Which products bring the customer back?

What a first order is really worth, over the months that follow it.

You know what a sale costs you to win. You do not know what it is worth. A product with a thin first-order margin can be the best thing you sell if a third of those buyers come back, and the fat-margin one can be the worst if nobody ever does. Judge advertising on the first order alone and you will quietly spend the most on the products that earn the least.

What we do about it

• Buyers are grouped by the month they first bought a specific product, and we follow what that group spends in every month after — so a second order six months later still lands against the campaign that won the customer.
• Cohorts run as deep as your history goes. There is no fixed six- or twelve-month window cutting the tail off a product that sells slowly and repeats well.
• Revenue is converted into one currency, so a product selling across several marketplaces reads as a single number instead of five you add up yourself.
• Group by marketplace or by parent ASIN when the repeat pattern belongs to a variation family rather than one SKU — the size someone reorders is rarely the size they bought first.

A real moment

Two products, near-identical first-order profit, and the advertising budget split evenly between them because there was never a reason to do otherwise. The cohorts show one earning a second order from roughly a quarter of its buyers within four months, and the other almost never repeating. Nothing about either product changed — but the budget stopped being split evenly that week, and the question moved from what the sale makes to what the customer makes.

A composite scenario, drawn from patterns we see repeatedly — not a named customer's result.