How to calculate customer lifetime value

customer lifetime value (LTV) tries to answer one question: how much is a customer worth over the whole time they stay with you? It matters because it's the only side of the ledger that makes a customer acquisition cost meaningful — spending $200 to acquire a customer is either a great trade or a loss, and you cannot tell which without an LTV number to compare it against. The trouble is that "LTV" gets calculated at least three different ways depending on the tool, and the differences are large enough to change a funding pitch or a channel decision.

Historical LTV: what customers have actually been worth so far

The simplest version looks backward: take a cohort of customers who joined in the same period, sum what they've paid to date, and divide by the number of customers. This is honest — it's not a forecast, it's an observation — but it systematically understates the value of recent cohorts, because customers who joined last month haven't had time to generate the revenue that older cohorts already have. Comparing a historical LTV number for a six-month-old cohort against a three-year-old cohort is comparing two different things.

A more useful historical formula ties LTV to margin and churn directly:

LTV = (average revenue per customer per period × gross margin) / churn rate

If a customer pays $50/month, gross margin is 80%, and monthly churn rate is 2%, LTV is (50 × 0.8) / 0.02 = $2,000. This formula assumes churn is constant over a customer's lifetime, which is rarely exactly true — churn is usually highest in the first few months and lower for customers who survive that period — but it's a reasonable approximation when you don't yet have enough history for a full survival-curve model.

Predictive LTV: modeling what customers will be worth

Predictive LTV fits a model — often a probabilistic model of purchase frequency and churn, sometimes a simpler regression on early behavioral signals — to project forward from a customer's first weeks or months of activity, rather than waiting years to observe their full lifetime. This is the only practical option for a growing business that needs to make acquisition decisions on customers who joined a month ago, but it is a forecast, with a forecast's uncertainty, and it should be reported with a confidence range rather than a single number wherever the tooling allows it. Predictive models trained on your early cohorts can also be systematically wrong if the behavior of early adopters differs from the customers you're now acquiring at scale — validate the model against a cohort old enough to check it, not just against the data it was trained on.

LTV only means something next to CAC

LTV in isolation is an interesting number; next to customer acquisition cost it's a decision. The LTV:CAC ratio — commonly cited with 3:1 as a rough, widely used rule of thumb for a healthy SaaS business — tells you whether the unit economics work at all, but it hides timing: a 3:1 ratio achieved over five years is a much worse business than the same ratio achieved in one year, because of the cash tied up in between. The CAC payback period — how many months of gross margin it takes to recover the acquisition cost — is the number that actually captures that timing risk, and it's usually the more actionable of the two for deciding whether to increase spend on a channel now.

Retention metrics that sit next to LTV

Two related numbers are worth calculating alongside LTV, because they explain why it moves: gross revenue retention measures revenue kept from existing customers before any upsell (a ceiling on how good retention can look), and net revenue retention adds expansion revenue back in, which is why a SaaS business can post net revenue retention over 100% even while losing some customers outright — the ones who stayed grew enough to cover the ones who left. If LTV is rising, check whether it's because retention genuinely improved or because expansion revenue from a shrinking set of accounts is masking churn.

Where the calculation breaks

• Averaging LTV across very different customer segments hides the fact that a self-serve customer and an enterprise account have nothing in common as a "typical" customer — segment before averaging, or the number is meaningless for any single decision.
• Ignoring gross margin and using revenue instead of margin-adjusted value. A high-revenue, low-margin customer can be worth less than a lower-revenue, high-margin one; LTV built on top-line revenue overstates the first and understates the second.
• Comparing LTV across tools with different churn or expansion definitions. Two products connected to the same billing system can report different churn numbers depending on how each classifies trials, downgrades and plan changes — know which convention you're reading before you compare periods or vendors.

Tools that calculate this from your billing data

Rather than building the formula from scratch against a raw billing export, most subscription businesses compute LTV from a tool that reads directly off the payment processor:

• A single Stripe-style billing system, wanting an out-of-the-box dashboard. Baremetrics connects directly and computes MRR, churn, LTV and cohort retention with no SQL required.
• More than one billing system to reconcile into one number. ChartMogul normalizes across sources like Stripe, Chargebee and Paddle so churn and LTV aren't calculated three different ways.
• Want a free starting point before committing to a paid tool. ProfitWell Metrics computes the same core metrics at no cost, with paid retention tooling available separately.
• B2B contracts with multi-year terms and usage components, where a simple monthly-churn formula doesn't fit. Subscript reconciles CRM and billing data for exactly this kind of contract structure.

Questions to ask before you trust a number

• Is this LTV historical (observed) or predictive (modeled) — and if predictive, has the model been checked against an older cohort?
• Does the LTV figure use revenue or margin? If revenue, what's the margin adjustment we'd apply ourselves?
• What counts as churn in this calculation — does a downgrade or a paused subscription count, and does that match how we think about it?
• Is LTV being compared against CAC and payback period together, or just the ratio in isolation?

Common mistakes

• Quoting a single company-wide LTV number when the business has customer segments with genuinely different economics.
• Treating a predictive LTV model's output with the same confidence as an observed historical number.
• Using revenue-based LTV to justify spend decisions that should be margin-based.
• Letting rising net revenue retention hide declining gross revenue retention — check both.

For the retention analysis that LTV calculations depend on, see how to run a cohort retention analysis. Browse SaaS and subscription metrics tools for the full field.