Why two-stage churn is common
Many products lose a meaningful share of customers early (activation/onboarding), then settle into a lower steady-state churn rate among customers who achieved product-market fit. A two-stage model captures that pattern better than constant churn.
Model structure
- Early phase churn: higher churn for the first N months.
- Steady-state churn: lower churn after customers survive the early phase.
- Translate retention into value using ARPA and gross margin.
How to use it for strategy
- Improving early churn often has outsized impact because it increases the base that can expand later.
- Segment by channel/plan to identify cohorts with steep early drop-off.
- Use sensitivity: small changes in early churn can compound over 12-24 months.
Common mistakes
- Applying blended churn to all segments (hides pockets of weak retention).
- Assuming steady-state churn never changes (it can worsen with product changes).
- Ignoring expansion and contraction when revenue retention is the real driver.