Customer value: LTV, cohorts and payback
Why an ad that loses money on the first order can still be a good ad, and how to tell when it is.
Net ROAS tells you whether an ad made money on the order it produced. It does not tell you what happens next — and for most shops, what happens next is where the profit is.
The first order is not the whole customer
Say it costs you 30 to win a new customer, and their first order leaves you 25 after costs. On that order alone you lost 5.
If that customer never returns, the ad was a mistake. If they come back twice more this year, it was one of the best things you bought. The ad is identical in both cases. The only difference is what the customer does afterwards, and that is what these three ideas measure.
LTV — lifetime value
LTV is the total revenue one customer brings you over time, not only on the first order. Molecule reports it per customer at a fixed point — for example, cumulative value at month 3 — so you can compare like with like.
On its own LTV is just a number. It becomes useful next to CAC, the cost of winning that customer:
LTV : CAC is the ratio to watch. 3 or higher is the usual target: each customer returns at least three times what they cost to acquire. Well below that and you are buying customers at a loss; far above it and you can probably afford to spend more.
Cohorts
A cohort is simply everyone who first bought from you in the same month. Tracking each month's group separately answers a question an average never can: are the customers you are buying now as good as the ones you bought before?
Reading the cohort table:
Each row is one month's new customers. Read along a row to see how that group behaves as time passes.
Retention % shows how many of them came back in each following month.
Cumulative revenue shows how much that group has spent in total so far.
Read down a column to compare groups at the same age. If recent cohorts are weaker at month 2 than older ones were, your newer ads are bringing in worse customers — even if Net ROAS looks fine.
That last point is the reason cohorts exist. A discount-heavy campaign can produce a strong ROAS and a cohort of one-time bargain hunters.
Payback period
Payback is how many months it takes for a customer's cumulative revenue to cover what you paid to acquire them.
It matters because of cash, not only profit. A channel that pays back in 2 months lets you reinvest the same money six times a year. One that pays back in 10 months can be perfectly profitable on paper while draining your bank account today.
Molecule marks 6 months or less as the target, and colours each channel: green for 4 months or less, amber for up to 6, red for longer — or for channels that do not pay back within 12 months at all.
Where to find it
All of this lives on the Cohorts & LTV page: the LTV and LTV : CAC tiles, the cohort table, LTV curves per acquisition channel drawn against your blended CAC, and payback per channel.
How to use it with Net ROAS
Short payback, strong LTV : CAC — you can accept a Net ROAS below your first-order break-even on that channel, because the customer repays it.
Long payback, or no repeat purchases — hold that channel to your first-order break-even Net ROAS. There is no second order coming to rescue it.