Key Takeaways

• Churn rate counts accounts, not dollars, so an enterprise cancellation and a trial-tier cancellation each register as one lost customer.

• The denominator is the customer count at period start, and the numerator counts only losses from that same group.

• Four defensible denominators applied to one identical month produce rates from 3.00% to 4.00%, which compound to 30.62% and 38.73% across a year.

• Annualizing by multiplying the monthly rate by 12 overstates the loss. Use 1 minus (1 minus monthly) to the 12th power instead.

• A benchmark is only comparable when it states its segment, contract length, denominator, and whether it counts payment failures.

What is the churn rate formula?

Divide the customers you lost during the period by the customers you had when it opened, then multiply by 100. The arithmetic takes four decisions:

1. Define the loss event. A cancellation, a non-renewal at term end, and a downgrade to a free plan each need an explicit yes or no.

2. Count the accounts on the first day of the period. That count is the denominator and it never moves.

3. Count how many of those same accounts left before the period closed.

4. Divide, multiply by 100, and label the result with its window. A rate without a stated period is meaningless.

Annualizing needs the compounding form: 1 minus (1 minus the monthly rate) raised to the 12th power. At 3% monthly that's 30.62% of customers lost across the year, not the 36% that multiplying by 12 implies, because each month's losses come off a smaller base than the last.

Which denominator and measurement window should you use?

Pick the start-of-period count with start-cohort losses only, because every other option lets acquisition volume distort a retention metric. Here's one month read four ways: 500 customers on day one, 60 signups during the month, and 20 cancellations, 15 from the opening base and 5 from those same-month signups.

The one-point monthly spread becomes eight points of annual difference. Teams that pad the denominator with new signups flatter their worst months, so define the choice in writing rather than accepting a dashboard default.

The window needs the same discipline:

• Monthly rates suit month-to-month plans, where a customer can leave on any renewal date.

• Annual cohort rates suit annual contracts, where nobody can churn for eleven months and a monthly rate reads as near zero until renewal season.

• Splitting voluntary losses from involuntary churn keeps a failed card from looking like a product problem.

What makes a churn rate benchmark comparable to your own number?

A benchmark applies to you only when it declares its segment, its contract length, its denominator, and whether it folded in payment failures. Strip any of those out and the number stops being a comparison.

Check these before quoting anyone's figure:

• Segment and average contract value, because self-serve and six-figure enterprise deals behave nothing alike.

• Monthly or annual framing, stated outright, since the compounding gap between the two can flip a verdict.

• Denominator definition, so you know whether mid-period signups sat inside or outside it.

• Sample size and collection window, so a figure from 40 companies in one vertical isn't read as an industry constant.

The Optifai pipeline study of 939 B2B SaaS companies, collected between Q2 2025 and Q1 2026, reports monthly logo churn of 3 to 5% for SMB, 1.5 to 3% for mid-market, 1 to 2% for enterprise, and under 1% for its top-performing cohort. Ranges like those tell you which order of magnitude you're in. They can't tell you whether your own 2.8% is healthy, because your denominator and contract mix aren't the study's. Pair the rate with MRR movements to see whether the accounts you lost were the ones that mattered.

Related terms

A churn rate is read alongside four or five neighbouring metrics, and each one answers a question this page doesn't.

• Revenue churn weights every loss by the recurring revenue attached to it instead of counting accounts.

• Involuntary churn isolates the cancellations caused by failed payments rather than by a decision to leave.

• Gross revenue retention states the same story as a retention percentage and caps at 100%.

• Net revenue retention counts upgrades from retained accounts, which is how it climbs past 100% while churn stays positive.

• MRR movements breaks a period's change into new business, expansion, contraction, churn, and reactivation.

• Subscription lifecycle defines the states a cancellation passes through before it counts.

FAQ

Is churn rate the same as customer attrition rate?

Yes, the two names describe the same calculation. Attrition rate is the usual phrasing in finance and HR reporting, churn rate in SaaS reporting. Some teams reserve attrition for annual figures and churn for monthly ones, but that's convention rather than a different formula.

How is churn rate different from revenue churn rate?

Churn rate weights every customer equally, while revenue churn weights each one by the recurring revenue it carried. A portfolio with a few large accounts can post a low churn rate and a painful revenue loss in the same month, so reports carry both figures.

Can churn rate be negative?

No. Customer churn rate has a floor of 0%, since you can't lose fewer than zero accounts. Negative figures belong to net revenue churn and net revenue retention, where expansion from retained customers can exceed the losses. Reactivated customers count as new business, not as negative churn.

Should paused or downgraded accounts count as churn?

Only if they've stopped paying you. A paused subscription with a scheduled resume date stays in the base and turns into churn only if that date passes without payment. A downgrade to a lower paid tier isn't logo churn, since the account is still there. Write both rules down, since inconsistent treatment across quarters is what makes a churn trend line lie.