Churn sets the size of the business
New signups get the attention, but with steady acquisition and steady churn the customer count does not rise forever. It converges on new customers divided by churn rate. A hundred a month against five percent churn levels off at two thousand; the same hundred against ten percent levels off at one thousand. Doubling the marketing budget doubles the ceiling and leaves the shape untouched, which is why churn work outranks acquisition work once a business is past its first phase.
Lifetime value has the same term in the denominator: contribution per month divided by churn. Moving churn from five percent to four raises lifetime value by twenty-five percent. Getting the same twenty-five percent from price means raising it by a quarter, which pushes churn back up. Retention is the lever with no equal and opposite reaction attached.
Measuring churn, and four ways to get it wrong
The definition is customers lost in a month over customers at the start of the month. The complications are all in the measurement.
- Fast growth hides it. When most of the base joined recently, the blended rate reflects a population that has not had time to leave. Growth slows, the rate appears to jump, and it was always that number.
- Annual plans distort it. An annual subscriber shows zero churn for eleven months and then all of it at once. Blend them with monthly customers and you lose sight of when the risk actually sits.
- Involuntary churn gets counted as rejection. Expired cards and failed charges commonly make up twenty to forty percent of departures. That is a payments problem wearing a product problem's clothes.
- Customer churn is not revenue churn. If leavers are small accounts and stayers upgrade, the count falls while revenue rises. B2B businesses usually track the revenue version for that reason.
Where the three-times rule comes from
The reason for a bar above one is that lifetime value here is contribution, not profit. Variable costs are already out; fixed costs are not. At exactly two times, each customer returns one CAC's worth beyond their acquisition cost, and that surplus has to cover engineering, support salaries, tooling and rent across the whole company. It usually does not stretch.
| LTV/CAC | What it means | Common response |
|---|---|---|
| Below 1x | Each new customer destroys value | Stop spending, fix price or retention first |
| 1x to 3x | Variable costs covered, fixed costs tight | Fix the model before adding scale |
| 3x to 5x | The band most operators aim at | Keep the channels running, expand carefully |
| Above 5x | Possibly underspending on growth | Push acquisition and watch CAC as it rises |
The figure above five is worth investigating rather than celebrating. Broadening channels normally raises CAC, so the practical exercise is finding how far spend can rise before the ratio drops through three.
Payback period is the one that runs out of cash
A healthy ratio with a long payback period still empties the bank. Spend $50 to win a customer contributing $13 a month and recovery takes nearly four months; at a hundred customers a month, $5,000 goes out now against money arriving over the following quarter. The faster you grow, the wider that hole opens, which is why twelve months is a frequent ceiling and six is common where funding is tight.
Payback longer than the average retained life is a straightforward failure: the typical customer leaves before covering their own acquisition, and the business only works if a minority of long-stayers subsidise everyone else. In that situation the distribution matters more than the average, and the average is what this page computes.
Where to spend the next hour
Retention work tends to beat acquisition work on cost per point gained, and it has a rough order to it. Payment recovery comes first — card expiry warnings, retry schedules, a second payment method on file — because it is cheap and the failures are not decisions anyone made. The first thirty days come next, since departures cluster there and getting a customer to use the thing once changes the curve. Annual billing locks in twelve months at the price of a discount, which usually raises lifetime value despite lowering ARPU. Cancellation flows recover a little, though making cancellation difficult buys a short-term number at a long-term cost.
For fixed costs against contribution, break-even sales takes it further. What a price rise does to the same figures is modelled in the price increase simulator, and the unit economics of a single sale sit in the margin calculator.
Questions people ask
Which churn figure should I enter when it moves every month?
An average of the last three to six months, with any month distorted by an unusual event left out. If the business is growing quickly the blended rate understates the true one, so where you can, split by joining cohort and use the rate among customers who have been on the books six months or more. That figure predicts the long run better than the blended average, which is dominated by whoever signed up most recently.
Why does LTV use contribution instead of revenue?
Because revenue-based lifetime value flatters the ratio enough to justify almost any acquisition spend. Serving more customers costs more in hosting, payment fees and support, so the money a customer actually leaves behind is revenue minus the costs that scale with them. At a thirty percent variable cost share the two versions differ by thirty percent, which is more than enough to move a business from one side of the three-times line to the other.
How should free trial users be counted?
Keep them out of the customer count, which should be paying customers only, and put their cost into CAC. Total marketing spend divided by converted customers is the honest denominator. Dividing by trial signups instead makes CAC look much lower and inflates the ratio by whatever your conversion rate happens to be. The hosting and support costs of running trials belong in acquisition spend too, since that is what they are.
The twelve-month projection does not match what happened. Why?
It holds new customers, churn, price and cost share constant for a year, and none of them stay constant. Channels saturate, so the same budget brings fewer signups over time. Seasons move signups around. Price changes shift churn. Use it for the ceiling and the direction, and if you need a plan, run it several times with different acquisition assumptions per quarter and treat the spread as the real answer.
What counts as a variable cost here?
Anything that grows when the customer count grows: hosting and bandwidth, payment processing, per-seat licences for tools your product depends on, content delivery, and the support cost that scales with volume. What does not: salaries, office, and the engineering effort of building the product, all of which continue whether you have five hundred customers or five thousand. Putting salaries into the variable share understates contribution and makes every downstream figure on this page too pessimistic.