What CAC payback tells you that LTV:CAC does not
CAC payback measures time, and time is the dimension that ratio metrics throw away. Two businesses can both return three dollars of lifetime value for every dollar of acquisition cost and have completely different funding requirements: one gets its money back in eight months and can grow from its own cash, the other takes thirty-four months and has to finance every new customer until then. The LTV:CAC ratio cannot tell them apart. Payback can.
That makes payback the cash metric of the pair. Every customer you acquire is a small investment: money out today, gross profit back over months. Until the cumulative gross profit crosses the acquisition cost, that customer is a hole in your bank balance, and if you are growing quickly you are digging new holes faster than the old ones fill. This is why fast-growing subscription businesses can be simultaneously profitable per customer and desperately short of cash — the arithmetic is working, but the timing is not.
The metric also has a diagnostic property that ratio metrics lack: it responds immediately. Lifetime value depends on a churn estimate for years you have not lived through yet, so it can be argued about indefinitely. Payback depends on this month's acquisition cost, this month's revenue per account and this month's margin, all of which you know. When a quarter goes badly, payback moves first and by an amount nobody can dispute.
One convention to fix before you compute anything: the revenue term must be multiplied by gross margin. Acquisition is repaid out of gross profit, not out of billings. A business at a 60% margin needs 1 ÷ 0.6 = 1.67 times as many months as one at 100% to recover the same CAC on the same revenue, and quoting payback on revenue rather than margin understates the period by exactly that factor.
Two formulas: the simple one and the honest one
The simple version divides CAC by monthly gross profit per account. It is the figure most companies report, and it is the right one when churn during the payback window is negligible — a long-contract enterprise product, say, where the customer cannot leave inside the first year.
The churn-adjusted version recognises that a cohort shrinks while it is repaying. If 2% of customers leave every month, then after twelve months only 0.9812 = 78.5% of the original cohort is still paying, and the cohort's cumulative gross profit is less than twelve times the monthly figure. Summing the geometric series gives cumulative gross profit after t months of GP × (1 − (1−c)t) ÷ c, and setting that equal to CAC and solving for t gives the log expression in the formula block above.
Because the adjustment removes revenue that the simple version assumed, the churn-adjusted payback is always at least as long as the simple one, and strictly longer whenever churn is above zero. That direction holds regardless of the numbers you enter, which is why the calculator reports both — the difference between them is the cost of churn expressed in months.
The churn-adjusted formula also exposes a ceiling the simple one hides. As t grows, (1−c)t approaches zero and cumulative gross profit approaches GP ÷ c. That limit is the lifetime gross profit of the cohort, and if CAC exceeds it the logarithm has a non-positive argument and there is no solution: the customers never pay back, at any horizon. The calculator returns a blank payback and says so explicitly rather than reporting a large finite number, because "never" and "in 200 months" are different business situations.
Worked example: $1,200 CAC at $100 of monthly gross profit
You spend $1,200 to acquire a customer who pays $100 a month. Take a 100% gross margin first so the margin term does not obscure the churn arithmetic, and assume 2% of customers churn each month.
- Monthly gross profit. $100 × 100% = $100.
- Simple payback. $1,200 ÷ $100 = 12.00 months.
- Lifetime ceiling. $100 ÷ 0.02 = $5,000. The CAC is well under this, so a solution exists.
- Churn-adjusted payback. 1 − (1,200 × 0.02 ÷ 100) = 1 − 0.24 = 0.76. Then ln(0.76) ÷ ln(0.98) = (−0.274437) ÷ (−0.020203) = 13.58 months.
- Check it. Cumulative gross profit at 13.58 months is $100 × (1 − 0.9813.58) ÷ 0.02 = $100 × 0.24 ÷ 0.02 = $1,200, which is the CAC. The formula is consistent.
- Cash at twelve months. $100 × (1 − 0.9812) ÷ 0.02 = $100 × 0.215283 ÷ 0.02 = $1,076.42, so the customer is still $1,200 − $1,076.42 = $123.58 underwater at the one-year mark.
Churn added 13.58 − 12.00 = 1.58 months, or 13.2% of the simple period. Now change the margin instead: at a 70% margin the monthly gross profit falls to $70, the simple payback rises to $1,200 ÷ $70 = 17.14 months, and the ceiling falls to $70 ÷ 0.02 = $3,500. Margin moves the answer far more than churn does at these values, which is the usual pattern and the reason a margin-blind payback figure is not worth quoting.
How to read your payback period
Compare the result to the cash you have, not only to a benchmark. The widely used rule of thumb in subscription software is that payback under twelve months is comfortable and beyond eighteen months requires either patient capital or a deliberate decision to slow growth. Treat those figures as conventions of the trade rather than as measured statistics — their usefulness is that investors and boards share them, not that they were derived from a published dataset.
The number that actually constrains you is the product of payback and your new-customer rate. If you acquire 100 customers a month at $1,200 each, you are committing $120,000 a month, and at a fourteen-month payback the peak cash tied up in unrepaid acquisitions runs to well over a million dollars before the earliest cohorts start funding the newest ones. Check that against your runway using the burn rate and runway calculator before you approve a growth plan.
Read the direction of travel as well as the level. Payback lengthening while acquisition volume grows is the normal, expected shape of scaling — cheap demand is captured first — and it is only alarming when it crosses the threshold your balance sheet can carry. Payback lengthening while volume is flat is a different diagnosis entirely: something has gone wrong in pricing, margin or acquisition efficiency, and the three inputs on this page will tell you which.
Finally, do not confuse a fixable payback with a broken business model. If the calculator says the cohort never pays back, the problem is structural: at that churn rate the customers simply do not live long enough to repay what they cost. No pricing tweak inside the current model fixes it — you need lower churn, higher price, higher margin or cheaper acquisition, and the churn rate calculator and CAC calculator are where those two conversations start.
Payback months by CAC and monthly gross profit (no churn)
| CAC | $50 GP/month | $100 GP/month | $200 GP/month | $500 GP/month |
|---|---|---|---|---|
| $500 | 10.0 | 5.0 | 2.5 | 1.0 |
| $1,000 | 20.0 | 10.0 | 5.0 | 2.0 |
| $1,200 | 24.0 | 12.0 | 6.0 | 2.4 |
| $2,000 | 40.0 | 20.0 | 10.0 | 4.0 |
| $5,000 | 100.0 | 50.0 | 25.0 | 10.0 |
| $10,000 | 200.0 | 100.0 | 50.0 | 20.0 |
Monthly gross profit is revenue per account times gross margin: a $250 subscription at a 40% margin belongs in the $100 column, not the $200 one. At 2% monthly churn the ceiling on repayment is GP ÷ 0.02, so the $50 column can never repay a CAC above $2,500 and several cells in it are unreachable in practice.
Ways this metric gets misreported
- Dividing by revenue instead of gross profit. The most common error, and it understates the period by the reciprocal of your margin — a 50% margin business reports half the true payback.
- Using a marketing-only CAC. Sales salaries, commission, sales engineering and marketing headcount all belong in acquisition cost. A media-spend-only figure flatters payback substantially in any business with a sales team.
- Ignoring churn during the payback window. The cohort shrinks while it repays. The adjustment is small at low churn and severe above about 5% a month, where the lifetime ceiling starts binding.
- Counting setup and implementation fees as recurring. A one-off fee shortens the first month only. Either subtract it from CAC or leave it out; do not amortise it into the monthly figure.
- Blending self-serve and enterprise cohorts. Their CAC and revenue per account can differ by two orders of magnitude, and the blended payback describes neither segment.
- Comparing a monthly-churn figure against an annual one. A 24% annual churn is roughly 2.3% monthly, not 24% monthly. Convert before entering it.
Payback next to the other unit-economics metrics
The three metrics that describe customer economics answer different questions and are not substitutes. CAC asks what a customer costs. LTV:CAC asks whether a customer is worth more than they cost. Payback asks when you get the money back. A business needs an acceptable answer to all three, and the cheapest way to fail is to have a strong ratio and a payback longer than your funding.
The three are algebraically linked. With a constant monthly gross profit and a constant churn rate, lifetime gross profit is GP ÷ c and payback is approximately CAC ÷ GP, so the ratio LTV ÷ CAC is approximately (GP ÷ c) ÷ (payback × GP) = 1 ÷ (c × payback). At 2% monthly churn, a twelve-month payback implies a ratio near 1 ÷ (0.02 × 12) = 4.2. That identity is a useful consistency check: if someone reports a twelve-month payback and a 2:1 ratio at 2% churn, one of the three numbers is computed on a different basis from the others.
For a recurring-revenue business, keep payback beside net revenue retention. Expansion revenue inside the existing base shortens effective payback, because the cohort's monthly gross profit rises over time instead of merely shrinking with churn — a business with net revenue retention above 100% is repaying acquisition with a growing annuity, and the simple formula on this page understates its performance. If your expansion is material, model the cohort's revenue explicitly rather than treating ARPA as constant.
Finally, payback is a period metric and should be recomputed every quarter on that quarter's cohort. A trailing-twelve-month blend hides exactly the deterioration you are trying to detect, because the good early cohorts keep propping up the average long after the acquisition environment has changed.