How this instrument works
Churn rate measures how many customers a subscription business lost during a period against the count it had the moment that period opened. The denominator is a fixed snapshot taken at the start, not an average of several points across the period — a distinction that matters, because it lets a revenue-operations team read the rate in real time as cancellations land, without waiting until the period closes to compute anything. A signup that arrives on day twelve does not enlarge the base a day-two cancellation is measured against.
SaaS finance teams, customer-success leads and boards track this figure because subscription revenue compounds on retention. The golden case here is instructive: a company starting a month with 500 customers and losing 25 of them logs a churn rate of 5%. Treated as a one-off, 5% sounds tame. Held steady month after month, it is not — retaining 95% of a base thirteen or fourteen times running erodes it to roughly half its starting size, which is exactly why a subscription business watches a monthly figure that a retailer would barely notice.
The ratio is a logo count, not a revenue count, and it says nothing about who left. Losing 25 of 500 trial-tier accounts and losing 25 of 500 largest accounts produce the identical 5% reading here, even though one event barely dents monthly revenue and the other guts it — that gap is what a separate revenue-churn figure, weighted by contract value rather than headcount, exists to close. This ratio also nets nothing against new signups gained in the same window; a business can post a positive 5% churn rate and still grow, if enough new logos arrived to outpace the losses.
- Count cancellations and non-renewals for the period and enter that total as Customers lost in the period.
- Enter the headcount you had on day one of that same period as Customers at start of period.
- Read Churn rate, % — the instrument divides the loss count by the starting count and multiplies by 100.
- Recompute every period on the same length and the same starting-count convention, so one reading compares cleanly against the next.
Worked example — 25 lost from a base of 500
Take a subscription business that opens the month with 500 paying customers and logs 25 cancellations before the month closes. Enter 25 as Customers lost in the period and 500 as Customers at start of period, and the instrument divides 25 by 500 and multiplies by 100, returning a churn rate of 5%.
Five percent looks tame as a single reading until it compounds. Keeping 95% of the base every month for about fourteen straight months multiplies out to roughly 0.95 raised to the 14th power, near 0.49 — meaning the original 500 would fall to close to 245 if nothing offset it. That is why a steady 5% monthly figure reads as urgent inside a SaaS finance team, while the same 5% quoted as an annual rate would barely register.
Questions
What counts as a customer lost for this formula?
Any account that canceled or did not renew inside the period counts once, regardless of how much it paid or why it left. The formula treats a trial account that lapsed the same as an enterprise account that canceled — it counts logos, not dollars, so a revenue-weighted churn figure is a separate calculation layered on top of this one.
Is churn rate the same thing as attrition rate?
No, even though both divide a loss count by a population. Attrition rate on this site tracks employees who left against a company's average headcount for the period. This instrument tracks customers who canceled against the count at the very start of the period — a different population, a different event being counted, and a different denominator convention, so the two figures are not interchangeable.
Why divide by the starting count instead of an average headcount?
A starting count is fixed the instant the period opens, so a team can watch the rate climb as cancellations land through the period without waiting for it to close. Averaging in the ending count would let new signups mid-period enlarge the denominator and quietly flatter the rate, masking losses behind growth that happened at the same time.
Does a 5% monthly churn rate mean roughly 60% churn over a year?
No — multiplying 5% by twelve overstates it, because retention compounds rather than adding. Keeping 95% of a base twelve months running works out to 0.95 raised to the 12th power, close to 0.54, so annual churn lands nearer 46%, not 60%. The shrinking base each month means each later month's 5% loss is 5% of a smaller number.
Does this number account for new customers gained during the period?
No. This is a gross figure — only losses against the starting base appear in it, and new signups never enter the ratio at all. A business can report a positive churn rate here and still grow its customer count, provided new logos in the same period outnumbered the cancellations this formula counted.
Can two companies share a churn rate and still be in different trouble?
Yes, because this ratio counts customers, not the revenue attached to them. Losing 25 of 500 low-value trial accounts and losing 25 of 500 top-spending accounts both read as 5% churn here, even though the second event can remove most of a month's revenue while the first barely moves it — that gap is what a dollar-weighted revenue-churn figure is built to catch.
References
- U.S. Small Business Administration — Marketing and sales guide
- New York University Stern School of Business — Aswath Damodaran's valuation research
Read this first: This instrument shows arithmetic, not advice. Real offers add fees, taxes and terms that vary by lender and place — verify the figures against your actual paperwork before deciding anything.