Churn measures loss over a defined period. Distinguish customer loss from recurring revenue loss, then state the starting population and calculation convention.
Calculate customer loss
A common convention divides customers lost during the period by customers present at its start. In a fictional example, four departures from one hundred starting customers produce 4% churn for that period. Newly acquired customers should not silently change the denominator.
Separate gross and net revenue loss
Gross revenue churn examines losses from cancellation or contraction. A net measure also considers expansion according to its definition. Strong expansion in a few accounts can therefore conceal many departing customers. Keep both views when they explain different product realities.
Inspect cohorts
Group customers by acquisition period or segment. An overall average can hide recent onboarding problems or a poor fit with an acquisition channel. Examine time to departure and known reasons without treating one survey answer as a definitive causal explanation.
Avoid period mismatches
Do not directly compare monthly and annual rates. Under a simplified constant-monthly-rate assumption, cumulative annual loss is 1 − (1 − monthly rate)^12. Real cohorts may differ substantially from that assumption.
What rate is acceptable?
There is no universal threshold independent of audience, contract and usage cycle. Start with a stable definition, then investigate changes and affected segments. A benchmark without matching context can encourage the wrong product decision.
Use a stable population
In a fictional month, 100 opening customers and five departures produce 5% customer churn for that population. Customers acquired during the month should not quietly enlarge that denominator. If the five departures account for $200 of $10,000 opening MRR, revenue lost to those cancellations is 2%. The measures describe the same period but answer different questions.
Specify whether cancellation counts when requested, when the paid term ends or when service actually stops. Treat test accounts, plan migrations and reactivations consistently. A stable definition is more useful than a flattering rate obtained by changing the population.
Read cohorts alongside reasons
Group comparable customer starts and observe retention over time. A blended rate can mix new accounts still discovering the product with established customers. For small populations, show absolute counts: one departure out of four is not the same operational situation as 250 out of 1,000.
Connect departures to observed use and known reasons without claiming every cause is established. “Too expensive” can mean low usage, mismatched expectations or a budget cut. These situations need different responses. Separate gross losses from expansion so upsells do not hide an underlying retention problem.
From this guide to your project
Growth ↗
A product finds its market when acquisition, activation and retention work together.
Tell us about your ideaExamples are educational and do not describe client results. Content prepared with AI assistance; see our editorial policy. Our method ↗