Most lifetime value formulas quietly assume a subscription: a monthly fee, a predictable churn rate, a clean multiplication that produces a tidy number. A clinic, a retail shop, an agency running discrete projects — none of these have a billing system that quietly tells them when a customer is still active and when they've left. There's no cancellation event to log. There's just a customer who bought once, and either comes back at some point or doesn't, with nothing in the system marking the moment the relationship actually ended.
Why the subscription formula doesn't transfer
The standard formula divides average revenue per customer by a churn rate, because churn rate is precisely the piece of information a subscription business has and a project-based one doesn't. Importing the formula without that input means guessing at churn, which turns a supposedly rigorous calculation into a number dressed up as more certain than it is. The honest response isn't to force the formula to fit — it's to build a version around the data these businesses actually generate: purchase or booking dates, spaced out over real history.
Building it from repeat-purchase patterns instead
Pull every customer's transaction history and look at the actual gap between their first and second purchase, if they had one, and every purchase after that. This produces something closer to the truth than a formula ever will: an observed distribution of how long customers who do return typically take to come back, and roughly what proportion return at all within a reasonable window. A clinic might see that a meaningful share of first-time patients return within eighteen months for a related service. A retail business might see repeat purchases cluster at three, six and twelve months for different product categories. This pattern is the real foundation — not a churn assumption borrowed from a different kind of business.
Putting a number on it
Multiply the average value of a single transaction by the average number of transactions a customer completes over a realistic window — two years, three years, whatever the actual repeat pattern supports rather than an arbitrary standard. This produces an estimate, not a certainty, and it should be treated that way: useful for comparing whether one acquisition channel brings in customers who return more than another, and useful for judging whether a given customer acquisition cost is sustainable, but not a number precise enough to bet the whole plan on down to the decimal.
What changes once you have it
A customer acquisition cost that looks expensive against a single transaction can look entirely reasonable against a realistic two-year value, and a channel that looked cheap can look worse once it's clear its customers rarely return at all. This is often the actual answer to why a channel that produces a low cost per lead still isn't building the business the way a smaller, pricier channel is — the second one is bringing in people who come back, and the first isn't, and neither fact is visible from acquisition cost alone.
Our consulting team builds this kind of repeat-value estimate with clients specifically because most of the businesses we work with have no subscription data to lean on, and treating every acquisition channel as if the first purchase were the whole story leads to decisions that look right in the short term and wrong within a year. This connects to the wider thinking in our SEO and content strategy, where the same principle holds — a channel's real value is rarely visible from the first transaction it produces.
