What churn rate measures
Churn is the share of customers who stop within a defined period. Count who was active at the start, count how many had gone by the end, and express the loss against the starting group. The arithmetic is easy; deciding what “gone” means is not. A subscription churns when someone cancels, which is a clean event with a date attached. A shop, a clinic or an agency has no cancellation, so you have to choose a stretch of silence long enough to count as leaving — and that choice, far more than the sum, decides the answer you get.
Two versions are worth keeping apart. Customer churn counts people. Revenue churn counts money, and the two can move in opposite directions: losing a crowd of small accounts while your largest ones grow will raise customer churn and lower revenue churn in the same month. Email lists have their own version again, where unsubscribes, hard bounces and quiet disengagement shrink the audience you can actually reach even though nobody has stopped being a customer.
Why churn rate matters
Churn sets the value of every customer you buy. Customer lifetime value is, in effect, margin divided by churn: the faster people leave, the less any acquisition is worth and the tighter your cost per customer has to be. A poor cost per lead can be repaired in a month. A leaking bucket cannot be outspent.
It also drives an arithmetic that owners often meet the hard way. Each month you have to replace what you lost before a single new customer counts as growth, so a churn rate creeping upwards quietly raises the acquisition target without anybody deciding to raise it. That is usually experienced as marketing getting more expensive, when the real change happened after the sale rather than before it.
Common mistakes with churn rate
Measuring only at the whole-business level is the first. Churn is rarely spread evenly. It concentrates in a particular acquisition source, plan, product line or intake period, and a blended figure hides exactly the group you needed to see. Broken out by cohort, the problem is often that you are buying a kind of customer you should stop buying.
The second is treating every departure as dissatisfaction. Some churn is structural: the customer moved city, the project finished, the student graduated. No service recovery changes that, and retention effort aimed at people who were always leaving is expensive and demoralising. The third is watching cancellations only, while ignoring the earlier signals — falling usage, unanswered emails, a reorder that never came. By the time someone formally cancels, the decision is old.
How to act on it
Write the definition of “churned” down and keep it still, because most alarming swings between months come from somebody quietly changing the rule. Then split the figure by cohort and by where the customer came from. If churn is far worse from one source, the fix belongs in acquisition rather than retention: the targeting or the message made a promise the product does not keep.
For everything else, work at the start of the relationship rather than the end. Most avoidable churn is decided in the first weeks, when a customer either reaches the result they came for or quietly concludes they will not. A dependable onboarding sequence, a check-in before the first renewal and an obvious route to help will beat any discount offered after someone has already decided. For the group that drifted rather than decided, a re-engagement campaign is still worth running, and keeping that audience reachable is one of the quieter arguments for email marketing being owned rather than borrowed from a social platform.
