ACoS and TACoS explained: the benchmarks sellers should hit

The two formulas, and why the difference matters

Spend $250 and generate $1,000 of advertised sales, and your ACoS is 25%. If total sales that month were $4,000 including organic, your TACoS is 6.25%. Same spend, two very different questions answered.

ACoS answers: is this campaign paying its way? It counts only sales Amazon attributes to an ad click inside the attribution window, which is why it never reconciles with your total revenue and why comparing it to anything other than your own break-even is close to meaningless.

TACoS answers: is the account getting healthier? Because the denominator includes organic sales, a falling TACoS at steady spend means organic revenue is growing underneath the advertising. That is the single best signal that your listing and keyword work is compounding rather than renting.

The failure mode of managing on ACoS alone is worth naming. Cutting bids will almost always lower ACoS, because you stop paying for the expensive marginal clicks. It will also lower sales velocity, which lowers organic rank, which lowers organic sales — and you will not see that in ACoS at all, because organic sales are not in the denominator. Accounts optimised to a beautiful ACoS have a habit of shrinking quietly.

Where ROAS fits

ROAS is the same relationship inverted: ad-attributed sales divided by ad spend. An ACoS of 25% is a ROAS of 4. Neither carries information the other lacks, so use whichever your team argues in and be consistent. The thing to avoid is a report that quotes ACoS in one section and ROAS in another, which reliably produces a meeting where nobody is comparing the same thing.

Nobody publishes a real benchmark. Including Amazon.

This is the part most articles skip, and it changes how you should read all of them.

Amazon does not publish ACoS or TACoS benchmarks. Its own guide states plainly that "there isn’t a definitive number for a good Amazon ACOS. It’s dependent on your industry, company size, and campaign frequency, among other variables". That is the platform owner, with all the data, declining to name a figure.

So every category table you have seen comes from an agency or a tool vendor reporting its own platform data. Those are not worthless, but they are samples of whoever happens to use that tool, with methodologies that are rarely disclosed and sample sizes that almost never are. And they disagree with each other by margins large enough to invert a decision:

• One widely-cited agency table puts apparel ACoS around 42%. Another puts the same category near 29%.
• Published category ranges across the better tables run from roughly 8% at the low end to 50% at the high end — a spread so wide that "your category averages X" carries almost no information.
• A tool vendor publishing a platform-wide average ACoS around 30% is describing its own customer base, not the marketplace.

If two independent sources disagree by fifteen percentage points on the same category, neither is a benchmark. They are two samples.

What to use instead

Amazon now gives you a real one, and almost nobody has turned it on. Benchmarks reporting went generally available worldwide on 18 May 2026, covering eight metrics for Brand Registry advertisers including click-through rate, cost per click and cost per purchase. Its Sponsored Brands companion is the one that gives you peer context directly: it reports your performance, ACoS included, against the category median, the lower-performing quartile and the top quartile.

That is your category, your window, Amazon’s own data. It beats any published average, and it is already in the console you are paying into.

The only number that actually matters: your break-even ACoS

Break-even ACoS is your contribution margin expressed as a percentage. Spend more than that on advertising and each advertised sale loses money; spend less and it makes money. It is arithmetic, not a benchmark, and it takes ten minutes.

Work down from the selling price:

Referral fees vary widely by category — Amazon’s published schedule runs from 5% to 45% with a $0.30 minimum — so use your own rate rather than the 15% in this example. If you are VAT-registered and selling into the UK or EU, strip VAT from the price before you start; Amazon reports sales gross and the VAT portion was never yours.

Two things follow from having this number that do not follow from any benchmark.

It tells you who you are competing against. A brand on a 45% contribution margin can profitably outbid a brand on 20% all day. If your category is dominated by manufacturers and you are a reseller, the gap in break-even is the real story, not the gap in tactics.

It converts a target into a decision. "Is 32% ACoS good?" has no answer. "Is 32% below our 42.5% break-even?" has one, and it is yes, so the campaign is profitable and the question becomes whether to spend more.

Working ranges by stage

With the caveat above firmly in place — these are our own planning ranges, not published benchmarks — here is roughly where we expect an account to sit at each stage, and more importantly what each range means.

Read the TACoS column first. It is the one that tells a story: 25–40% at launch means most of your revenue is bought, 5–10% at maturity means most of it is earned. An account that has been running two years and still sits at 25% TACoS has not built anything — it is renting its position, and the rent goes up every year.

When a high ACoS is the right answer

Three situations, all of them deliberate.

Launch. You are buying sales velocity to establish relevance and rank on terms you intend to own organically. A 70% ACoS in month one that produces organic position on eight commercial terms is a good trade. The mistake is failing to define when it ends.

Defending a term you own. If a competitor is bidding on a keyword where you hold organic position one, conceding the paid slot above you costs more than the inflated ACoS of holding it.

Clearing stock. Storage fees and ageing inventory have their own cost. An ACoS above break-even on a product you need gone is not a loss, it is a cheaper exit than long-term storage.

When a low ACoS is a warning

A 6% ACoS almost always means bids are too low to win meaningful volume. It is profitable on nearly nothing, and it usually coexists with flat or falling total sales. If ACoS is excellent and revenue is not growing, you have optimised the ratio rather than the business.

It can also be an artefact. A very low ACoS on a brand-name campaign is mostly measuring people who were already going to buy from you, and reading it as advertising performance flatters the whole account. Split branded and non-branded reporting before you draw any conclusion from a number that looks too good, because a blended figure carried by branded traffic hides whatever the acquisition campaigns are really doing.

Reading the TACoS trend, which is where the money is

TACoS is only useful as a direction of travel. A single month’s figure tells you almost nothing; three months of it against stable spend tells you whether the account is compounding. There are four shapes it makes, and each one means something specific.

TACoS falling, total sales rising. The one you want. Organic revenue is growing faster than ad spend, which means the listing and keyword work is landing and advertising is doing less of the heavy lifting each month. When you see this, the correct response is usually to spend more, not less — you have found something that compounds and you are underfunding it.

TACoS falling, total sales flat. You cut spend and organic did not move. This looks like efficiency and is usually contraction: you have stopped buying marginal sales without replacing them. Check organic rank on your main terms before congratulating anyone.

TACoS rising, total sales rising. Growth bought with money. Legitimate during a launch, a new marketplace or a seasonal push, and a problem if it has been true for six months, because it means every increment of revenue costs more than the last and nothing is accumulating.

TACoS rising, total sales flat or falling. The alarm. Spend is going up and nothing is coming back. Almost always either a conversion problem on the detail page or a competitor who has entered the auction with a better margin than yours.

The measurement mistakes that waste the metric

Comparing across ASINs at different life stages. A launch product at 30% TACoS and a mature one at 7% are not comparable, and averaging them produces a number that describes neither. Track TACoS per product family, then roll up.

Reading it weekly. Attribution windows, restocks, promotional spikes and the ordinary lumpiness of retail all move a weekly figure far more than your management does. Monthly is the shortest honest period.

Forgetting that a price change moves it. TACoS has revenue in the denominator, so discounting raises it and a price increase lowers it, neither of which has anything to do with advertising. If you ran a coupon, say so in the report.

Why your ACoS suddenly jumped

In rough order of likelihood, and each has a different fix.

Your conversion rate fell, not your ad efficiency. ACoS has two inputs and most people only look at one. If a competitor cut price, your reviews took a hit, you went out of stock on a variation, or your main image changed, the same clicks now produce fewer sales and ACoS rises without a single bid changing. Check conversion rate before you touch bids.

You lost the Featured Offer. Ads do not serve when you are not winning the Buy Box, and on a shared listing the spend can shift to whichever of your ASINs still is — usually a worse-converting one.

Seasonal competition. Cost per click is an auction, and in Q4 or during a Prime event the auction fills with advertisers who will accept a worse ratio than you. Your ACoS rising in the second week of November is not a campaign problem.

Automatic campaigns drifted. Without regular search term harvesting and negatives, auto campaigns spread into adjacent, poorly-converting queries. This is the most common cause in accounts nobody has touched for a quarter.

A new competitor entered. Someone launching in your category is running the 50–80% launch ACoS in the table above, and they are bidding against you with a budget that does not need to make sense yet.

The diagnostic order is: conversion rate, then Featured Offer, then search terms, then bids. Most people start at bids, which is the one lever that cannot fix any of the first three.

Two structural causes people miss entirely

Attribution shifting between campaigns. If you launched a Sponsored Brands or Sponsored Display campaign recently, some sales that Sponsored Products used to be credited with are now credited elsewhere. Sponsored Products ACoS rises, nothing has actually got worse, and the only way to see it is to look at blended performance across all campaign types rather than one line.

A change in placement mix. Top-of-search clicks cost meaningfully more than product page placements and convert differently. If your bid adjustments or the auction pushed a larger share of your impressions to top of search, your average cost per click rises and ACoS follows, without any change to your keywords or your listing. Check the placement report before concluding anything about targeting.

What to change, and in what order

Fix the detail page before the campaign. If conversion rate is the cause, no bid change recovers the ratio, it just buys fewer clicks at the same bad conversion. Then apply negatives to the search terms that spent without converting, which is almost always the cheapest single win available. Then lower bids on keywords that convert but above break-even, rather than pausing them — pausing a term you rank for organically can cost you the rank as well as the paid position. Bid changes should be the last step and should move in increments of ten to fifteen percent, not halvings, because the auction needs a few days to settle before the new figure means anything.

Setting a target you can defend

A target ACoS is a business decision, not a marketing one, and it should be written down before the quarter rather than explained after it.

If you want profit now: target roughly 70–80% of break-even. On the 42.5% worked example above, that is 30–34%. It leaves genuine margin per advertised sale and accepts that you will not win every auction.

If you want growth: target break-even itself, and judge the quarter on total revenue and TACoS rather than on ACoS. You are converting margin into position on purpose.

If you are launching: ignore ACoS for the first thirty days entirely. Set a total budget you are willing to lose, watch organic rank on your target terms, and review at day 30. Our launch process treats the first month as data collection explicitly, because a launch judged on ACoS in week two always gets throttled before it has learned anything.

Review on a schedule, not on a feeling

Weekly: search term report, negatives applied, obvious bid corrections. Monthly: ACoS against target by campaign, and TACoS for the account. Quarterly: recalculate break-even, because fees and cost of goods move — Amazon raised FBA fees by an average $0.08 per unit from 15 January 2026, and announced no new FBA fee types. A break-even calculated before that date is now slightly wrong.

If you would like the arithmetic done against your actual numbers, our profit and break-even calculator is free and takes about a minute, and a written audit comes back within 48 hours with your targets set against your own margin rather than someone else’s category average.

Questions people also ask