency units of recurring revenue. During the month, 8 opening-cohort customers leave and their recurring revenue totals 1,200 currency units. New customers are reported separately.
8 / 200 = 4%
- 8
- Customers lost from the opening cohort (customers) — user scenario
- 200
- Opening customers (customers) — user scenario
1,200 / 20,000 = 6%
- 1,200
- Recurring revenue lost from the opening cohort (currency units per month) — user scenario
- 20,000
- Opening recurring revenue (currency units per month) — user scenario
The 6% revenue result being above the 4% customer result indicates that the lost customers had above-average opening revenue. It does not identify why they left.
Check mistakes and decision boundaries
- Counting a delayed renewal as churn in one period but not another.
- Combining voluntary cancellations with payment failures without preserving the distinction.
- Comparing customer churn with revenue churn as though the units were interchangeable.
- Treating a customer survey response, support contact or usage change as causal proof.
- Celebrating a lower churn percentage without checking contribution, discounts, support load or delayed cancellations.
Subscription churn review questions
- What is a good churn rate?
- There is no provider-neutral rate that is automatically good for every product, cohort, price point and billing cadence. Start with a reproducible internal baseline and test the contribution and operating effect of a change.
- Should new customers enter the churn denominator?
- For an opening-cohort measure, keep them separate so the denominator is stable. Another method can be valid, but it must be documented and used consistently in every comparison.
- Does a lower churn rate prove that a retention change worked?
- No. It is evidence of a change under the stated measurement rules, not causal proof. Check timing, cohort mix and other changes, then use a pre-defined follow-up measure and contribution boundary.
Methods used in this guide
- Margin101 Subscription Churn Impact methodology — Margin101: Definitions, units, assumptions and calculation boundaries for the linked planner.
- Margin101 Customer Retention Profit Impact methodology — Margin101: Contribution and programme-cost boundary for a retention scenario.