The cohort that never crossed 1.0
A company reports a 14-month CAC payback. By the benchmark most decks cite, that is healthy. What the benchmark does not tell you is which cohort produced it, whether the recent cohorts look anything like the early ones, or whether the customers stayed long enough for the payback to actually happen.
A single payback number is a conclusion. The curve underneath it is the evidence. And when the evidence is plotted cohort by cohort, the conclusion often falls apart.
What the curve shows that the number hides
The way to test payback is to track it over time for each cohort separately. For every group of customers acquired in the same month, plot the cumulative gross-margin-adjusted lifetime value per customer as a ratio against the cost it took to acquire them. Month by month, the curve rises as those customers pay. When the ratio crosses 1.0, the cohort has earned back its acquisition cost. Everything above 1.0 is real return.
A healthy cohort's curve rises steadily and crosses 1.0 within a reasonable window. A cohort in trouble rises slowly, then flattens. The flattening is customers leaving. Once enough of them are gone, the remaining revenue cannot push the curve any higher, and the ratio sits below 1.0 permanently. That cohort never paid back.
A single payback figure cannot distinguish between these two shapes. It reports the average, which is anchored by the early cohorts that crossed quickly and hides the recent ones that may not cross at all.
Why the early cohorts flatter the number
The first customers a company acquires are almost always its best. They found the product when the market was thinner, they needed it the most, and they were often sold by the founders at near-zero sales cost. Those cohorts cross 1.0 fast and keep climbing.
The cohorts acquired two years later, further down the demand curve, won with heavier spend and competing against more alternatives, usually look different. The curve rises slower because CAC is higher. It flattens sooner because retention is lower. Blend them with the early cohorts and the payback figure stays reassuring. Separate them and the recent trajectory is visible.
An investor plotting the curves will see whether each successive cohort is crossing 1.0 faster or slower than the one before it. That direction is the unit economics trend, and no single payback number captures it.
Gross margin is the denominator that matters
There is a second place the number gets flattering. Payback computed on revenue rather than gross margin overstates how fast recovery happens. Revenue is what the customer pays. Gross margin is what you keep after delivering the product. The gap between them is the cost of delivery, and ignoring it makes payback look shorter than it is.
A company running 70% gross margin that quotes a 12-month payback on revenue is really looking at roughly 17 months on the margin it actually keeps. That is the denominator that pays back acquisition cost, not the top line.
The gmLTV/CAC ratio accounts for this by adjusting lifetime value for gross margin before dividing by CAC. It answers the question in real dollars retained, not dollars invoiced.
What "no payback reached" means
The worst case is not a long payback. It is no payback at all. When a cohort's curve flattens below 1.0 and stops rising, it means the customers in that cohort left before their cumulative gross profit covered what it cost to acquire them. The company paid to acquire customers it will never earn back.
This is not visible in a blended payback number. A blended figure can show 14 months while the most recent cohorts show no payback reached, because the early cohorts crossed so quickly they pull the average above the threshold. The average is real. It describes a cohort mix that no longer exists.
Expansion can rescue a slow curve, but check where it comes from
A cohort that crosses 1.0 late is not necessarily a problem. If the customers who stay expand their spend over time, the curve keeps climbing past 1.0 and the late crossing is an investment that eventually compounds.
The question is where the expansion is concentrated. If it is spread across the surviving base, the slow payback is a feature of a land-and-expand motion that works. If it is concentrated in a handful of accounts while the median customer flatlines, the aggregate curve looks healthy and the typical customer still loses money. The curve's shape tells you which story is true.
What this means before the raise
An investor will not accept a single payback number. They will plot the curve by cohort, computed on gross margin, and look for two things: whether recent cohorts are crossing 1.0, and whether each new vintage is crossing faster or slower than the last. The direction of that trend is the unit economics verdict.
Levian plots the gmLTV/CAC ratio for each cohort over time, computed from your transaction data and P&L, with lifetime value adjusted for gross margin. You see which cohorts cross 1.0, which ones flatten below it, and whether the trajectory is improving or deteriorating, before an investor runs the same analysis from your data room.