Your Top-Selling SKU May Not Be Your Most Profitable. Here's What to Check.

A product is selling well.

It is bringing in revenue. Ads are generating orders. It may be one of the products your team looks at first when reviewing marketplace performance.

So the obvious question is: Should we put more money behind it? Not necessarily.

When you look at profitability at SKU level, a product that looks strong on revenue can tell a different story after product cost, marketplace fees and advertising are taken into account. And once inventory enters the picture, there is another question: should we actually buy more stock, or do we already have enough stock sitting somewhere else?

We saw both situations while reviewing a real UAE marketplace account across Amazon and Noon. The customer and product names have been removed, but the numbers below come from the actual account over a 90-day period.

The key takeaway

High revenue does not automatically mean high profitability. Before scaling a marketplace SKU, look at what remains after product cost, marketplace fees and advertising. Then look at where the inventory is.

• Revenuetells you how much you sold.
• Contributiontells you what the sale left behind.
• Inventorytells you whether the next decision should be to buy, move or hold stock.

The account's top-selling Amazon SKU was not its highest-contributing SKU

Let us start with the number-one individual SKU by Amazon revenue during the 90-day period.

AED 12,276 in revenue is a meaningful result. The advertising numbers do not immediately look alarming either — ROAS was 4.46x, TACoS was only 7.4%. Yet the product left AED 2,800 in contribution. Now compare that with another SKU in the same account. It generated lower revenue — approximately AED 10,113 — but left approximately AED 3,359 in contribution, at a 33.2% contribution margin.

The SKU producing the most revenue was not producing the most contribution. That is the first reason revenue rank and profit priority should not automatically be treated as the same thing.

Where was the money actually going?

For the top-selling SKU, the economics were roughly:

Another way to look at it: for every AED 100 of revenue, approximately AED 23 went to product cost, AED 47 to marketplace fees, AED 7 to advertising, and AED 23 remained as contribution.

That changes the diagnosis. If the team looked only at revenue and advertising, it might ask: "This product is selling well. Should we increase PPC?" If the team looked only at contribution margin, it might ask: "Should we reduce advertising?" But neither gets to the main issue. Advertising was not the largest cost line. The recorded marketplace fee line was.

So the first investigation should not automatically be "How do we cut ad spend?" It should be "What is actually sitting inside this fee burden, and is there anything we can change?"

That may involve referral fees, fulfilment charges, refunds, promotions, storage-related charges or other marketplace costs. The SKU-level margin view available here gives us the total fee burden, not every individual fee component, so it would be wrong to guess which specific Amazon fee caused it. The useful thing is that we now know where to investigate — already a much better starting point than simply telling the PPC team to reduce ACoS.

Was advertising actually the problem?

Not obviously. For this SKU:

Advertising should still be reviewed, particularly before more budget is added. But the numbers do not point to advertising as the first or biggest source of margin pressure. That distinction matters. A product can have reasonable advertising economics and economics that are less attractive than another product in the same catalogue — this is why isolated metrics can be misleading.

• ROASHow much ad-attributed revenue you generated for each dirham of ad spend.
• ACoSHow much of ad-attributed sales was consumed by advertising.
• TACoSHow much of total product revenue was consumed by advertising.
• ContributionWhat remained after the costs included in the analysis.

The next decision was not only about advertising. It was about inventory.

A high-selling SKU does not just attract more advertising budget. It usually gets more inventory attention too. When a product is ranked number one for revenue and stock cover starts looking tight, the instinct is often: reorder it quickly.

The problem was not "We do not have enough product." It was "The product needs to be in the right place." That is why profitability and inventory should not be managed as separate conversations. The question is not simply: Is this selling fast enough to reorder? It is: Is this SKU worth keeping in stock, how much stock do we already have, and where should that stock sit?

The decision gets easier when these numbers are looked at together

A marketplace operator would normally need to piece together several different views to answer these questions. Sales tell you the product is moving. Advertising tells you how much paid demand is supporting it. Settlement and fee data tell you what marketplace costs are taking out. Inventory tells you what is available and where. Returns tell you whether sales are coming back.

The useful part is not putting all of those numbers into one dashboard. It is being able to move from one signal to the next decision:

• Revenue saidthis is the top-selling SKU.
• SKU economics saidit is contributing, but less efficiently than some lower-revenue products.
• Cost analysis saidthe fee line deserves more attention than advertising.
• Inventory saiddo not automatically place a new PO. Move existing stock first.

That is a much more practical operating conversation — and it does not require someone to manually rebuild the analysis from separate advertising, settlement and inventory spreadsheets every time the question comes up.

A high-selling product can still deserve a different priority

This does not mean the top-selling SKU is a bad product. It is producing positive contribution. It has genuine demand. And it deserves to remain available. The point is that "top seller" is not enough information to decide how aggressively to scale it.

Another product may generate less revenue but leave substantially more contribution from each dirham sold:

Which one should receive the next AED 5,000 of advertising? There is still not enough information to answer that conclusively — we would also want to understand available inventory, incremental advertising opportunity, conversion, marketplace position, product objective, pricing, returns and growth potential. But one thing is already clear. Revenue alone cannot answer the question.

The same problem becomes even harder across marketplaces

Another product family in the same UAE account shows why. Over the same 90-day period, it generated AED 13,453 in combined Amazon and Noon revenue with an overall contribution margin of only 8.0%. At the combined level, it looks like one weak-margin product. But split it by marketplace and the picture changes.

Same product family. Same period. One marketplace produced positive contribution. The other produced a loss. That does not mean Noon is inherently the problem — it means the product needs a different diagnosis on Noon. In this case, advertising clearly deserves attention. And when we later looked inside the Noon advertising, the problem was not simply "ads are bad": some targets were producing orders, others were consuming spend without producing sales. That creates a more useful action: reduce the waste, preserve what is working, and decide whether the remaining economics justify further investment.

Returns can change the answer too

The same product family also had a 7.7% return rate across the 90-day combined view. One Amazon child SKU showed a return rate of 13.3%. When we drilled into that SKU, some of the recorded return reasons included: defective, quality unacceptable, not compatible. One customer comment specifically referred to the product leaking.

That matters. If a SKU has weak economics because advertising is inefficient, the answer may sit in campaign optimisation. If customers are returning the product because of a product-quality or expectation problem, increasing advertising could simply push more demand into the same issue.

The wrong question is "Which metric is red?" The right question is "Why is it red?"

There is no universal "good ACoS"

This is why generic advertising benchmarks need to be used carefully. There is no single ACoS or ROAS target that makes sense for every product. A SKU with plenty of margin before advertising can support more ad spend. A product carrying a high fee burden or high returns may not have the same room. Use benchmarks to ask: "Is this number unusual enough to investigate?" Use your own unit economics to answer: "Does this number actually work for this product?"

What should you check before scaling a marketplace SKU?

Before putting more advertising or inventory behind a successful-looking product, check the full decision.

1. RevenueIs the product genuinely selling well over a meaningful period? Do not make the decision from a few good days.
2. ContributionAfter the costs you are measuring, what is actually left? And how does that compare with other products competing for the same capital?
3. Product costHow much of every sale disappears before marketplace and advertising costs even begin?
4. Marketplace feesWhat percentage of revenue is being consumed by marketplace-related costs? If unusually high, investigate before assuming PPC is the problem.
5. AdvertisingLook at ACoS, TACoS and ROAS in the context of the product's margin, not in isolation.
6. ReturnsAre customers keeping the product? If return rates are high, understand why before paying to generate more demand.
7. Contribution per unitWhat does another unit sold actually add to the business?
8. Marketplace differenceDoes the same product produce the same economics on Amazon and Noon? Do not assume it should receive the same strategy everywhere.
9. Inventory positionHow much stock exists across the entire network? A marketplace may look close to stockout while adequate inventory is sitting in a local warehouse.
10. The objectiveAre you optimising for contribution, launch, visibility, market entry, stock movement or another deliberate business goal?

The real lesson from this UAE account

The useful insight was not that the top-selling SKU is bad, or that marketplace fees are always too high, or that advertising should be cut whenever contribution looks weaker than expected. The lesson was more practical.

The number-one revenue SKU produced AED 12,276 revenue and AED 2,800 contribution. A lower-revenue SKU generated AED 10,113 revenue and AED 3,359 contribution. For the top-selling SKU, advertising represented only 7.4% of total revenue, while recorded marketplace fees represented 46.7%. And when inventory cover looked tight, the answer was not another purchase order — there was already stock in the warehouse.

Those are different decisions coming from the same SKU. That is why marketplace profitability should not stop at revenue, ACoS or ROAS. The answer may be to investigate fees, change advertising, fix a return problem, move inventory, reorder, change marketplace allocation — or simply not push this SKU harder yet. The job is to identify which one.

Frequently asked questions