Subscriptions
LTV (customer lifetime value)
LTV is the total contribution a customer is expected to generate over their whole relationship with a brand, best measured net of goods and fulfilment rather than gross.
Gross-revenue LTV is the common version and the misleading one, because it flatters subscriptions with thin margins. Contribution-margin LTV is what can actually be compared against acquisition cost.
It should be segmented by acquisition channel and by customer type. Subscription customers and one-time buyers have structurally different curves, and so do customers acquired through paid social versus email — blending them produces a number that justifies no specific decision.
The formula, and what each term hides
The standard identity is average revenue per period, multiplied by margin, multiplied by average customer lifetime — where average lifetime is taken as one divided by the churn rate. Written out with numbers it is unambiguous, which is both its virtue and its trap.
Take a subscription billing $45 a month at a 40% contribution margin, so $18 of contribution per renewal, losing 6% of subscribers each month. Average lifetime is one divided by 0.06, or 16.7 months. LTV is $18 divided by 0.06, which is $300.
Now change one input. At 8% monthly churn the same customer is worth $18 divided by 0.08, or $225 — a quarter less, from two percentage points of churn. LTV is more sensitive to the churn input than to anything else in the formula, which is why it is a useful management number and also why it is so easy to move by choosing a flattering churn definition.
The identity carries an assumption known to be false: it treats churn as constant across a customer's life. Real retention curves are steep early and flatten later, so a single blended rate applied uniformly undervalues the customers who survive the first months and overvalues those who do not. The two errors do not cancel in any predictable way, which makes the formula a serviceable estimate and a poor forecast.
LTV is meaningless without a stated horizon
One divided by churn is a sum to infinity. It assumes the business collects from a customer for as long as the arithmetic runs, which no business does — funding, inventory and planning all happen inside a finite window, and a number that ignores the window cannot inform a decision made inside it.
The correction is to quote LTV with the horizon attached: LTV at 90 days, at 12 months, at 24 months. For any cohort old enough those are measured rather than modelled, and they are what an acquisition decision is actually made against, because cash spent on advertising this month has to come back within a period somebody can name.
The gap is larger than it looks. Using the same example — 1,000 customers, $18 of contribution a month, 6% lost each month, the first month counted in full — the cumulative contribution per acquired customer is set out below. A business quoting $300 against a $100 acquisition cost is claiming a three-to-one return. The same business at a twelve-month horizon is at roughly 1.6 to one, and has to fund the difference for another two years. Both numbers are correct; only one of them is a plan.
| Horizon | Contribution per acquired customer | Share of the unbounded figure |
|---|---|---|
| 3 months | $50.82 | 17% |
| 12 months | $157.22 | 52% |
| 24 months | $232.05 | 77% |
| Unbounded ($18 / 0.06) | $300.00 | 100% |
Contribution-margin LTV versus revenue LTV
The largest single source of inflated LTV is using revenue where contribution belongs, and it is not a rounding difference.
Continuing the example: a $45 order costing $14 in goods, $8 in pick, pack and shipping, $1.60 in payment processing and $3.40 in a discount and returns allowance leaves $18 of contribution. Revenue LTV at 6% churn is $45 divided by 0.06, or $750. Contribution LTV is $300. Against a $100 acquisition cost the first says 7.5 to one and the second says 3 to one — one of those justifies doubling spend and the other says the business is near its line.
What belongs in the deduction: cost of goods, fulfilment and shipping, payment processing, promotional discount, returns and refunds, and any per-order support cost that can honestly be attributed. What does not: fixed overhead, which does not vary with the customer, and acquisition cost itself, which LTV is compared against rather than reduced by. If you do deduct acquisition cost, label the result net LTV, because the two quantities are constantly confused for one another.
Why cohort LTV beats one blended number
A blended LTV is computed across everyone currently in the book, which means it is computed across a population selected for having survived. The customers who left are not in it. The result drifts upward whenever acquisition slows and downward whenever it accelerates, for reasons that have nothing to do with what a customer is worth.
Cohort LTV asks a different question: of everyone acquired in a given month, how much cumulative contribution has each of them produced by month 1, 3, 6, 12? The denominator is fixed at acquisition and includes the customers who churned, so the number cannot be flattered by mix.
It also makes change visible when it happens rather than a lifetime later. A shift in acquisition channel, a deeper introductory discount or a new entry product shows up in the next cohort's month-3 figure immediately, while a blended average takes as long as the average customer lifetime to reflect it.
The practical form is a grid: cohorts down the side, months since acquisition across the top, cumulative contribution per acquired customer in the cells, and modelled values only beyond the months actually observed — marked as modelled. Anything past the observation window is a forecast and should look like one on the page.
Frequently asked questions
- What does LTV mean?
- LTV, or customer lifetime value, is the total value a customer is expected to produce across the whole of their relationship with a business. It is most useful measured as contribution — revenue after goods, fulfilment, payment processing, discounts and returns — because that is the figure that can honestly be compared against what it cost to acquire them.
- How do you calculate LTV?
- The standard identity is revenue per period, times margin, divided by the churn rate for that period. A subscription billing $45 a month at a 40% contribution margin, losing 6% of subscribers monthly, gives $18 divided by 0.06, or $300. The division by churn assumes an unlimited horizon and a constant churn rate, so treat the result as an estimate.
- What is a good LTV to CAC ratio?
- Three to one is a convention rather than a law, and it means nothing without two further facts: the horizon the LTV was measured over, and the payback period. A three-to-one ratio realised over three years and one realised over six months describe entirely different businesses, and only one of them can be funded from its own cash flow.
- Should LTV use revenue or profit?
- Contribution — revenue less cost of goods, fulfilment, payment processing, discounts and returns. Revenue LTV systematically flatters thin-margin subscriptions: the same customer worth $300 in contribution can be worth $750 in revenue, and comparing the revenue figure against acquisition cost produces a spending decision the margin does not support.
- What is the difference between LTV and AOV?
- AOV is the average value of one order. LTV is the total value of every order a customer is expected to place. AOV is a checkout metric, moved by bundling, upsells and shipping thresholds; LTV is a retention metric, moved mainly by churn. A business can raise AOV and lower LTV at the same time by discounting its way into larger, worse orders.
- Why does LTV change depending on who calculates it?
- Three inputs are usually left unstated. The horizon: an unbounded model returns a far larger number than a twelve-month measurement of the same cohort. The margin basis: revenue or contribution. And the churn rate: which denominator, over which period, blended or by cohort. Publish all three beside the number and the disagreements resolve themselves.