Hey,

Quick one today.

Picture two brands. Same $60 CAC, same $80 first order.

On every dashboard you normally look at, they're twins.

One of them should double spend going into Q4. The other one is slowly dying, and the difference lives in a number most founders have never pulled: what a new customer is worth by day 90.

If you read the scorecard email two weeks ago, this is one of the two numbers the question marks always cluster on.

Brand one's customers come back.

By day 90, the average new customer has generated $150 of gross profit against that $60 CAC. Ratio: 2.5.

Every dollar into acquisition returns two and a half within a quarter. That brand has permission to scale hard, even when day-one ROAS looks mediocre.

Brand two's customers buy once.

Day-90 gross profit per customer: $70. Ratio: 1.2.

They're paying $60 to earn $70 over three months, before a single fixed cost. Scaling that harder in Q4 means scaling a leak during the most expensive quarter of the year.

Identical on every dashboard, and headed in opposite directions.

And here's the part that makes this urgent: the trend matters more than the number itself.

Pull your last six monthly cohorts.

For each month, take the 90-day gross profit per new customer and divide it by that month's CAC. Write the six ratios in a row.

That row tells you one of three stories:

• Climbing or steady above 2: your machine works, scale with confidence.
• Hovering between 1.5 and 2: you're fine, but fix retention before you pour Q4 fuel on it.
• Sliding month over month: every cohort you buy is worse than the last one, which usually means scaling has pushed you into weaker audiences, and Q4 will accelerate the damage exactly when it's hardest to see.

That third pattern is the dangerous one, because blended dashboards hide it completely.

Revenue grows, ROAS looks acceptable, and meanwhile each month's customers are worth less than the month before. By the time it shows up in the P&L, you're two quarters deep.

One nuance so you read your row fairly: a single weak cohort is noise, especially if you ran a big promo that month, because discounted cohorts almost always grade worse.

Three sliding cohorts in a row is a trend, and a trend this close to Q4 deserves budget consequences.

We built a free grader that does the math and tells you which of the three stories you're in.

One row of six numbers, one evening of work.

Then set your November budget from that row instead of from a ROAS screenshot, and you'll make a better Q4 decision than most brands twice your size.

Talk soon,
Andrej

PS: Can't find 90-day gross profit per cohort in your stack? Reply "cohorts" and tell me what you run on (Shopify, Triple Whale, Lifetimely, spreadsheets), and I'll send you the exact steps to pull it.