POAS (profit on ad spend)

POAS is gross profit divided by ad spend. The arithmetic is identical to ROAS with margin substituted for revenue in the numerator, and that single substitution changes which campaigns look good — which is the entire reason to compute it.

The problem it solves is a bidding problem, not a reporting one. Revenue-based ROAS treats a 70%-margin sale and a 15%-margin sale as equal outcomes, so Smart Bidding optimizing toward revenue will systematically buy more of whichever product converts most cheaply, regardless of what it contributes. The algorithm is not making a mistake. It is maximizing exactly what you told it to maximize, and you told it revenue.

Work an example. Campaign A returns $50,000 on $10,000 of spend, a 5.0 ROAS, selling accessories at 18% margin: $9,000 of gross profit, so POAS is 0.90 and the campaign is destroying value. Campaign B returns $24,000 on the same $10,000, a 2.4 ROAS that looks mediocre next to A, selling a 62% margin range: $14,880 of gross profit, a POAS of 1.49. Ranked by ROAS, A is the hero and B is a candidate for cuts. Ranked by POAS, the ranking inverts. The first time an account is scored this way, at least one flagship campaign usually drops from hero to passenger.

A POAS of 1.0 is breakeven by definition — gross profit exactly equals ad spend — which makes the number easier to reason about than a ROAS target that changes with every margin band. Above 1.0 advertising is contributing; below it, it is consuming.

The mistake that stalls most POAS projects is waiting for perfect data. Teams decide they need exact landed cost per SKU, including freight and duty, before they can begin, and the project dies in a spreadsheet. Margin bands — high, mid, low, applied across the catalog — re-rank campaigns far more truthfully than exact revenue values ever will, because the error you are correcting is measured in tens of percentage points and the error introduced by banding is measured in single digits. Start banded, refine later.

Implementation runs through conversion values rather than reports: send margin-reflecting values to the platform instead of order revenue, and bidding starts pursuing contribution on its own. Two cautions. Values must stay consistent, because switching mid-flight resets the algorithm’s learned economics. And POAS uses gross margin, so it still ignores fixed costs — a portfolio of POAS-positive campaigns can coexist with an unprofitable business if overhead is heavy enough.