Churn Signal

A churn signal is an observable behaviour that recurs before someone stops using a product and marks a higher than average likelihood of leaving.

Common examples are a lengthening gap between visits, usage narrowing to a single feature, shorter sessions, and repeated encounters with errors. The relationship is correlational, not causal: most people who leave show the signal, yet most people who show it do not leave, and that asymmetry is why many alerts are false. The costly mistake is treating the signal itself as the problem, since sending notifications to raise login counts moves the number while leaving the underlying dissatisfaction untouched.