Customer Churn Rate
The share of paying customers who stop paying in a month.
Put this metric to work: explore FlowMRR's MRR retention analytics to track revenue movements, retention and forecasts from your Stripe account, or see how FlowMRR compares as a Baremetrics alternative for Stripe.
The share of paying customers lost in a month, also called logo churn. FlowMRR counts people, not subscriptions or Stripe customer records: records that share an email are one customer, and only customers who pay something count. A customer churns in the month their last paying subscription ends; cancelling one of two subscriptions is contraction, not churn. The base is the previous month's closing count, so customers acquired during the month do not dilute the rate and customers who come back do not offset it. With no paying customers at the start, there is no rate.
Formula
Customer Churn = Customers who stopped paying this month ÷ Paying customers at the close of last monthWorked example
400 paying customers at the close of March; 8 of them stop paying entirely in April → Customer Churn = 8 ÷ 400 = 2.0% for April. Over a year at that pace: 1 − 0.98¹² ≈ 21.5%, not 24%.
Other definitions you will see
Subscriber churn counts subscriptions instead of people: a customer who cancels one of two subscriptions adds a churned subscription but no churned customer, so the two rates can differ. FlowMRR shows subscriber churn as its own figure on the Retention page. Other tools divide by the average number of customers over the month, or by customers at the start plus those acquired during it, which lowers the rate in fast-growing months; some leave failed-payment churn out of the headline figure.
Illustrative comparisons
B2B SaaS monthlyThese unsourced ranges are illustrative scenarios, not measured industry benchmarks, percentiles or recommended targets. Their qualitative labels are editorial interpretations. Compare your own cohorts using the same segment, period, currency and measurement basis.
Why it matters
Customer churn caps your maximum customer count. At 5% monthly churn, the base levels off around 1 / 0.05 = 20× the customers you acquire each month: you cannot grow past that ceiling without reducing churn or acquiring faster forever. A growth problem is often a churn problem in disguise.
Common mistakes
- Excluding involuntary churn (failed cards) from the headline number — that's a measurement choice, not a definition.
- Annualising monthly churn by multiplying by 12 (it's compounding: 1 − (1 − m)¹²).
- Mixing B2B and B2C benchmarks — a healthy B2C number is a death sentence for B2B.
Calculating it from Stripe data
Customer churn counts people, not subscriptions. Start from Stripe subscriptions and keep only paying ones: exclude trialing, incomplete, incomplete_expired and paused, and skip customers whose subscriptions bill nothing, such as a free plan or a 100% discount. Then group by customer. The same buyer can sit under several Stripe customer records, so grouping by email or company gives a truer count than raw customer IDs.
A customer churns in the month their last paying subscription ends. Cancelling one of two subscriptions is contraction: lost revenue, same customer. A past_due subscription is still a customer while Stripe retries the card; one that ends unpaid churns at the end of the last period it actually paid for, not when the dunning cycle finishes. A cancellation scheduled for the end of the current period is not churn yet; it churns when that period ends.
The denominator is the customers paying at the close of the previous month. Customers acquired during the month stay out of it, so a busy acquisition month does not dilute the rate, and customers who return after cancelling do not offset the month's churned count. For a stable figure, average the monthly rates over the last twelve complete months and leave the current month out: it is partial and reads artificially low. FlowMRR computes customer churn this way for its LTV, and reports subscriber churn, which counts subscriptions instead of people, as a separate figure.
Reading the result
Customer churn tells you how many relationships end, whatever their size. Read it next to gross MRR churn: high customer churn with low revenue churn means you are mostly losing small accounts; the reverse means a few large accounts are leaving, which is usually more urgent.
Split it into voluntary churn (the customer cancelled) and involuntary churn (payment failed until the subscription ended). Involuntary churn is fixed with payment recovery, such as smarter retries, dunning emails and card updates, not with product work. For voluntary churn, look at when customers leave: churn concentrated in the first months points to onboarding or to expectations set during the sale; churn spread across all tenures points to value, pricing or competition. A cohort table shows which pattern you have.
Frequently asked questions
What is a good customer churn rate for SaaS?
It depends heavily on who you sell to and how they pay: monthly plans for small businesses churn faster than annual contracts with larger companies. Compare with businesses that have a similar segment and billing mix, using monthly churn measured the same way, and annualise with compounding: 2% a month is about 21.5% a year, not 24%.
What is the difference between customer churn and revenue churn?
Customer churn divides lost customers by customers at the start of the period; revenue churn divides lost MRR by MRR at the start. In FlowMRR, gross MRR churn counts only customers who stopped paying, with downgrades reported separately as contraction; some tools include downgrades in gross MRR churn. Losing one large account barely moves customer churn but can move revenue churn a lot.
Does a failed payment count as churn?
Not immediately. While Stripe is retrying, the customer is still counted. If the retries fail and the subscription is cancelled or marked unpaid, the customer churns as of the end of the last period they paid for; that is involuntary churn.
Should customers who come back be subtracted from churn?
No. A customer who cancels and later returns counts as churned in the month they left and as a reactivation when they come back. Netting the two hides how many customers actually leave.