How To Calculate Churn Rate

Churn rate is the share of customers who leave during a period, and the monthly version is simply customers lost divided by the customers you had at the start. Lose 48 of 1200 and the monthly churn rate is 48 / 1200 = 0.04, or 4%. The 1152 who stayed are the retained customers, which is the same fact written the other way round. One division is the whole calculation; everything else on this page concerns what that single number actually implies. It is among the cheapest metrics to compute and one of the most expensive to get wrong, because it feeds every downstream estimate of how long a customer stays and what they are worth.

Churn compounds, and that is the first surprise. A rate of 4% a month does not become 48% a year, because each month's loss is taken from a smaller base than the month before. The correct annual figure is 1 - 0.96^12 = 0.38729024267023293, so 38.729024267023293% of the starting customers are gone after twelve months, not 48%. The gap between 48% and 38.73% is not a rounding artefact; it is the arithmetic of a shrinking base.

The compounding runs through the retention factor. If 4% leave each month then 96% remain, so the retained fraction after twelve months is 0.96^12 = 0.6127097573297671. Slightly under 61.27% of the original customers are still there a year later, which means 1 - 0.6127097573297671 = 0.38729024267023293 have churned. Multiplying the monthly rate by twelve, the intuitive shortcut, ignores that every month the survivors are the only ones still exposed to churn.

Average customer lifetime follows directly from the monthly rate. If 4% of the base leaves each month, the expected lifetime is 1 / 0.04 = 25.0 months. The reasoning is that a steady churn rate of c implies an average of 1/c periods before a customer departs, which is why a small change in churn moves lifetime sharply: halving churn to 2% doubles the average lifetime to 50 months, while doubling it to 8% halves it to 12.5 months. The relationship is exact only while the rate is genuinely steady, so a business whose churn is falling month by month has a longer true lifetime than the current rate alone implies.

The denominator has to be stated, because start-of-period and average counts are not the same thing. Taking 48 losses against the 1200 you began with gives 0.04. Taking the same 48 against an average of (1200 + 1152) / 2 = 1176 gives 48 / 1176 = 0.040816326530612245, slightly higher. Both are legitimate conventions, but a rate without a declared denominator cannot be compared with anything, and mixing the two across months produces a series that quietly drifts.

Rates say nothing about scale, which is where churn figures mislead most reliably. Losing 60 of 1500 and 36 of 900 are both exactly 0.04 a month, yet the first business is shedding 60 customers and the second only 36. The identical percentage hides a 24-customer difference in absolute loss, and that difference is what decides how much replacement revenue the sales team must generate. Quote the absolute loss beside the rate whenever the base sizes differ.

Logo churn and revenue churn answer different questions and often disagree. Logo churn counts customers, so 48 of 1200 leaving is 0.04. Revenue churn weights those departures by what they spent, so losing 48 small accounts may cost only 2% of monthly recurring revenue while losing a handful of large ones costs far more. A business can hold logo churn flat while revenue churn worsens, or the reverse, and reporting only one of them hides half the story.

Churn is the mirror of retention, and the two must be reconciled before either is quoted. If retention is 96% then churn is 4%, but the identity only holds when both are measured over the same window and on the same population. Net revenue retention complicates this further, because expansion revenue from existing customers can push it above 100% even while customers are leaving. A 4% logo churn alongside strong upsells can coexist with net revenue retention of 105%, and neither figure is wrong.

Finally, churn is the input that sets customer lifetime value, so its precision matters more than its headline suggests. At a monthly churn of 0.04 the average lifetime is 25.0 months, and at that same 4% the annual churn is 0.38729024267023293 rather than the 0.48 a naive multiplication would give. Always record the period, the denominator and the definition you used, because a churn figure stripped of those three details is not a measurement but a guess. Two analysts can compute the same 0.04 from the same raw data and still disagree about what it means, purely because one measured over a calendar month and the other over a rolling thirty-day window.