ROAS vs POAS. Same budget, two completely different decisions.

TTwo founders, same budget, same ROAS. One is profitable, the other is bleeding.

Why? Because ROAS is a map with no elevation. It shows you the road, but not whether you are going uphill or down. And that costs you.

It is a pattern you see all the time. An e-commerce account, weekly call with the agency. All good, 3.8x ROAS, best in six months. A few months later the business is in the red. The ROAS had grown. The profit had vanished.

It is the kind of story you hear four or five times in a founder lifetime. And each time it feels new.

Let us get into the numbers. You will see exactly where the logic breaks, and why ROAS on its own never tells you whether you are making a good call. It was not meant to. ROAS is a volume metric, not a profit metric. Confusing the two costs you dearly.

Two campaigns, same ROAS. Opposite outcome.

Online store with two categories. Electric scooters at 35% net margin. Accessories (helmets, chargers) at 8% net margin. Agency reports separately. 3x ROAS on both. On the report, identical.

Speed does not help if you are headed the wrong way. Here is what the numbers show.

The accessories campaign pulls EUR 3,800 out of the company every month. But the report says 3x ROAS, good performance.

If the agency scales accessories from EUR 5,000 to EUR 15,000 (and calls it ROAS-based scaling), the loss hits EUR 11,400 a month. With weekly meetings full of nodding.

I have seen exactly this scenario, and more than once. The agency reports aggregate ROAS. 3x. It looks good. The owner is happy. At year-end, after six months, he discovers half the budget went into campaigns that lost money. Nobody told him. Nobody calculated POAS per campaign.

That is why, when you sign with an agency, the first thing you must demand is POAS separately for each campaign. Not aggregate ROAS.

What changes once POAS enters the room

If POAS had been the decision metric, that meeting looked completely different.

• Scooters, 1.05 POAS. Clean break-even. You do not raise budget. You work on creative and CPM optimization to push POAS toward 1.4 before scaling anything.
• Accessories, 0.24 POAS. Obvious hemorrhage. Either kill the campaign, raise price 20% and test elasticity, or find a cheaper supplier. But you do not scale anything until one of those moves.

That is the fundamental difference. ROAS tells you: I sold. POAS tells you: I won or I lost.

In business, only the second question matters. The first is interesting for reports. The second is essential for survival.

IN execution, AT strategy, ABOVE the business

IN execution, with ROAS, decisions sound simple but dangerous. Scale what is big, cut what is small. Do what the report says.

AT the strategy level, with POAS, nuance appears. The campaign with high revenue but low margin does not get budget. The campaign with modest revenue but 2x POAS gets everything you have. This is where the real call is made.

ABOVE both, POAS becomes a tool for redesigning the business itself. You stop optimizing the campaign. You change the product, the supplier, the price. We got this wrong too. We tried optimizing campaigns that could never work because the margin underneath was too thin. The lesson: you cannot make an ad good enough to compensate for 8% margin.

What decisions look like with POAS in front of you

1. 01POAS between 1.5 and 2.5, scale. Raise budget 20% a week and watch whether POAS holds. If it drops below 1.5, stop. If it holds, continue.
2. 02POAS between 1.0 and 1.5, hold. Iterate on creative and audiences, do not raise budget. Work on efficiency. Test new angles, new messages, new landing pages.
3. 03POAS between 0.8 and 1.0, verify. If it brings high-LTV customers who repurchase, maybe keep it. But temporarily. And with a clear plan to reach 1.0 within 60 days.
4. 04POAS below 0.8, pause the campaign and run a separate analysis. No discussion, and no one more week of testing. You stop.

This is what we do at every weekly meeting. We open a spreadsheet. Columns: campaign, budget, revenue, net margin on dominant product, POAS. Sort descending. Top half gets more budget. Bottom half gets analysis. Bottom quarter gets stopped.

It is not complicated. It is just math that many avoid because the results are not always pleasant.

What to do when margin varies by product

In most stores, margin is not a number. It is a spectrum. That is exactly why you calculate POAS at the campaign level, not across the whole account. Meta Ads Manager will not do it. GA4 will not either. It is manual work.

Practically, it is a Google Sheet. Columns: campaign, revenue, dominant product, product net margin, budget. One formula per row: POAS = (revenue x margin) : budget. Twenty minutes a month. The only twenty minutes that actually matter. The rest of reporting is decor.

A predictable objection: but I have 50 products in the same campaign. The answer is a weighted average of margins on the products that generated 80% of that campaign's revenue. Pareto rule. You do not need to calculate every product, only the ones that matter.

Example: how you stop a hemorrhage hidden in the budget

Picture a home products store, 40 SKUs. The agency reports 3.2x aggregate ROAS. Happy owner. Until someone actually looks inside the account.

You open the account. You make the spreadsheet.

Half the campaigns can have POAS below 0.5, only one above 2, the rest between 0.3 and 0.9. A 3.2x aggregate ROAS easily hides the fact that a good part of the budget goes into campaigns that lose money.

Does the agency not know? Or do they know and say nothing?

The right call: stop the losing campaigns. Redistribute budget to the ones with POAS above 1.0. Cut the products with margin below 10%.

What can happen in a scenario like this: total revenue drops a little, but profit climbs. From loss to gain. Same budget. Just allocated intelligently.

When ROAS is genuinely useful

I will not tell you ROAS is useless. It is not. It has its role, and an important one. You just need to know where to use it and where it lies.

• Quick benchmarking. 2x versus 4x ROAS gives you a rough sense of relative performance without digging into margin. Useful for a first impression, not for decisions.
• Algorithmic optimization. Meta and Google optimize on purchase value, not on POAS. ROAS is the signal they read. You use it to see if the algorithm is heading the right direction, not whether you are profitable.
• Reporting to investors or partners who do not know the margin details. Here you label it clearly as ROAS, not POAS, so you avoid confusion. But you know, behind the scenes, that POAS is the real verdict.

The trouble starts the moment ROAS becomes the only decision metric. That is when you start making choices that look good in the presentation but empty the bank account. Most agencies do this, because it is easier to sell a 4x ROAS than a 0.8x POAS.

The Friday-evening test

Take the biggest campaign you are running right now. Four steps, ten minutes.

1. 01Ask for net margin on the dominant product. No estimates, no approximations. An exact number, calculated with every variable cost.
2. 02Pull last month's ROAS.
3. 03Apply the formula. POAS = ROAS x net margin.
4. 04If the result is above 1, you have room to scale. Between 0.8 and 1, optimize. Below 0.8, you have a serious issue and it needs to be addressed this week, not next month.

10 minutes that can save you hundreds of thousands of lei per year.

If you do not do this every week, you do not know what is happening in your account. You only know what the agency shows you. And the agency has an interest in showing you what looks good.