The base churn formula
Churn rate is the share of customers, or revenue, you lose in a set period. The base formula is simple: take the number you lost during the period and divide it by the number you had at the start, then multiply by 100. If you began the month with 200 customers and 10 cancelled, that is (10 ÷ 200) × 100, or 5%. Count new sign-ups separately — never net them against losses, or a churn problem hides behind fresh growth. For the one-line definition, see churn rate.
| Measure | At start of period | Lost in period | Churn rate |
|---|---|---|---|
| Customers | 200 | 10 cancelled | 5.0% |
| MRR | $20,000 | $1,400 lost (cancels + downgrades) | 7.0% |
Customer churn vs revenue churn
Customer churn counts logos; every lost account weighs the same. Revenue churn weights each loss by dollars, so one enterprise cancellation can outweigh fifty free-trial drop-offs. In the table above the two rates diverge — 5% of customers left, but they took 7% of monthly recurring revenue with them, meaning your larger accounts churned. Track both. Customer churn tells you how many relationships you are losing; revenue (MRR) churn tells you what those losses cost.
Gross vs net revenue churn
Gross revenue churn counts only what you lose — cancellations plus downgrades (contraction) — divided by MRR at the start of the period. By definition it can never fall below zero. Net revenue churn subtracts what existing customers add back through upgrades and reactivations. When that expansion outpaces your losses, net churn turns negative, which ChartMogul describes as a strong signal that your existing base is growing on its own. Report gross to see the leak; report net to see whether expansion is patching it.
Voluntary vs involuntary churn
Voluntary churn is a decision — the customer chooses to leave. Involuntary churn is mechanical: an expired card, insufficient funds, a failed retry. This is the slice small teams miss, because it never arrives as an angry email. Paddle estimates failed payments cause as much as 40% of churn — treat that as a rough ceiling, not a fixed number, since it varies by pricing and payment mix. It is also the cheapest churn to recover: dunning and card-update prompts often win a large share back. Feed the rest into your win-back flows.
Why one number misleads, and what counts as good
A single period-wide rate mixes customers of every tenure into one figure, which flatters a fast-growing list — a flood of new sign-ups with low early churn can mask heavy loss among older accounts. A cohort-based view is truer: group customers by when they joined and watch each group decay. As for a good rate, it varies enormously by model, price point, and stage; sources like ChartMogul put healthy monthly net revenue churn in the low single digits as a rough reference — early-stage companies typically run higher — but there is no universal target to copy. Your own trend, watched by a Data Analyst, matters more than any benchmark.