Define churn: what churn rate means, with customer and revenue churn formulas

Churn is the loss of customers or recurring revenue over a period: a customer who cancels, or doesn't renew, has churned. The churn rate is that loss as a percent of what you had at the start of the period, measured either in customers (customer churn) or in recurring revenue (revenue churn).

The two rates can tell very different stories, which is why you should track both.

Churn rate formulas

Customer churn rate   = customers lost in the period
                        / customers at the start of the period × 100

Gross revenue churn   = (churned MRR + contraction MRR)
                        / MRR at the start of the period × 100

Net revenue churn     = (churned MRR + contraction MRR − expansion MRR)
                        / MRR at the start of the period × 100

Three rules keep the number honest:

• Only customers who were paying at the start of the period are in the denominator and the numerator. New customers who sign up and cancel in the same month are a separate problem (activation), not churn of your base.
• Trials that end without a payment are not churn. They never paid.
• Net revenue churn can be negative. That happens when upgrades from customers who stay are larger than what leaving customers took with them.

Customer churn vs revenue churn: a worked example

On 1 September a company has 400 paying customers and $40,000 of MRR. During September:

Customer churn      = 12 / 400                          = 3.0%
Gross revenue churn = (2,400 + 300) / 40,000            = 6.75%
Net revenue churn   = (2,400 + 300 − 1,500) / 40,000    = 3.0%

Customer churn says 3%. Gross revenue churn says almost 7%, because the twelve customers who left paid $200 a month on average against a company average of $100. That gap is the signal: you are losing larger accounts, and the reason usually sits in onboarding, support or a missing feature that bigger customers need.

Voluntary and involuntary churn

Split churn by cause before you act on it:

• Voluntary: the customer chose to cancel. Exit surveys and cancellation reasons tell you why: price, missing features, switched to a competitor, stopped using it.
• Involuntary: a card was declined or expired and the payment was never recovered. Card updater services, retries and a polite email fix much of this without any product change.

Stripe, for example, records payment_failed as a cancellation reason on the subscription, and the reason a customer picks when they cancel, for example in the customer portal (too_expensive, missing_features, switched_service and others), so both splits come from data you already have.

A churn benchmark you can cite

SaaS Capital's 2025 retention survey, run in the first quarter of 2025 across more than 1,000 private B2B SaaS companies, put median gross revenue retention at 91% (December 2024 revenue from customers who were already customers in December 2023). SaaS Capital describes churn as 1 minus gross retention, so the median company lost about 9% of its starting revenue to cancellations and downgrades over that year. Companies with the largest contracts retained the most.

Common churn mistakes

• Adding new customers to the denominator. Using the end-of-month count makes churn look lower in a growing month.
• Multiplying monthly churn by 12. 3% a month is about 30.6% a year, not 36%, because each month's loss comes from a smaller base: 1 − 0.97^12 = 0.306.
• Mixing monthly and annual plans. Annual customers can only churn at renewal, so a blended monthly rate swings with the renewal calendar.
• Counting the cancel click instead of the end date. A customer who clicks cancel on 20 September but has paid through 14 October is still a customer in September. They churn in October, when the subscription ends.
• Hiding failed payments. Lumping them with voluntary cancellations sends you fixing the product when the problem is billing.

Related metrics

• Net revenue retention: NRR and GRR, the retention view of the same movements.
• Monthly recurring revenue: churned MRR is one of its four movements.
• LTV to CAC: customer lifetime is roughly 1 divided by the churn rate.
• ARPU: comparing churned ARPU with your average explains a gap like the one above.

Get churn automatically every month

The Stripe monthly churn report skill runs on the 1st. It rebuilds the paying base and its MRR as of the first day of last month, finds the subscriptions that ended during the month, and posts customer churn and revenue churn (canceled MRR over starting MRR) next to the month before, with payment_failed kept apart from customers who chose to leave and the cancellations grouped by the reason they gave.

For the involuntary side, Stripe failed payments lists each weekday the invoices whose payment failed, with the decline reason and the next retry, so you can chase them before they turn into churn.