SaaS churn rate: what good looks like, how to calculate it and why your tools disagree

What is a good churn rate for SaaS?

There is no single number, because churn tracks what customers pay. The best public dataset on this is ChartMogul's SaaS Retention Report, built on anonymised data from more than 2,100 SaaS businesses, and its message is consistent: the higher the average revenue per account (ARPA), the better the retention.

• By price point: in the same report, top-quartile gross revenue retention is above 90% for ARPA over $500 a month, but only 60–70% for ARPA under $50.
• For mid-market B2B: SaaS Capital's 2025 benchmarks put median net revenue retention at 102% for private companies with annual contracts of $25k–50k (top quartile 111%). SaaS Capital treats churn as roughly one minus gross retention.
• As an alarm bell: David Skok, in SaaS Metrics 2.0, reads net revenue churn above 2% a month as a sign that something in the business is wrong.

The practical rule: compare yourself with companies selling at a similar price to a similar customer. A $9-a-month tool for freelancers measured against enterprise software will always look broken, and vice versa.

How do you calculate SaaS churn rate?

A hypothetical month. You start with 500 customers and $40,000 MRR. 10 customers cancel, taking $1,200 with them; others downgrade by $400 in total; expansion adds $2,000. Customer churn is 10 ÷ 500 = 2%. Gross revenue churn is ($1,200 + $400) ÷ $40,000 = 4%. Net revenue churn is ($1,600 − $2,000) ÷ $40,000 = −1%: negative churn, and a monthly net revenue retention of 101%.

Notice how the same month reads three ways. The leavers were bigger than average, so revenue churn is double customer churn; expansion then more than covers the loss. Report only one of these and you are telling a partial story.

How do you convert monthly churn to annual churn?

Not by multiplying by 12. Churn compounds, because each month you lose a share of an already smaller base. The conversion is annual churn = 1 − (1 − monthly churn)^12: 1% a month is about 11.4% a year, 2% about 21.5% (Skok rounds it to roughly 22% of revenue), 3% about 30.6% and 5% about 46%.

Stripe's documentation shows the same effect on a cohort: at 10% monthly churn, 100 subscribers who joined together become 90 after a month, 81 after two and about 59 after five. It is also why churn dominates lifetime value: in the common formula LTV is monthly margin divided by monthly churn, so doubling churn halves LTV. The LTV:CAC ratio inherits every error in your churn estimate.

Why do Stripe, your spreadsheet and your analytics tool disagree?

Because each makes different choices about the denominator and the edge cases. Stripe's Billing analytics divides subscribers churned in the last 30 days by subscribers 30 days ago plus new subscribers in that window, which gives a lower figure than the textbook formula. It counts a past-due subscription as still active and treats it as churn only when it is cancelled or marked unpaid. And if you subtract discounts from MRR, a customer on a 100% coupon counts as churned until the coupon goes.

Cohorts add one more fork. In his cohort-analysis spreadsheet guide, Christoph Janz, co-founder of SaaS investor Point Nine Capital, takes a cohort that has shrunk from 100 customers to 50 and loses 5 more in month seven: that is 5% churn against the original cohort, or 10% against the previous month. Both are valid; they answer different questions.

None of this is wrong. What is wrong is switching definitions halfway through the year, or comparing your number with a benchmark built on another one. The same discipline applies to revenue figures; see MRR vs ARR.

What is involuntary churn, and how do you reduce it?

Involuntary churn is customers lost without deciding to leave: an expired card, a declined payment, a subscription that lapses on its own. Measure it separately, because the fix is operational rather than product work.

1. Retry failed payments automatically. Stripe's Smart Retries pick the best moment to try again; the recommended default is 8 attempts within 2 weeks.
2. Ask for a new card before the old one expires, and make updating it a one-click job.
3. Decide what happens when retries run out. Cancelling, pausing or leaving the subscription past due changes when, and whether, the customer shows up as churn.

Voluntary churn is a different animal. Heavy cancellations in the first months usually mean the product is not yet essential to anyone, which is a product-market fit question before it is a retention one.