Why cohort-based LTV beats flat-churn LTV
Flat-churn LTV uses a single monthly churn rate (ARPU × margin ÷ churn) and assumes the rate is constant across customer age. It's a useful back-of-envelope, but it overstates true LTV by 25–40% for most SaaS — because retention curves bend. Customers who stick past month 3 are dramatically more likely to stick at month 12. Flat-churn ignores this.
The cohort method
Cohort LTV sums each cohort's actual retention curve, weighted by ARPU and margin, projected forward. The result is the dollar value a customer is expected to generate, given how the customer base actually decays — not how a uniform churn rate would predict.
Reading the heatmap
Each row is a monthly cohort. Each column is age in months. Darker accent = higher retention. Three things to look for: (1) whether newer cohorts retain better than older ones (product-led improvement), (2) the "leak month" where the steepest drop occurs (typically month 1–2 onboarding), (3) whether decay flattens after month 3 (signal of true product-market fit).
2024 SaaS retention benchmarks
- SMB SaaS — month 1 retention 80–90%, month 12 retention 50–65%.
- Mid-market — month 1 retention 92–97%, month 12 retention 80–88%.
- Enterprise — month 1 retention 97–99%, month 12 retention 90–95%.
When to trust this analysis
With fewer than 50 customers per cohort, retention curves are statistically noisy — single cancellation swings can mislead you. For low-volume SaaS, focus on the trend across cohorts (improving or worsening) rather than the absolute curve shape.