Cohort-based LTV builds lifetime value from evidence rather than assumption. Group customers by when they were acquired, track the gross profit each group produces every month, and read lifetime value off the cumulative curve at a fixed horizon.
Why the formula version breaks in B2B
The standard calculation divides ARPA by gross margin and a churn rate. It carries two assumptions that B2B data contradicts. The first is that churn is constant, when cancellations cluster around the first renewal date and then fall for the customers who survive it. The second is that one average customer stands in for the base, when a self-serve account and a six-figure enterprise contract behave nothing alike.
A single blended churn rate averages those populations into a number that describes no real customer.
How to build the curve
1. Group customers by acquisition month or quarter. 2. Record monthly gross profit for each cohort, including expansion and contraction. 3. Divide by the original customer count to get gross profit per acquired customer. 4. Accumulate the series to produce a cumulative curve. 5. Read LTV at a capped horizon rather than extending to infinity.
| Cohort | Cumulative gross profit per customer, month 12 | Month 24 | Month 36 |
|---|---|---|---|
| Q1 acquisitions | $9,400 | $17,100 | $23,600 |
| Q2 acquisitions | $8,200 | $14,900 | $20,300 |
What the curves reveal
Three things surface immediately and stay invisible in the formula version.
Whether expansion offsets logo churn, which appears as a cumulative curve that keeps rising after the first-year drop. Whether newer cohorts are worse than older ones, which usually traces to pricing changes or drift in who the company sells to. Whether one segment carries the entire average, which happens often enough that a company-level LTV should be treated as suspect until the segment curves are checked.
Pairing it with CAC
Compare each cohort's LTV against the CAC of the same acquisition period. That is the only version of the ratio where the numerator and denominator describe the same customers, which removes the timing mismatch that distorts blended comparisons in any quarter where spend and closings moved at different rates.
Cohort curves also give net revenue retention a value interpretation, since a cohort with NRR above 100% produces a cumulative curve that never flattens. Feed the result into the LTV to CAC ratio and the ratio stops being an assumption and starts being a measurement.
Frequently Asked Questions
How is cohort LTV different from the standard LTV formula?
The standard formula divides ARPA by a churn rate, which assumes churn is constant forever and that one average customer represents the base. Cohort LTV drops both assumptions and reads cumulative gross profit directly from what each acquisition group paid, month by month.
How far out should you extend the curve?
Cap the horizon at the longest period your own cohorts have actually been observed for. Extending a fitted curve beyond that produces a number no one can verify and inflates every ratio built on it. A capped LTV is smaller and defensible.
How many customers does a cohort need?
Enough that one departure does not move the curve. In enterprise, where cohorts are small, group by quarter or by half-year instead of by month, and read segment-level curves rather than company-level ones when deal sizes vary widely.
Does cohort LTV include expansion revenue?
Yes, and that is most of its value. Cohort curves capture contraction and expansion in the same series, so a base where expansion outruns churn shows a rising cumulative curve. The formula version cannot represent that at all.
Put these metrics to work
ORM builds custom revenue forecast models that turn concepts like cohort-based ltv into prescriptive action for your team.
Schedule a Demo