A 4.03x ROAS. Just 1.7% Contribution Margin. What Did ROAS Miss?

If someone told you an Amazon campaign was delivering a 4.03x ROAS, would you call it a good result?

Probably. AED 1 spent on advertising generated just over AED 4 in attributed sales. ACoS was below 25%. At first glance, the advertising looks reasonable.

But when we looked at the full economics of the SKU over 90 days, only 1.7% of revenue remained as contribution. That was just AED 36.73 from AED 2,188.21 in revenue.

Nothing was wrong with the ROAS calculation. The problem was expecting ROAS to answer a question it was never designed to answer. ROAS tells you how efficiently advertising generated attributed revenue. It does not tell you how much money the product ultimately left behind. Once we looked underneath the ROAS, the reason for the weak contribution became much clearer.

The key takeaway

A high ROAS can still sit inside weak product economics. Before deciding whether 3x, 4x or 5x ROAS is "good", look at how much of the product's revenue is already being consumed by product cost and marketplace fees — then look at how much advertising consumes.

• Product costconsumed 18.6% of revenue.
• Marketplace feesconsumed 53.4% of revenue.
• Advertisingconsumed another 26.4% of total revenue. What remained? 1.7%.

The advertising number looked healthy

Here is the Amazon UAE SKU over the 90-day period.

If the advertising team saw only 4.03x ROAS and 24.8% ACoS, there would be nothing obviously alarming. The campaign was generating attributed sales. But the product had very little economic room underneath those sales. That is the part that changes the decision.

What actually happened to the revenue?

Start with AED 2,188 revenue, then subtract costs in order:

The largest cost line was not advertising. Recorded marketplace fees represented more than half of the SKU's revenue during the period — 53.4%. That should trigger a separate investigation before assuming PPC is the main problem.

So if the first recommendation had been "Reduce PPC because margin is weak", we would have skipped over the largest cost line entirely. We cannot attribute that fee figure to a specific component — the available SKU-level data does not break it down. But we now know where to investigate first. That is already a much better starting point.

A 4x ROAS can mean very different things for different products

Imagine two products. Both show 4x ROAS.

The same advertising efficiency is sitting on top of two completely different businesses. That is why the question should not be: "Is 4x ROAS good?" It should be: "Is 4x enough for this SKU?"

The same SKU makes the point even more clearly on Noon

The same SKU was also selling on Noon. Its Noon ROAS was 3.93x — almost identical to Amazon's 4.03x. If we ranked the two marketplaces only by ROAS, Amazon would look slightly better. But look at the full economics:

Amazon had the slightly higher ROAS. Noon had the substantially higher contribution margin. The difference was underneath the advertising metric — Noon's combined product-cost and recorded-fee burden was materially lower, leaving more room for advertising to work within. That is exactly why ROAS should not be used as a shortcut for profitability.

ROAS and contribution are answering different questions

• ROASHow much attributed revenue did advertising generate for every AED 1 spent?
• ACoSWhat share of ad-attributed sales was spent on advertising?
• TACoSWhat share of total product revenue was spent on advertising?
• ContributionWhat remained after the costs included in the analysis?

Both ROAS and contribution are useful. Neither should replace the other. A marketplace team still needs ROAS to understand advertising efficiency. But once the question becomes "Should we increase the budget?" — ROAS needs context. That context includes product cost, marketplace fees, total advertising dependence, returns, pricing, inventory, contribution and marketplace objective. A campaign can be efficient at generating attributed sales while the underlying product economics remain unattractive.

What about break-even ROAS?

A useful way to think about ROAS is to start with the margin available before advertising. In simplified terms: Break-even ROAS ≈ 1 ÷ margin available before advertising.

This is a simplified operating approximation, not final accounting profitability. ROAS is based on ad-attributed sales. Contribution uses total recorded revenue and costs. They answer different questions and should not be expected to reconcile as the same number.

For this SKU, approximately 28% of total revenue remained after product cost and recorded marketplace fees. Advertising represented 26.4% of total revenue. That left 1.7%. This is why TACoS is useful alongside ROAS — it tells you how much of the product's total revenue was being spent on advertising, not just ad-attributed sales.

So should the advertising budget be reduced?

Possibly. But that is not the first conclusion to make from this SKU. The bigger questions are:

• Why is the recorded marketplace-fee burden 53.4% of revenue?
• Can the product support the current selling price and fulfilment economics?
• Can advertising become more efficient without materially reducing sales?
• Does the SKU have enough strategic value to tolerate a very low contribution margin?

Only after that should the team decide whether to reduce spend, change bids, change price, review fulfilment, investigate marketplace charges, or simply stop pushing the product harder. "4.03x ROAS" is not the decision. It is one input into the decision.

The objective still matters

A low contribution margin does not automatically mean a campaign should be stopped. A business may deliberately accept weaker short-term economics because it is launching a new product, entering a marketplace, building rank, acquiring customers, clearing inventory, or testing whether a product can gain traction. Those may all be valid decisions. That is why there is no universal ROAS number that should automatically trigger "scale" or "stop" — the difference is whether weak economics are deliberate and understood, or simply going unnoticed.

What does this mean for marketplace budget allocation?

Imagine two SKUs:

Which one should receive the next AED 1,000 of advertising? There is still not enough information to make the final decision — you would also want to know inventory, incremental demand, conversion, campaign headroom, organic sales, strategic importance and customer value. But one thing is clear: you should not rank the opportunity purely by ROAS.

Advertising and profitability need to meet in the same decision

Advertising teams naturally look at ACoS, ROAS, CPC, conversion and campaign sales. Finance looks at cost, fees, margin and contribution. Marketplace operations looks at inventory, pricing, listings and availability. The product does not experience those as separate departments — it experiences all of them at once.

What should you check before increasing spend on a high-ROAS SKU?

1. Contribution marginWhat is actually left after the costs you are measuring?
2. Margin before advertisingHow much room does the SKU have before advertising spend begins?
3. Product costHow much of every sale is already committed to the product itself?
4. Marketplace feesIs the fee burden unusually high? If so, investigate before assuming advertising is the main problem.
5. TACoSHow much of total SKU revenue is being consumed by advertising?
6. ReturnsIs a good advertising result feeding into a product that customers are returning too often?
7. Marketplace differencesDoes the same SKU have similar ROAS but very different profitability somewhere else?
8. InventoryCan the product actually support more demand?
9. PricingDoes the current selling price leave enough room for the cost structure?
10. ObjectiveAre you trying to maximise contribution now, or deliberately investing in a longer-term goal?

The real lesson from the UAE account

The useful discovery was not that 4.03x ROAS is bad. It was almost the opposite — 4.03x was a perfectly real advertising result. But the SKU's underlying economics meant there was very little left after the sale.

There is no contradiction in those numbers. The contradiction only appears when we expect ROAS to mean profit.

• ROAStells you whether advertising is generating attributed revenue efficiently.
• The cost stacktells you how much room the product has.
• Contributiontells you what was left. Then the operator decides what to change.

Frequently asked questions