Stop worshipping ROAS: the metric that actually predicts profit

Every reporting call starts the same way: someone reads out a ROAS number like it settles the argument. It doesn’t.

ROAS tells you revenue per euro spent. It says nothing about margin, refunds, or whether that revenue was incremental in the first place. A 10x ROAS on a product with 15% margin can lose money the moment you factor in returns and payment fees. A 2x ROAS on a 60%-margin product can print cash.

Here’s what I actually calculate before I call a campaign “working”:

• Contribution margin per order — revenue minus product cost, shipping, payment fees, and returns. This is the number that pays rent, not revenue.
• Marginal ROAS — what happens to return when you add the next €500 of budget, not the blended average across everything you’ve spent this month.
• Payback window — for subscription or repeat-purchase businesses, a “bad” first-order ROAS can be excellent if customers come back three times.

None of this makes ROAS worthless — it’s a fast, comparable signal, and I still put it at the top of every dashboard. The mistake is treating it as the finish line instead of the first question. If a client’s whole reporting stack is one blended ROAS number, that’s usually the first thing I rebuild.