Amazon Ads vs MercadoLibre Ads: the differences that actually change your operation

Any seller working both channels eventually asks the same question: why does my advertising perform so differently on Amazon than on MercadoLibre if it’s the same product and the same budget? The comfortable answer is “Meli just converts differently”. The useful answer is that you aren’t running two versions of the same tool: you’re running two systems whose buying logic, controls and reports do not line up.

That confusion has a concrete cost. The structure that worked on Amazon — campaigns split by intent, term harvesting, negatives so you don’t bid against yourself — can’t be copied over to Mercado Ads, because the pieces holding it up don’t exist the same way on the other side. And in reverse: the simplicity of setting a target and letting the algorithm work doesn’t transfer to Amazon without giving up exactly the control that makes the difference there.

There’s also a reporting trap that sinks a lot of decisions: the ACoS from both platforms looks like the same number, gets placed side by side in a spreadsheet and gets compared. They are not directly comparable. Attribution windows differ, what each platform counts as an “attributed sale” differs, and what gets deducted from price before reaching your pocket differs too.

This post isn’t trying to decide which platform is better — that question almost never has a general answer. It’s trying to order the differences that genuinely change how you operate: campaign types, control, available signals, attribution and reporting. And then close with the one thing that is comparable across channels: what’s left over for the product after the campaign is paid for.

two philosophies: buying terms or buying a target

The root difference is what you are buying when you put money in.

On Amazon you buy terms and placements. You can declare a keyword with its match type and bid, target a specific competitor ASIN, or let the automatic campaign propose and then harvest what worked. The whole system is built around the per-query auction, and your job is deciding which queries you enter and at what price.

On Mercado Ads, by contrast, Product Ads revolves around an efficiency target and a budget. You pick a strategy — typically presented as profitability, growth or visibility, each tied to a target efficiency level — and the system decides where and when to show your listings inside Meli’s ecosystem. There is no keyword list you administer listing by listing.

This doesn’t make one platform “better”. It makes the seller’s job different: on Amazon the job is selection and pruning; on Meli the job is product selection and target calibration. Forcing a keyword strategy onto Meli is pushing against the system’s design, and running Amazon as if it were a black-box target means giving up the only real advantage it offers.

One important note, because this is the ground that shifts most: Mercado Ads has been changing how that target is expressed — the traditional target-ACOS setting has been giving way to a target expressed as ROAS — and Amazon adds and renames controls frequently. Before rebuilding anything, open your Mercado Ads campaign manager and your Amazon Ads console and confirm what options you have, today, in your account and your country.

the campaign types you have today on each platform

On the Amazon side, the family revolves around three self-service formats plus programmatic buying. Sponsored Products promotes individual listings and is where nearly all of a mid-sized seller’s spend lives; it runs in automatic or manual mode. Sponsored Brands takes prominent placements with your own creative — headline, image or video — and normally requires a registered brand. Sponsored Display works by audience and by product, on and off Amazon, and is the natural format for retargeting. Above all that sits DSP buying, which is another conversation entirely.

On the MercadoLibre side, Mercado Ads groups three products with analogous but not identical roles. Product Ads is the performance one: it promotes listings and is the functional equivalent of Sponsored Products. Brand Ads takes premium positions in search results and is built for brands with an official store. Display Ads covers banners and high-impact formats with audience targeting inside the ecosystem.

The mental mapping is tempting — Product Ads is Sponsored Products, Brand Ads is Sponsored Brands, Display Ads is Sponsored Display — and it helps you get oriented, but it breaks the moment you go into detail. Access requirements aren’t the same, placement inventories aren’t the same, and above all the degree of control they hand you isn’t the same.

I’ll repeat the obvious because it costs money: format catalogs change, country pilots launch, and options get retired without much notice. Check each console before planning a whole quarter around a format you may not even have enabled.

how much real control you have over spend

This is the difference you feel most in daily operations.

On Amazon control is granular to an uncomfortable degree: bid by keyword, bid by targeting group in automatic campaigns, placement adjustments, budget per campaign, negatives at campaign and at ad group level. You can decide that exact pays more than broad on the very same term. That granularity is an advantage only if you use it; otherwise it’s an endless source of unsupervised spend.

In Product Ads control is aggregate: budget, strategy, efficiency target and which listings are included. Fine calibration isn’t in your hands, but neither is the maintenance of hundreds of bids. In exchange, the weight of the decision moves to two earlier questions: which products deserve budget and what target is realistic for each one’s margin.

The practical consequence is that the same mistake is paid for differently. On Amazon, a bad bid setup is paid in expensive clicks on irrelevant terms, and it’s fixed with negatives. On Meli, a target that’s too aggressive is paid in missing exposure, and one that’s too loose is paid with an acquisition cost that eats contribution without any keyword being to blame.

The intervention rhythm changes too. A well-run Amazon account gets touched weekly. A Product Ads campaign fiddled with every day never learns: constantly changing the target restarts the system’s learning and produces numbers that mean nothing.

Glossary: ACoS is ad spend divided by the sales attributed to those ads; it measures spend efficiency, not profit.

what signals each one gives back

An ad platform doesn’t just spend: it reports. And what it reports defines what you can optimize.

Amazon gives you the search term report, the single most valuable asset in the whole operation: the literal list of what people typed before buying. That’s where keyword harvesting comes from, where the negative list comes from, and a good chunk of your listing SEO work. It also gives you impressions, clicks, CTR, CPC and attributed sales by term and by placement.

Mercado Ads gives you a different set of signals. Alongside impressions, clicks and conversions, it explicitly separates direct sales — someone clicked your ad and bought that product — from indirect sales — sales of your promoted listings that arrived by another route, typically organic or through another ad type. That separation is more useful than it looks: it’s the raw material for reasoning about advertising’s total effect on a listing, which on Amazon you have to reconstruct yourself by crossing ads against total sales.

What you don’t get on the Meli side is the granular query list you use on Amazon to build keyword strategy. That absence isn’t an oversight: it’s consistent with a system where you don’t buy terms. But it means the discovery lever that lives in the automatic campaign on Amazon lives somewhere else on Meli: in the title, the attributes and the listing itself.

attribution and reporting: why the two ACoS numbers don’t compare

This is the difference that produces the most bad decisions, because it’s invisible in a spreadsheet.

On Amazon, Sponsored Products attributes sales within a click window that for seller accounts runs seven days, while Sponsored Brands works with fourteen. That alone means the same product advertised in two formats shows different efficiency without either one “performing better”: one simply has twice as long to collect conversions.

In Product Ads, the documented attribution window is fourteen days, and attribution prioritizes the click over the impression. In other words: even though both reports show you a box labeled ACoS, the numerator and the denominator weren’t built with the same rules or the same time horizon.

The net effect is that a 20% ACoS on Meli and a 20% ACoS on Amazon are not the same economic fact. Putting them side by side and concluding “Meli is more efficient” is comparing two measurements taken with different clocks. And the most important part is still missing: neither figure knows what category referral fee, what fulfillment fee or what withholding was deducted before the money reached you, and those genuinely do differ between channels.

The practical rule: compare each campaign against itself over time within its own platform, and compare channels only in the currency that really is shared — the money left per unit sold.

a campaign is judged against contribution, on both channels

The only honest common denominator across platforms is the contribution of the product the campaign finances.

An example with invented but consistent numbers. A product you sell for 1,160 pesos including VAT: the pre-tax price is 1,000. Suppose on channel A the category fee takes 130 and fulfillment 95; on channel B, the fee takes 160 and logistics 70. Product cost is 420 on both and the returns provision is 25. On channel A you’re left with 330 of contribution before advertising; on channel B, 325. Nearly identical in this made-up case, but the point is the method: that number, not ACoS, is the ceiling of what each channel can pay in ads.

With that base, comparison becomes real. If you run at 18% ACoS on Amazon you keep 15 points of contribution per unit; if the same product runs at 24% on Meli, you keep roughly 8.5. Now you can say which channel leaves you more per unit sold, because you’re comparing money that survived every deduction, not percentages calculated under different rules.

This calculation demands something that sounds boring and turns out to be decisive: that the same physical product be identified as one thing even though it carries an ASIN on one side and an MLM on the other. Without that unification, each channel hands you its own version of the truth and there’s no way to add them up.

Glossary: a unified catalog ties the same real product to its listings on every channel, so sales, stock and costs add up without duplication.

how this reads in iqseller

In the panel, the starting point isn’t the campaign but the product. The Parent → Model → SKU tree ties the same item to its listings on each channel, so the Amazon ASIN and the MercadoLibre MLM stop being two separate stories. On that base, the Profitability module builds net margin per SKU and per channel from the COGS you load, the commissions that come from the Amazon settlement and from MercadoLibre orders, FBA and Full fees, shipping, and the VAT and withholding breakdown.

That’s the number that makes two platforms with different reporting comparable. Once ad spend is deducted against each channel’s real contribution, the question stops being “where do I have the better ACoS?” and becomes “where do I keep more per unit after paying for the ads?”. Those are different questions and they often have opposite answers.

The Alerts module helps with the other half of the problem, the half that gets mistaken for advertising and isn’t: efficiency that degrades overnight usually comes from a price that moved or real available stock that ran out, not from the campaign. With inventory, price and advertising on the same board, the diagnosis takes minutes instead of an afternoon of exports.

what changes in your operation when you run both

Three things, in order of importance.

• Don’t transfer structures, transfer criteria. What copies from one channel to another isn’t the campaign tree: it’s the spend ceiling each product can pay. That ceiling comes from margin, and margin travels between channels even when platforms don’t.
• Separate the intervention calendars. Amazon wants frequent review of terms and bids; Product Ads wants patience so the system can learn, with spaced-out target adjustments. Touching both at the same rhythm damages one of them.
• Close the reporting in one place. As long as the answer lives in two consoles plus a spreadsheet, you’ll be deciding with numbers calculated under different rules. The common close isn’t ACoS: it’s profit per unit and per channel.

And the usual warning, because it applies to both sides: formats, controls, configuration metrics and attribution windows change. Before committing peak-season budget, confirm in Seller Central, in your Amazon Ads console and in the Mercado Ads campaign manager what you actually have available today. A strategy built on an old assumption degrades without warning, and in advertising the warning arrives as an invoice.