Churn rate
Retention metric. Do they keep coming back?
App churn rate is the share of users or subscribers you had at the start of a period who are gone by its end: customers lost ÷ customers at the start × 100. Losing 800 of 10k starting subscribers in a month is 8% monthly churn, which compounds to about 63% a year, not 96% (example inputs).
What is churn rate?
Churn rate is the share of users or subscribers at the start of a period who are lost by its end.
Voluntary churn: a customer choosing to cancel.
Involuntary churn: a customer lost to a failed payment or expired card without choosing to leave.
How to calculate Churn rate
Churn rate = Customers lost in the period ÷ Customers at the start of the period × 100- Customers lost
- starting customers who cancelled, lapsed or went inactive during the period
- Customers at the start
- customers at the beginning of the period; new customers added during it are left out
Where definitions differ
| Variant | Formula | When to use it |
|---|---|---|
| Subscriber churn | Cancelled or not renewed ÷ Subscribers at start × 100 | For subscription apps. |
| Activity churn | Users inactive for n days ÷ Active users at start × 100 | For free apps with no cancellation event; the n is your choice. |
| Involuntary churn | Lost to failed payments ÷ Subscribers at start × 100 | Report separately from voluntary cancellations. |
| Monthly to annual | 1 − (1 − monthly churn)^12 | Monthly churn compounds; it does not multiply by 12. |
App churn rate worked example
Example inputs, not a benchmark. A made-up subscription app over one month, to show the method.
- The app starts the month with 10k paying subscribers.
- During the month 800 of them cancel or fail to renew, and 1.5k new subscribers join. The new joiners are left out of this month's calculation.
- Monthly churn = 800 ÷ 10k × 100 = 8%.
- Annual churn = 1 − (1 − 0.08)^12, which is about 63%, not 8% × 12 = 96%.
- At a constant 8% a month, the average subscriber stays about 1 ÷ 0.08 = 12.5 months.
For a free app with no cancel event, swap "cancelled" for "inactive for n days" (activity churn) and keep the same n in every report.
Churn rate calculator
Example inputs: change them to your own numbers. The examples only make the calculator show a worked answer; they are not benchmarks or targets. Runs in your browser; nothing you type is sent anywhere.
- Monthly churn
- 8%
- Annualised churn (compounded)
- 63.2%
- Expected lifetime (months, at constant churn)
- 12.5 months
How do you measure churn for an OTT app?
For subscriptions, count subscribers at the start of the month who cancelled or failed to renew by its end, and divide by the starting count. Report involuntary churn from failed payments separately, because it needs a billing fix, not a content fix.
How Churn rate differs by vertical
| EdTech | Subscription churn spikes after exams and at the end of terms. |
|---|---|
| Fintech | Activity churn (no transaction in 30 or 90 days) is more common than cancellations. |
| OTT and streaming | Subscriber churn after a big release ends is the classic pattern. |
| Gaming | Activity churn on days since last session; players rarely cancel anything. |
Common mistakes
- Including customers acquired during the month in the starting count.
- Multiplying monthly churn by 12 to get annual churn.
- Changing the inactivity window for activity churn between reports.
Questions
How do you calculate monthly churn?
Divide the customers lost during the month by the customers you had at its start and multiply by 100. Losing 800 of 10k starting customers is 8% monthly churn.
How do you turn monthly churn into annual churn?
Annual churn is 1 minus (1 minus monthly churn) to the power of 12. 8% monthly churn is about 63% a year, not 96%.
Is churn the opposite of retention?
Roughly, but they are measured differently: retention follows an install cohort by day, while churn usually measures a customer base over a calendar period.
What is the difference between a metric and a KPI?
A metric is any measured quantity; a KPI is a metric a team has chosen as a target for judging performance. Most metrics here become KPIs only for the teams whose work moves them.
Do I need an MMP to measure these metrics?
Not for all of them. Engagement and monetisation metrics come from product analytics and billing data, while metrics split by ad channel, such as CPI, ROAS or retention by source, need installs attributed to campaigns.
Are the calculator's example inputs benchmarks?
No. The example inputs only make the calculator show a worked answer when the page loads; they are not typical, good or target values. Replace them with your own numbers.