Cohort LTV forecasting: churn, expansion, discounting (practical model)

A practical guide to cohort-based LTV: why it beats simple churn formulas, how to choose assumptions, and how to interpret discounted LTV.

Updated 2026-01-23

Try it in a calculator

Why cohort-based LTV is worth it

The classic shortcut LTV ~ (ARPA * gross margin) / churn can mislead because it assumes constant churn, ignores expansion, and can explode when churn is small. A cohort model makes assumptions explicit and is easier to scenario test.

A simple cohort model

  • Start with ARPA and gross margin to compute monthly gross profit per account.
  • Apply a monthly retention factor (1 - churn) to model survival.
  • Apply monthly expansion to surviving accounts to model upgrades/seat growth.
  • Optionally discount future cash flows to compute discounted LTV.

Choosing assumptions

  • Logo churn: start with trailing monthly churn by plan/segment if possible.
  • Expansion: use observed net retention patterns (expansion often varies heavily by segment).
  • Discount rate: pick a consistent annual rate (e.g., cost of capital) if you want a present-value lens.

Common mistakes

  • Mixing time units (annual churn plugged into monthly churn).
  • Confusing logo churn with revenue churn (different denominators).
  • Using blended averages when segments behave differently (plan, channel, cohort).

FAQ

Should I use revenue or gross profit for LTV-
For unit economics decisions, gross profit is usually better because it accounts for COGS and variable delivery costs. If you use revenue, you can overstate value and set CAC targets too high.
How long should the horizon be-
Use 36-60 months for many SaaS products as a practical horizon. For high retention businesses, also run scenarios since long tails dominate the sum.

More in saas metrics

Cohort analysis playbook: retention curves, LTV forecasting, and payback
Cohort payback curves: how to model payback with early churn