Cohort LTV:CAC and Payback Period: When a Customer Cohort Is Mature Enough to Judge

This post is written for ECOM / DTC operators. The subscription-business version of this math (recurring MRR, churn-driven retention) is meaningfully different — SaaS readers should see our SaaS marketing attribution post.

LTV:CAC is the ratio every DTC investor asks about and the one most brands compute wrong — not because the arithmetic is hard, but because they measure it at the wrong time. A 30-day LTV:CAC and a 12-month LTV:CAC for the same cohort can tell opposite stories, and acting on the early number is one of the most expensive mistakes in DTC finance. This post is about cohort-based LTV, how it pairs with CAC, why payback period is the more honest near-term metric, and — above all — when a cohort is mature enough to judge. The glossary entries on cohort LTV and payback period define the terms.

The technical reality — why cohorts, not a single LTV number

A single store-wide "LTV" figure is nearly useless because it blends customers acquired last week with customers acquired two years ago, who have had wildly different amounts of time to make repeat purchases. The two-year customers inflate the average; the last-week customers haven't had a chance to repeat at all. The blended number tells you neither what a new customer is worth nor whether your economics are improving or decaying.

Cohort analysis fixes this by grouping customers by acquisition period (the month they first purchased) and tracking each cohort's cumulative value over time since acquisition. This reveals three things a blended LTV cannot:

1. The repeat-purchase curve. How a cohort's cumulative revenue-per-customer accrues month-over-month after first purchase — steep early, flattening later, with the shape varying enormously by category.
2. Whether cohorts are improving or decaying. If your March cohort is worth more at month-3 than your January cohort was at its month-3, your acquisition quality or retention is improving. A blended LTV would hide this entirely.
3. The maturation timeline. How long it takes a cohort to reach a stable, projectable value — which is the answer to when you can trust an LTV:CAC ratio for that cohort.

CAC — getting the denominator right

CAC is the fully-loaded cost to acquire a new customer: total acquisition spend (ad spend plus the directly-attributable costs of running acquisition) divided by new customers acquired in the same period. Two disciplines matter:

• New customers only. Dividing total spend by all orders (including repeat purchases from existing customers) understates CAC and flatters your economics. The denominator is new customers.
• Acquisition spend only on the cost side. Retention spend (email/SMS to existing customers, loyalty programs) is not acquisition cost — folding it into CAC inflates the number and mis-frames retention as acquisition. Keep them separate, mirroring the new-vs-returning discipline from our AOV post.

CAC is best read by channel, because acquisition cost varies sharply across prospecting, branded search, and so on — the same logic behind reading AOV by channel.

Payback period — the honest near-term metric

Here is the core problem with LTV:CAC for a young brand: true LTV is a projection into the future that you cannot observe until enough time has passed. A brand that's been acquiring customers for eight months simply does not have an observed 24-month LTV — any 24-month LTV:CAC it quotes is a forecast resting on assumptions about repeat behavior it hasn't witnessed yet.

CAC payback period sidesteps the projection problem. It asks: how many months of a cohort's cumulative contribution margin does it take to recover that cohort's CAC? It is computed entirely from observed data — you watch a cohort's cumulative margin accrue and note the month it crosses its acquisition cost. No long-horizon forecast required. This is why payback period is the more honest metric for near-term decisions and cash planning: it tells you when you get your acquisition dollar back, using only what has actually happened. The glossary entry on payback period walks through the calculation.

The relationship: payback period is the near-term, observed discipline; LTV:CAC is the long-term, partly-projected ambition. A healthy brand watches payback period for cash-cycle decisions and uses cohort LTV:CAC for strategic ones — knowing the latter is only trustworthy once cohorts have matured enough to project.

When is a cohort mature enough to judge?

This is the question that separates good DTC finance from wishful thinking. A few principles:

• A cohort younger than your repeat-purchase cycle is structurally unjudgeable. If your typical customer's second purchase comes at month 3–4, a cohort that's only six weeks old has barely begun its repeat curve. Its "LTV" is essentially first-order AOV, and its LTV:CAC will look artificially low. Don't kill a channel on a six-week-old cohort's ratio.
• Use the repeat-purchase curve's flattening point as the maturity signal. Plot the cohort's cumulative revenue-per-customer over months since acquisition. Early on it rises steeply; eventually the marginal monthly addition shrinks. The point where the curve flattens enough to project reliably is the cohort's maturity threshold — and it's category-specific. Consumables flatten faster (frequent repeats stabilize quickly); considered durables take much longer.
• Compare cohorts at the same age, never at the same calendar date. "January cohort vs March cohort" is only meaningful if you compare them at the same number of months since acquisition (e.g. both at month-3). Comparing a 5-month-old cohort to a 1-month-old cohort at today's date is comparing different points on the maturation curve — a guaranteed mis-read.
• Treat early LTV:CAC as a floor, not the answer. A young cohort's observed value is the minimum it will reach (assuming non-negative future repeats). It's a useful floor for "are we at least not catastrophically underwater," but not the basis for a confident scaling decision.

Methodology — building a trustworthy cohort LTV:CAC

Step 1 — Group by acquisition month and track cumulative margin

Bucket customers by first-purchase month. For each cohort, track cumulative contribution margin per customer (revenue minus COGS and variable order costs) at month 1, 2, 3, … since acquisition. Use contribution margin, not revenue — the same lesson as the AOV post: revenue that doesn't survive COGS and shipping doesn't fund anything.

Step 2 — Compute CAC for each cohort, new-customers-only

For each acquisition month, divide that month's acquisition spend by that month's new customers. This gives a per-cohort CAC you can line up against the cohort's cumulative-margin curve.

Step 3 — Read payback period off the curve

For each cohort, find the month where cumulative contribution margin crosses CAC. That month is the cohort's payback period — fully observed, no forecast. Watch whether payback is shortening across successive cohorts (improving economics) or lengthening (decaying).

Step 4 — Only project LTV once the curve has flattened

For mature cohorts whose repeat curve has flattened, you can project a defensible LTV and compute LTV:CAC. For young cohorts, don't — report their observed cumulative value as a floor and explicitly label it as not-yet-mature.

Step 5 — Compare cohorts at equal age, segment by channel

Always compare cohorts at the same months-since-acquisition. Where you have the data, build per-channel cohorts — acquisition channel often predicts retention, and a channel with a high CAC but a short payback can be better than a cheap channel with customers who never return.

Step 6 — Re-read as cohorts age

A cohort's verdict isn't final until it matures. Re-read your cohort table on a cadence so a young cohort's floor gets replaced by its real, observed payback as time passes.

Illustrative scenario

Imagine a DTC brand that pauses a paid channel because its 30-day LTV:CAC came in below 1.0 — customers acquired through it appeared to be costing more than they were worth in the first month. The decision looks prudent on the 30-day number.

A cohort view at equal age tells the opposite story. That channel's customers had a longer consideration cycle and a later but steeper repeat curve — their second and third purchases clustered around months 3–5, well outside the 30-day window. Read at month-5, the cohort's cumulative contribution margin had crossed CAC comfortably; its payback period was longer than the brand's fastest channel but well within an acceptable range, and its mature LTV:CAC was actually among the brand's best. By killing the channel on the 30-day number, the brand cut off a high-LTV acquisition source because it judged the cohort before the cohort's repeat cycle had even begun. The figures here are illustrative; the pattern — a channel that looks bad on a too-early window and good once the cohort matures — is the recurring, expensive trap that payback-period and equal-age cohort discipline exist to prevent.

What we do at Admaxxer

Admaxxer reports cohort LTV by acquisition month with the cumulative-margin curve, CAC computed new-customers-only, and payback period read directly off observed data — not a borrowed forecast. Cohorts are comparable at equal age and segmentable by channel, so a young cohort's floor is never mistaken for its verdict. The marketing acquisition view ties CAC by channel to the cohort curves, and our revenue tracking pipeline keeps the underlying order and margin data tied to your canonical commerce source. For the first-order side of the economics, see our AOV post; for the efficiency framing, our blended MER vs ROAS guide. Pricing is on the pricing page.

FAQ

What is a good LTV:CAC ratio?

There is no universal number — it depends on your gross margin, repeat-purchase economics, payback tolerance, and cash position, and it's only meaningful for mature cohorts. More important than hitting a ratio someone published is (a) computing it on mature cohorts at equal age, (b) using contribution margin not revenue, and (c) watching the trend across cohorts. A ratio quoted on a too-young cohort is a forecast dressed up as a fact.

Why is payback period better than LTV:CAC for a young brand?

Because payback period uses only observed data — you watch a cohort's cumulative margin cross its CAC and note the month — whereas a long-horizon LTV:CAC requires projecting repeat behavior you haven't witnessed yet. A brand that's been acquiring for eight months has no observed 24-month LTV; any 24-month LTV:CAC it quotes is an assumption. Payback period tells you when you actually get your acquisition dollar back, no forecast required.

How long should I wait before judging a customer cohort?

At least past your typical repeat-purchase cycle, and ideally until the cohort's cumulative-margin curve has flattened enough to project. This is category-specific: consumables flatten in a few months; considered durables can take much longer. The concrete signal is the curve's flattening point — judging a cohort younger than that means judging it before its repeat behavior has played out.

Should I use revenue or contribution margin for LTV?

Contribution margin. Revenue that doesn't survive COGS, shipping, and payment fees doesn't fund acquisition, so an LTV built on revenue overstates what a customer is actually worth to the business. Build the cohort curve on cumulative contribution margin per customer, and compute CAC payback against that same margin.

How do I compute CAC correctly?

Total acquisition spend divided by new customers in the same period — not all orders (which includes repeats from existing customers and understates CAC), and with only acquisition spend on the cost side (retention spend like email to existing customers is not acquisition cost). Read it by channel where you can, since acquisition cost varies sharply across channels.

Why does comparing cohorts at the same calendar date mislead?

Because cohorts acquired in different months are at different ages today — a 5-month-old cohort has had far more time to repeat than a 1-month-old one, so comparing their cumulative values at today's date measures the age gap, not a real quality difference. Always compare cohorts at the same months since acquisition (month-3 vs month-3) to see genuine improvement or decay.

Does acquisition channel affect LTV?

Frequently, yes — the channel a customer came through often predicts their retention and repeat behavior, so a high-CAC channel with loyal, repeat-prone customers can beat a cheap channel whose customers never return. This is why per-channel cohort LTV and per-channel payback are worth building: the cheapest channel by CAC is not always the best channel by lifetime contribution.