Churn rate calculator

The churn rate measures the share of customers (or revenue) you lose over a given period, usually a month. There are two distinct churn figures, and they rarely tell the same story.

Customer churn (or logo churn) is the simplest: customers lost divided by customers at the start of the period. It treats a $19/month customer exactly like a $990/month one — which can be misleading if you're mostly losing small accounts.

Revenue churn corrects for that: it divides MRR lost by MRR at the start of the period. A revenue churn lower than customer churn is a reasonably good sign — it means the customers who left were, on average, smaller than your base. The reverse signals you're losing your biggest accounts first, which is more concerning.

Worked example: you start the month with 200 customers and $18,000 of MRR. Six customers leave, representing $450 of lost MRR. Customer churn is 6 / 200 = 3.0%. Revenue churn is 450 / 18,000 = 2.5%. Since revenue churn is lower, the customers who left were slightly below average ARPU — a fairly neutral signal.

The projection uses a compound-decay formula: if you lose a constant percentage p of customers every month, the number of months to fall to half your base is ln(0.5) / ln(1 − p). At 3% monthly churn, that's roughly 23 months — a rate that feels slow viewed one month at a time, but erodes nearly a quarter of the base every year.

Two common mistakes. The first: confusing monthly and annual churn. A 3% monthly churn doesn't equal 36% annually — compounding gives roughly 30% instead (1 − (1 − 0.03)^12). The second: calculating churn on too small a customer base or too short a period, which makes the percentage very volatile — losing 1 customer out of 10 looks like 10% churn without being a reliable signal.

When the metric stops meaning much: under about twenty customers, a single departure moves the percentage disproportionately — track the raw count of departures instead of the rate while the base is small. Churn calculated over a single month is also sensitive to one-off events; a 3-month rolling average smooths that out and gives a more reliable read on the real trend.