ROAS benchmarks: the number you actually need, not the industry average

ROAS is the easiest metric to brag about and the easiest to misread. “My campaign is doing 4x” sounds like a clear win, and in many businesses it is. In others, that exact same 4x means every advertised sale returns less money than it cost to bring in. The difference is not in the campaign. It is in the margin of the product you are advertising.

That is the underlying problem with the ROAS benchmarks floating around. They all hand you a number — 3x, 4x, “anything above 2.5x is fine” — while knowing nothing about your product cost, your category commission or your fulfillment rate. It is like being told how fast to drive without anyone knowing whether you are on a highway or a school street.

On Mexican marketplaces the gap widens, because between the price the shopper sees and the money that lands in your account there are four layers of cost no ad platform knows about. The ROAS Amazon Ads or MercadoLibre Product Ads reports to you is calculated on gross sales, not on what you keep.

This article does two things. First it reviews which benchmarks actually exist, which market they come from and how solid their methodology is. Then it walks you through the only number you really need: your break-even ROAS, derived from your own contribution margin, which tells you the point where your ad starts making money instead of just moving it.

what ROAS measures and what it does not

ROAS stands for return on ad spend: revenue attributed to advertising divided by advertising spend. Invest $10,000 and have the platform attribute $40,000 in sales to it, and your ROAS is 4.0, or 4x.

It is the exact inverse of ACoS. A ROAS of 4.0 equals a 25% ACoS; a ROAS of 2.5 equals a 40% ACoS. The math is simple: ACoS = 1 ÷ ROAS. They carry the same information, and the preference is cultural — the Amazon world leans on ACoS, the Meta and Google world leans on ROAS.

What matters is what neither of them measures:

• Neither measures profit. The numerator is gross sales, before product cost, commission, fulfillment and returns.
• Neither measures incrementality. Some of those attributed sales would have happened anyway, especially when you bid on your own brand.
• Neither measures halo effect. The organic sales your advertising pulled along do not appear in the numerator, which is exactly why TACoS exists.
• Neither measures customer lifetime value. A 1.8 ROAS on first purchase can be excellent if the customer buys three more times.

With those four caveats on the table, benchmarks can be discussed without doing damage.

Glossary: ACoS, the share of the sale that goes to advertising →

which benchmarks actually exist

Honesty about data quality matters here, because ROAS is one of the metrics where the most figures circulate and the least methodology gets published.

For ecommerce in general, the most frequently cited measurements put average ecommerce ROAS at around 2.87x in 2025, with a median near 2.04x in 2024. The gap between average and median is the interesting part of that data: it means a handful of very high-ROAS accounts pull the average up, and the typical store sits considerably below what the public conversation suggests. These are figures from mostly US ecommerce, primarily Meta and Google, not from Mexican marketplaces.

By platform, the same aggregators report ranges that diverge sharply: paid search above display, Meta in the 2x to 4x band, TikTok consistently lower. Polar Analytics, which publishes medians across more than 4,000 Shopify brands with data going back to December 2022, reports new-customer ROAS by industry — a more useful cut than blended ROAS, because it isolates acquisition from repeat purchase — also across brands selling through their own store.

For Amazon Ads specifically, the ranges published by tools and agencies place Sponsored Products at 20% to 35% ACoS and roughly 3x to 5x ROAS, depending on category and account maturity. Treat them carefully: almost none of them publish sample size, period or country, and the differences between one source and the next are enormous. They tell you the order of magnitude your category moves in, not a target.

And what does not exist: no public ROAS cut for Mexican sellers on Amazon México or MercadoLibre Product Ads. The AMVO Estudio de Venta Online 2026 publishes market and shopper behavior data in its open version, but the operational performance cuts are for members only. Any number presented as “the average Mexican seller’s ROAS” did not come from a published study.

why a “good” ROAS can still lose money

Here is the heart of the matter. Suppose your Sponsored Products campaign reports a 4.0 ROAS and your category benchmark sits around 3.5. You are above average. Congratulations.

Now land one sale. Price $899. Your COGS is $340. The 14% category commission takes $125.86. The fulfillment fee is $85. You reserve 3% for returns, $26.97. Added up, variable costs excluding advertising come to $577.83. Your contribution margin before advertising is $321.17, or 35.7% of the price.

A 4.0 ROAS means you spent a quarter of the sale on advertising: $224.75. Since your contribution before advertising was $321.17, you are left with $96.42 per unit. You survive, but storage, software, payroll and taxes still come out of that. That comfortable 4x left 10.7% of the price.

Now change the product. An electronics SKU with an 18% contribution margin before advertising: the same 4.0 ROAS spends 25% of the sale on ads against 18 points of contribution. You lose seven points on every advertised sale. The ROAS is above the industry benchmark and the campaign is still funding sales with your money.

That is the point: the industry benchmark does not know your margin. A 4x is excellent for a brand at 45% contribution and a disaster for a reseller at 18%. Someone else’s average cannot decide your budget.

your break-even ROAS: the formula

Break-even ROAS is the point where the advertised sale neither gains nor loses. It comes from one division:

• Break-even ROAS = 1 ÷ contribution margin before advertising

Where contribution margin before advertising is, in decimals, what is left of the price after COGS, commission, fulfillment, shipping and the returns reserve — everything variable except advertising.

With the examples above:

• 35.7% contribution → break-even ROAS = 1 ÷ 0.357 = 2.80. Below 2.80x you lose money on every advertised sale.
• 18% contribution → break-even ROAS = 1 ÷ 0.18 = 5.56. That 4x, which looked perfectly healthy, is well below break-even.
• 45% contribution → break-even ROAS = 1 ÷ 0.45 = 2.22. Here a 2.5x already makes money.

Three products in the same business, three thresholds ranging from 2.22 to 5.56. No industry benchmark table can give you that, because the number depends on your cost, your category and your channel.

Break-even ACoS is the mirror image: 35.7% contribution means a 35.7% break-even ACoS. Same math, same conclusion.

how to use break-even ROAS in practice

Knowing the threshold is half the job. The other half is deciding what to do with it, and the answer is rarely “aim exactly at break-even.”

• Set a target above break-even, not at it. If your break-even ROAS is 2.80, aiming precisely at 2.80 leaves you at zero. A reasonable target leaves enough room to pay fixed costs — say, a ROAS target that still leaves eight or ten points of contribution after advertising.
• Calculate it per SKU, not per account. An account-level break-even ROAS blends 45% products with 18% products and pushes you to raise budget exactly where you should not.
• Recalculate it per channel. The same product carries different commissions and fulfillment on Amazon than on MercadoLibre, so it has two different thresholds.
• Go below it on purpose, with an end date. During a launch, a ranking recovery push or a review-gathering phase, operating below break-even makes sense. What does not make sense is doing it unknowingly and without an exit date.
• Revisit it whenever a cost changes. A supplier increase or a fulfillment rate change moves the threshold. If the campaign target does not move with it, you are optimizing against an old number.

And a warning about attribution: the ROAS the platform reports includes sales that might have happened anyway. Whenever you can, check spend against the SKU’s total sales, not only against attributed sales.

Glossary: real net margin, with everything deducted →

how this reads in iqseller

The practical obstacle to calculating a break-even ROAS per SKU is not the formula: it is having an up-to-date contribution margin, per product and per channel. That figure lives split across the Amazon settlement, MercadoLibre orders and your costing sheet.

In iqseller that is the part that gets solved. The Profitability module builds the waterfall per SKU, starting from the COGS you load and subtracting commissions from the Amazon settlement and from MercadoLibre orders, FBA and Full fees, shipping, advertising, and the VAT and withholding breakdown. Alongside it, the advertising metrics — ACoS, ROAS, spend, CPC — read against that same margin instead of against gross sales, and that is the comparison that changes the decision.

Because the catalog is organized in a Parent → Model → SKU tree, you can drill down to the variant burning budget above its threshold, or move up to the family to see whether the whole category can carry the investment. Alerts cover the most expensive case: the SKU whose margin moved and left the campaign operating below break-even without anyone touching a thing.

the number that is actually yours

ROAS benchmarks are good for one thing: giving you context on the order of magnitude your category moves in, and flagging when you are far outside the range. As a target they are useless, because they are calculated on businesses with a different cost structure, in other markets, almost always without published methodology and almost never on Mexican marketplaces.

The number that is actually yours is break-even ROAS, and it comes from a division you can run today with data you already have. Calculate it per SKU and per channel, add the cushion you need to cover fixed costs, and use that threshold — not the industry average — to decide where to raise budget and where to switch off.