What is churn?

Voluntary and involuntary churn

The distinction matters because the fixes are completely different, and involuntary churn is usually larger than teams expect.

Voluntary churn is a decision: the user no longer finds the product worth it. The remedy is product and value.

Involuntary churn is a failure: an expired card, a failed payment, a lapsed authorisation. The remedy is dunning, retry logic and payment method updating — which is unglamorous engineering that frequently recovers several percent of a subscription base.

Reporting them as one number hides the cheapest available improvement behind the hardest one.

Where churn actually happens

Overwhelmingly early. Most products lose the majority of users who will ever leave within the first days, and a much smaller trickle thereafter.

This is why activation matters more than any later intervention. A user who reached the moment the product became useful behaves differently from one who never did, and win-back campaigns aimed at people who never activated are attempting to recover users who were never acquired in any meaningful sense.

The corollary is that churn reduction budget is usually better spent on the first session than on the ninetieth day.

Reducing it

• Fix activation first. The largest single lever in almost every product.
• Handle involuntary churn properly. Retries, card updating, and a dunning sequence that is not hostile.
• Intervene on leading indicators, not on the churn event — declining session frequency precedes departure by weeks.
• Make cancellation easy. Hostile flows raise measured retention slightly and reputational damage considerably, and are increasingly regulated.
• Measure [re-engagement](/resources/glossary/re-engagement) causally, since campaigns to lapsed users over-claim badly.