POAS, not ROAS: measure profit from ads, not revenue

POAS (Profit on Ad Spend) measures how much real profit is left from every unit of currency put into advertising; ROAS (Return on Ad Spend) measures only the revenue that ad touched. The difference is fundamental: two campaigns with an identical ROAS of 4.0 can leave one in solid profit and the other in loss, because ROAS cannot see product margin, payment processor fees, the cost of returns or logistics. If you scale budget on ROAS and your net margin is falling, that is not bad luck, it is a flaw built into the metric. Below I show how POAS differs from ROAS, how to calculate it on your own numbers, and why chasing revenue alone stops being enough in 2026.

POAS versus ROAS: one number in the numerator changes everything

Both metrics divide by the same thing: ad spend. They differ in the numerator, and that is the whole story. ROAS is revenue divided by ad spend: a ROAS of 4.0 means every unit of ad spend brought in four units of gross sales. Nothing more. It says nothing about how much of those four is left after cost of goods, fees and returns.

POAS is gross profit divided by ad spend, where gross profit is revenue minus cost of goods sold and the other variable costs attached to the order. A POAS of 1.8 means every unit of ad spend brings 1.80 of profit, out of which you still have to cover fixed costs. The line is sharp: a POAS below 1.0 means the ad alone costs more than the product earns on it, so every extra sale deepens the loss. ROAS never shows that line, because in its world every sale looks equally good.

What ROAS cannot see

ROAS treats revenue as the result, but those are two different things. Between revenue and the profit that actually stays in the account sit several layers of cost that ROAS, by definition, ignores:

• Product margin. Selling a 200 item at 55% margin and one at 25% margin looks identical to ROAS. To the result it is the difference between profit and a fight for zero.
• Payment processor fees. Every card, wallet or pay-by-link is a fraction of the basket handed to the gateway. At low margins that fraction carries real weight.
• Return costs. A returned product is not just lost revenue but shipping both ways, handling, and sometimes a drop in the value of the goods. In some categories returns eat the entire surplus.
• Logistics and fulfilment. Packing, shipping, subsidising free delivery. A fixed cost per order that, on a cheap basket, can exceed the margin.
• Different margins per product and market. A campaign mixes products and markets of wildly different profitability under one averaged ROAS. The average hides the fact that you are mostly selling the least profitable range.

Each of these layers is invisible to ROAS and real to the P&L. The lower the margin, the bigger the gap between what the ad panel shows and what the accounts show at month end.

Same ROAS, different POAS: a worked example

Take two campaigns with identical revenue and identical budget, therefore identical ROAS of 4.0. One thing separates them: product A has a 55% gross margin, product B has 28%. All the numbers below are illustrative, chosen for the example, not drawn from any specific store or study. Drop in your own and the mechanism stays the same.

Illustrative numbers, chosen only to show the relationship. The same ROAS of 4.0, two different results: A earns, B loses on every extra sale.

How to calculate POAS for your store

POAS is only as good as the data you put into it. To calculate it honestly, rather than on averages that hide the problem, you need four things at the level of a product or product group, not the whole store:

• Gross margin per product. Selling price minus real cost of goods (purchase, duty, per-unit packing). Without it you compute POAS on an average, and the average is exactly what hides the least profitable SKUs.
• Return cost. The real return rate for the category times the cost of handling one return. In high-return categories this is often the single largest correction.
• Fees and transaction costs. The payment gateway, any marketplace commissions, the cost of fulfilment per order.
• Real CAC. The cost of acquiring a customer from a given campaign, ideally split between new and returning, because mixing them inflates how profitable acquisition looks.

The formula itself is simple: POAS = gross profit on ad-attributed sales / ad spend. The difficulty is not in the division, it is in getting margin per product, return costs and fees to actually flow into one place where you can set them against ad spend. If that data lives across three systems and a spreadsheet, the problem is not a lack of knowledge about POAS, it is a broken data layer. We show the same cost-breakdown logic on the example of automation in the anatomy of savings: an honest number comes from the breakdown, not from one averaged metric.

The scaling trap

The most expensive mistake is not looking at ROAS. It is scaling budget on ROAS. As long as a campaign runs on a small budget, a negative POAS hurts only symbolically. The trouble starts when the panel says "ROAS 4.0, room to add" and you add. On a low-margin product, every extra unit of budget at a constant ROAS grows not the profit but the loss. The metric that was meant to be a signal to grow becomes a signal to burn cash faster.

So express your break-even in POAS, not ROAS. "Scale while ROAS is above 3" is an instruction with no knowledge of margin. "Scale while POAS is above the threshold that covers fixed costs and leaves the planned net margin" is an instruction that does not lead downhill. The target POAS is not universal: it depends on your fixed cost structure and how much net margin you want left at the end.

The end of growth at any cost

This part is our read on the market, not a metric: in 2026 the era in which "growth at any cost" was the default eCommerce strategy is over. With dearer money, higher acquisition costs and pressure on profitability, the game shifts from maximising revenue to defending real net margin. ROAS is a metric of the cheap-growth era, because it rewards turnover. POAS is a metric of the discipline era, because it rewards the result. This is not about fashion for a new acronym, it is that at thin margins the gap between turnover and profit stopped being a rounding error and became the line between a business that earns and one that grows to zero.

Moving to POAS rarely fails at the level of the formula. It fails at the level of data: margin per product out of date, return costs counted nowhere, fees scattered. The same thing shows up in an honest total-cost calculation, when instead of the list price you count the real cost of ownership and the saving over three years; you can run it in the Blueprint Check calculator. And keeping that data layer in a shape where POAS computes itself, rather than once a quarter in a spreadsheet, is ongoing work, the same kind we handle in the Maintenance and Growth stage.

Where to start

Do not swap ROAS for POAS overnight across the whole store. Start with one thing: calculate POAS for the three campaigns you currently consider your best, using real margin per product rather than the store average. There is a good chance the ranking of "best" campaigns reshuffles, and at least one ROAS winner turns out to be on the edge or below the line. That reshuffle is the whole value of POAS: it does not promise higher sales, it shows which sales actually leave money behind. Whether your platform and data even let you join margin, returns and ad spend in one place is a separate question, worth asking before you optimise anything.