The Same Product Made Money on Amazon and Lost Money on Noon. What Changed?

The same product. The same business. The same 90-day period.

On Amazon UAE, it made money. On Noon, it lost money.

At first, that sounds like a marketplace problem. It could be tempting to conclude: "Amazon works for this product. Noon does not." But that would be too quick.

When we looked underneath the numbers, the underlying product economics on Noon were not actually worse. The bigger difference was advertising. Then another signal appeared: returns. And when we went inside the Noon campaigns, the problem became more specific again — some targets were generating orders, others were consuming spend without producing sales.

That changed the question from: "Should we stop advertising this product on Noon?" to: "What exactly is going wrong, and what should we change first?" The customer and product names have been removed, but the numbers below come from a real UAE marketplace account.

The key takeaway

A product does not automatically deserve the same strategy on every marketplace. Before changing price, cutting ads or moving budget elsewhere, separate the economics by marketplace and understand what is actually causing the difference.

• Amazonwas profitable at +18.2% contribution margin.
• Noonwas loss-making at −19.7% contribution margin.
• The causewas not the product itself — advertising was consuming too much of Noon's available margin.

The combined view looked weak, but not disastrous

Across Amazon and Noon together, the product family generated:

AED 13.5K of revenue. Positive contribution. But only an 8% contribution margin. At the combined level, the first conclusion might simply be: "This is a low-margin product." That is true. But it does not tell you what to do. For that, we need to separate Amazon and Noon.

The same product told two very different stories

Same product family. Same period. Amazon at +18.2% contribution margin. Noon at −19.7%. So what changed?

The product economics before advertising were not worse on Noon

This is the part that matters. Take advertising out for a moment.

The product actually had more room before advertising on Noon, not less. Then advertising was added, and Noon advertising consumed 61.3% of total revenue versus 15.0% on Amazon. That moved the final contribution from +18.2% to −19.7%. So the first measurable problem to investigate on Noon was not product cost. It was not the recorded marketplace-fee burden either. It was advertising.

But "advertising is the problem" is still not enough

This is where analysis often stops too early. We could look at 61.3% TACoS and 1.06x ROAS and conclude: "Cut Noon ads." But that still does not tell an operator what to change. Should the whole campaign be paused? Should the budget be reduced? Should bids change? Are particular keywords wasting money? Are there targets that should actually continue receiving spend? To answer that, we need to go inside the campaign.

One child SKU made the problem even clearer

Within the same product family, one Noon child SKU generated approximately AED 1,319 in revenue during the 90-day period. Advertising spend against it was approximately AED 1,892. Its TACoS was 143.5%. Contribution was approximately −AED 1,368, or −103.8%.

That is a serious signal. Advertising spend was greater than the SKU's total marketplace revenue during the period. But even here, the right next move is not automatically: "Switch the ads off." The next question is: Where inside the advertising is the money going?

The shorter optimisation window showed where the waste was sitting

The 90-day economics showed that the problem was persistent enough to investigate. A shorter 14-day Noon Optimizer window then gave a more current campaign-level view.

Reduce budget. Most of the campaign was not earning enough return — but some targets were still worth keeping. The recommendation was not to stop the campaign, but to reduce spend while protecting what was working.

Some targets were producing orders. Others were not.

Inside the same campaign, the evidence divided clearly between what was working and what was waste.

Elsewhere in the campaign, the Optimizer was also recommending bid reductions where observed ROAS was below the target — for example, "water bottles" (phrase) at an observed ROAS of 0.47x had its bid recommended down from AED 1.97 to AED 1.48, approximately −25%.

So the practical action was no longer: "Noon advertising is too expensive." It became:

• reduce the overall budget while the campaign is inefficient
• protect the targets still producing useful orders
• reduce bids where returns do not justify the current bid
• block terms that repeatedly consume clicks without sales
• then watch whether the economics improve

That is much closer to an operating decision.

Returns were another clue, but they told a different story

Across Amazon and Noon together, the family had a 7.7% return rate over the 90-day period. That deserves attention.

Drilling into that child SKU, some of the recorded return reasons included defective, quality unacceptable, not compatible. One customer comment specifically said the product was leaking. That is a genuine product or customer-experience signal.

If the team simply increased advertising without looking at this, it could end up paying to bring more customers into the same issue. Importantly, Noon return information is less complete in the available dataset — it would be wrong to use Amazon return comments to explain Noon performance. Both issues deserve attention; they are not the same problem.

This is why root cause matters

Imagine the team had only seen the combined 8% contribution margin. They might have tried reducing PPC everywhere, raising the price, or stopping the product. None of those decisions follows automatically from the data. Once the account is broken down:

• AmazonPositively contributing. Return and product-quality issue deserves investigation.
• NoonUnderlying economics before advertising leave room. Advertising consumes too much of that room.
• Inside Noon adsSome targets deserve protection. Some need lower bids. Some consuming clicks without sales should be blocked.
• ResultThe operator has somewhere specific to start — not a vague budget cut.

Same product does not mean same advertising target

This is also why generic ACoS or ROAS targets can mislead. A product may tolerate one level of advertising on Amazon and a very different level on Noon. The marketplace environment changes: competition, search demand, conversion, CPC, product visibility, organic demand, campaign structure, available targeting, and customer behaviour. The underlying product may be exactly the same. The operating answer does not have to be.

Should we move the budget from Noon to Amazon?

Maybe. But not automatically. Amazon is currently producing positive contribution for this product. Noon is not. That makes Amazon the stronger economic position during this period. But budget allocation should still consider whether Amazon has room to absorb more spend efficiently, whether the Noon campaign can be repaired, inventory available on each marketplace, organic demand, strategic importance of Noon, pricing, product ranking, and whether the business is deliberately investing in marketplace entry or visibility. The question is not: "Which marketplace won?" It is: "Where will the next dirham create the best business outcome?"

The combined number can hide the decision

Remember the starting point: AED 13,453 revenue, AED 1,070 contribution, 8.0% contribution margin. If management sees only that number, it looks like one product with weak economics. But underneath it: Amazon +AED 1,788 contribution, Noon −AED 718 contribution. The combined figure had not calculated anything incorrectly. It had simply averaged together different situations. That is why marketplace reporting needs to be unified without becoming blindly blended.

What should an operator check when the same SKU performs differently by marketplace?

1. RevenueHow much demand does the product actually have on each marketplace?
2. Product costIs the cost base genuinely the same across channels?
3. Marketplace feesHow much does each channel take out before advertising?
4. Advertising intensityHow much of total revenue is being consumed by ads on each marketplace?
5. Campaign structureIs spend concentrated behind useful targets or spread across weak ones?
6. Search terms and targetsWhich ones are generating orders? Which are consuming clicks without sales?
7. ConversionIs the traffic reaching the product actually converting?
8. ReturnsAre customers keeping the product? And if not, why?
9. InventoryDoes each marketplace have enough stock to support the strategy?
10. ObjectiveAre you trying to maximise contribution now, launch a product, gain visibility, acquire customers or establish the marketplace?

This is where the operating workflow becomes useful

The useful part of this example was not one dashboard showing Amazon and Noon side by side. It was the sequence.

So the workflow moved from Something is wrong → Where is it wrong? → Why might it be happening? → What should we change first? That is the point. The system can surface the evidence and narrow the action.

The real lesson from this product

Over 90 days, the same product family produced:

Before advertising, Noon actually had more margin available than Amazon. Advertising changed the outcome. But the answer was not simply to switch Noon advertising off. The shorter campaign analysis showed that some targets were still working while others were consuming money without producing sales. At the same time, the Amazon returns data surfaced a separate product-quality issue that advertising optimisation would never fix.

That is why the useful question is not: "Which marketplace is better?" It is: "Why is this product behaving differently here, and what can we actually change?"

Frequently asked questions