One Catalog. Three Marketplaces. Three Completely Different Problems

Run the same catalog across three marketplace channels at the same time and the economics stop agreeing with each other. Same brand, same products, same price points — wildly different returns across all three. That disagreement is one of the most useful signals available — and it only shows up when the channels are looked at side by side, which most day-to-day reporting isn't built to do.

Here's where this account started. Three channels, three failure types — not three numbers you'd politely skim past. One looked too good to question. Two looked weak enough that everyone had already decided it was just how the channel performs — and neither had the one number attached that would have made ignoring it obviously wrong: how much was actually being spent there.

That third row is the one worth sitting with. Noon Minutes wasn't a small, forgettable line — it was 62% of total ad spend, the largest single channel in the account, running at a number nobody would call good or bad enough to escalate. That's the most dangerous kind of blind spot: not a hidden problem, a visible, mediocre one, sitting on top of the most money in the account.

None of that is visible from inside a single channel's own dashboard. It only shows up when you put all three next to each other and ask why the same product converts at 16.1% on one platform and 10.8% on another. So that's what we did — audited all three, from the outside in. Here's what was under the hood.

Channel 01 — Noon CoreThe budget was gone before anyone was shopping

Noon Core had been running under a single automatic campaign, untouched, for 20 months. That alone isn't the story — plenty of campaigns coast. The problem was a setting that had only gone live in the last three to four months of that stretch, and it was quietly torching the account every single day.

A Top-of-Search bid multiplier was set at 200%. In practice that meant the campaign was paying roughly AED 10 per click for the premium top slot when the same conversions were available further down the page for around AED 3.60 — nearly 3x the cost for the identical buyer. And because the daily budget was finite, that inflated top-slot bidding drained the whole day's spend in the morning, on low-intent browsers. By the time the UAE's real shopping window opened in the evening, the account was dark. Every day.

The fix was one number. We pulled the multiplier down and stopped overpaying for a slot that wasn't converting any better than the cheaper ones. Then we rebuilt the single auto-campaign into a structured set — brand defense, generics, category — each with its own job and budget instead of one campaign trying to do everything.

ROAS moved from 1.04x to 2.70x in twenty days, with spend cut roughly in half. Six reporting cycles later it's holding at 3.43x. The point isn't "we found a bad setting." It's that the setting had only been live a few months — a channel that looks like it's always run at breakeven can actually be a recent, fixable break.

Channel 02 — Amazon UAEThe number looked great because it was paying for demand it already had

Amazon was the opposite problem. The topline looked healthy — strong blended ROAS, a solid 18% ACoS on brand terms. Nothing about it asked to be touched. Which is exactly why it needed to be.

Under the hood, brand defense was eating 59% of total ad spend — bidding hard on searches for the brand's own name. Those searches convert whether you pay for them or not; someone typing your exact product name has already decided. The campaign was buying customers who were walking in the door anyway, and the healthy ACoS was the receipt for demand the brand already owned.

So we tested it, deliberately. We stepped the biggest brand keyword's bid down and watched what happened.

On a dashboard, that's a disaster — a keyword you drove into the ground. In reality it's the proof. The paid clicks vanished and the customers didn't: store-wide revenue rose over the same window, TACoS compressed from 29% to 8%, and in the cleanest stretch of the test, every dirham of ad spend "lost" on that keyword came back as more than a dirham of organic revenue — as high as 4:1. We weren't losing customers. We were just going to stop paying to find the ones we already had.

The keyword going dark wasn't the ad failing. It was the ad finally getting out of the way of demand that was already there.

Freed from defending its own name, the budget went to work finding new buyers — category and competitor terms. And then the real test of the whole restructure arrived, unplanned.

When the hero SKU died, the account didn't

Mid-optimization, the account's single biggest product — a SKU driving more than half of total revenue — went out of stock. Revenue fell off a cliff in the same week the brand-defense test was running. Two drops, same seven days, same dashboard, and it would have been easy to blame the strategy for a fire that had nothing to do with it.

Telling them apart before reacting is the actual job. We checked the ROAS figure against three independent sources and threw out the one reading an inflated 5.56x — Amazon's console report hides low-traffic rows, which quietly biases the visible number upward. We traced the collapse to one SKU's fulfillable inventory hitting zero, not a listing problem and not a demand problem. And while that SKU stayed dark, we pushed spend to the other SKUs still in stock.

Revenue rebounded 107% off the trough — without the product that normally drives half the business. That's the real dividend of cutting the brand-defense crutch: an account that no longer lives or dies on a single hero SKU, because the budget knows how to flex to whatever's actually available to sell.

Channel 03 — Noon MinutesA cheap SKU was winning the auctions it should have been losing

Minutes looked the most ordinary of the three — a middling ROAS bouncing in a band, no alarm bells. But "ordinary" was hiding two separate problems that a single-period glance would never catch.

First, the channel's best-performing keyword had been switched off entirely — dark for three straight reporting periods before anyone noticed the account's strongest converter was simply gone. Second, and more interesting: a cheap AED 18 entry-price SKU was quietly swallowing the majority of the account's largest campaign, running at roughly 0.85x — losing money on every dirham — while a premium SKU worth six times as much sat in the same campaign converting at 10x.

That's not bad luck. When a cheap, clicky product and a premium one share the same campaign and keywords, the auction structurally favors the one that gets clicked most — and an AED 18 impulse buy gets clicked. The entry-price SKU wasn't winning because it was better. It was winning because it was cheap, and the shared campaign let it hoard the budget on that alone.

The fix isn't fewer products — it's separating them by price tier so the entry SKU and the premium SKU stop bidding into the same auction. The entry SKU keeps a real job: cheap customer acquisition, in its own lane. The premium SKU gets its own campaign and a bid tuned to win on the economics it actually has. Same SKUs, radically different budget split.

And while digging through Minutes, we found something nobody asked us to look for: nine Arabic-language search terms buried inside more than 1,400 low-value ones, converting well enough to deserve their own campaign. It's now running at 3.16x and climbing — a channel nobody knew they had.

Minutes isn't finished, and the raw ROAS number undersells what's actually happening. Ad spend on the channel came down 16% group-over-group — and total channel revenue barely moved, down just 3.6%, because organic revenue grew 11% in the same window. Organic share of the channel moved from 36% to 41%. That's the same pattern as the Amazon brand-defense test: pull back paid spend, and demand that was always there keeps showing up. TACOS — the cleaner signal here, same logic as before — improved from 24.9% to 21.6%.

The price-tier split is live. We'll update this section with the results once it's cycled through a full reporting period.

What changed, and where it's still changing

Same three channels, re-audited. Findings, fixes, and where each one stands today — because none of this is a one-time pass. The auditing doesn't stop once a fix lands.

Three platforms. Three completely different root causes — a bid setting bleeding out before peak hours, a healthy number propped up by demand it already owned, a cheap SKU hoarding a shared auction. None of them showed up on a topline dashboard. They showed up because someone audited every channel, including the ones that looked fine — especially those.

One drop was the plan working. One was a fire. Same account, same week, same chart. Telling them apart before you react is the entire discipline.

That's the work, and it doesn't stop. The findings compound because the auditing doesn't pause between quarters — it's how an embedded operator runs, not a report you get and file.