What Is ACoS on Amazon? Meaning, Formula, Good ACoS, and Break-Even

ACoS (advertising cost of sales) is the share of your ad-attributed sales that went back into Amazon ads to win them, calculated as ad spend ÷ ad sales × 100. A 20% ACoS means you paid 20 cents in ads for every dollar of sales Amazon credits to those ads, and inside Amazon Ads you'll see it written in capitals, as ACOS.

Whether 20% is good depends on a number your ads console doesn't have: your margin. And the usual fix for a high ACoS, spending less, can cost more than it saves. On the $40 kitchen scale below, cutting ACoS from 20% to 15% loses $525 a month.

ACoS formula: how do you calculate ACoS?

Divide ad spend by ad sales and multiply by 100. That's Amazon's own ACOS formula, and it works the same for a keyword, a campaign, or the whole account.

ACoS = ad spend ÷ ad sales × 100

Take a $40 kitchen scale sold through FBA in Home and Kitchen. Last month its campaigns spent $1,500 and Amazon credited them with $7,500 of sales, so $1,500 ÷ $7,500 × 100 gives a 20% ACoS. Is that good? You can't say until you know what the scale costs to sell. The referral fee is Amazon's 15% for the category, and we made up the rest.

So each sale leaves $14, or 35% of the price, before a cent goes on ads. At last month's 20% ACoS, the ads took $8 of every $40 ad sale and the scale kept $6.

ACoS vs ROAS vs TACoS: differences and how to convert

The difference between ACoS, ROAS, and TACoS is which sales figure sits next to ad spend, and which way up the fraction goes. ROAS (return on ad spend) is ACoS turned upside down, ad sales ÷ ad spend, which is why Amazon calls it the inverse of ACoS. TACoS (total advertising cost of sales) keeps ad spend on top but divides it by everything the product sold, through ads and organically. For the kitchen scale, that was $30,000 last month against $7,500 of ad sales.

ROAS = 1 ÷ ACoS, and ACoS = 1 ÷ ROAS (with ACoS as a decimal)

The scale's 20% ACoS is a ROAS of 1 ÷ 0.20 = 5, and a ROAS of 4 would be an ACoS of 1 ÷ 4 = 25%. TACoS is the one that can drift on its own, because it's ACoS times the share of sales that came through ads (20% × $7,500 ÷ $30,000 = 5%). If the scale's ACoS held at 20% while its TACoS climbed, more of its sales would be arriving through ads and fewer on their own, which is the drift our guide to Amazon TACoS shows you how to spot across a whole catalog.

A good ACoS on Amazon is any ACoS under your break-even ACoS, the point where ad spend eats your whole margin, as long as it still brings in enough sales to be worth running. Amazon's ACOS guide won't name a figure. It says there isn't a definitive number, only that to keep a profit your ACoS "needs to be lower than your profit margin."

What about the benchmark figures that get quoted? SellerStack, which sells PPC software, reported a median ACoS of 31.5% across 69 US accounts that each spent $3,000 or more on ads from June to August 2026. The middle half ran between 23.8% and 44.4%. Those are one tool's customers, and 30 of the accounts belong to an agency group linked to its parent company. Run the scale at that median and it keeps $1.40 on each $40 ad sale. A product with a 25% margin, running the same 31.5%, loses $2.60 on every one.

How to calculate break-even ACoS and target ACoS

Break-even ACoS is the ACoS at which ad spend uses up your whole margin before ad costs, so each ad sale makes nothing and loses nothing.

Break-even ACoS = margin before ad costs ÷ price × 100

We use the margin before ad costs because ad spend is the cost ACoS measures. For the scale that's $14 ÷ $40 × 100 = 35%, the point where each ad sale spends $14 on ads to earn $14 of margin.

Your target ACoS sits somewhere under that line, and you pick where. For the scale, that's closer to 20% than to 15%, because the 20% Moderate setting earns the most from ad sales in the settings table below. A launch can run above 35% on purpose, for a set period.

Is a lower ACoS always better?

No, a lower ACoS isn't always better. ACoS is a ratio, and the easiest way to lower it is to spend less, which also shrinks the ad sales your margin is earned on. How far under break-even to aim depends on what each extra dollar of spend brings back, and that's easier to see in dollars.

Profit from ad sales = ad sales × margin before ad costs − ad spend

Now run it for the scale at four levels of monthly spend. We made these numbers up too, assuming each extra dollar of spend buys less than the dollar before it.

Tell an automation to "get ACoS down" and nothing else, and it'll walk the scale from Moderate to Tight. ACoS improves by 5 points, and you're $525 a month worse off.

Push the other way and the mistake is harder to spot. Wide still reports 27%, eight points under break-even, so nothing on the dashboard looks wrong. Yet the extra $1,500 it spends over Moderate brings only $3,500 more in sales. That's a 43% ACoS on the extra money, above the scale's 35% line, and it leaves Wide $275 behind Moderate.

Which step was worth paying for? From Tight to Moderate, the extra $1,050 brought $4,500 of sales, a 23% ACoS on the new money and comfortably under 35%. From Moderate to Wide it was 43%. Add spend only while the extra spend's own ACoS stays under break-even, which for the scale means stopping at Moderate.

None of those four rows counts organic sales. Across the whole product last month, $30,000 of sales at a 35% margin came to $10,500, and $9,000 once the $1,500 of ads came out. Call that profit after ads. If the ads lift organic sales too, the lift lands in that $9,000 and never in the table.

What is break-even ROAS?

Break-even ROAS is the lowest ROAS at which your ads don't lose money: 1 divided by your break-even ACoS.

Break-even ROAS = 1 ÷ break-even ACoS = price ÷ margin before ad costs

For the $40 kitchen scale that's 1 ÷ 0.35, or $40 ÷ $14, about 2.86. You'll need that number the moment you touch Amazon's own automation, because rule-based bidding for Sponsored Products chases a ROAS goal you set. To get there it can raise a bid to as much as 5 times your adjusted bid, and Amazon says it doesn't guarantee you'll hit the target.

Budget rules can add budget when a metric such as ROAS clears a level you pick, and our guide to AI PPC management on Amazon explains how rule-based bidding sets its bids. Put four ROAS targets through the scale's numbers and this is what each $40 ad sale keeps.

A ROAS target of 3 sounds healthy, yet it leaves the scale just 67 cents on a $40 ad sale. Leave the target blank and Amazon says rule-based bidding "will automatically determine an appropriate ROAS target to maximize sales." For the scale, chasing maximum sales can mean the Wide and Aggressive rows, where the extra sales are bought above break-even. Aggressive loses $100 a month, so leave the target blank only when that's a trade you've chosen.

A good ACoS can still lose money when the ACoS your bid rules read is wrong, or when ads stop showing without ACoS changing. The number reads wrong when recent days are still filling in, when attributed sales ignore cart discounts, or when other products' sales get credited to the ad. The ads stop altogether when you lose the Buy Box or run low on stock, and then nothing reaches the ACoS report at all.

The last 7 days of ACoS read too high

Yesterday's ACoS reads high because yesterday's sales aren't all in yet. For sellers' Sponsored Products ads, Amazon credits a sale to a click for 7 days after it and reports the sale on the day of the click, and a conversion can take up to 12 hours to show. Brand campaigns such as Sponsored Brands generally look back 14 days.

Say Sunday's $100 of clicks eventually brings $400 of sales. If only $250 has landed by Monday morning, Sunday reads 40%, and a rule that cuts the scale's bids above its 35% break-even fires on Monday, against a day that reads 25% a week later. Let your rules act only on days whose window has closed. Before you set a ROAS target at all, Amazon's rule-based bidding help recommends at least two weeks of data.

Cart discounts aren't in ACoS's sales figure

Amazon's attribution help page says attributed sales "don't reflect any discounts applied in the cart." Put $6 off the $40 kitchen scale and you collect $34, the referral fee drops to $5.10, and $8.90 is left before ads (before any coupon fee). The report still divides by $40, so during the promotion your break-even on the ACoS you see is about 22%.

A promotion week reporting 25% looks comfortable and loses $1.10 on every ad sale, because each one costs $10 in ads against $8.90 of margin. Check promotion weeks by hand against that 22% line instead of the usual 35%.

Sales of your other products count in ACoS

Say a shopper clicks the scale ad and buys your $12 measuring cups instead. The ad still gets the sale, because under Amazon's attribution rules a seller's ad gets credit for the advertised product "as well as any other products sold by the Seller Central merchant." That $12 lands in the scale's ad sales with the cups' margin attached, and the scale's 35% break-even only fits one of the two products. Before you judge that ACoS against 35%, look at which products the ad's sales came from.

Losing the Buy Box stops ads without raising ACoS

Losing the Buy Box stops your Sponsored Products and Sponsored Display ads, and your ACoS doesn't rise to warn you. Those ads only run while your product holds the Featured Offer, what most sellers still call the Buy Box. When another seller takes it, Amazon says the campaigns "generate 0 impressions" even while they show "Delivering". At last month's pace that's up to $250 a day of the scale's ad sales that never happen. Sponsored Brands keeps serving as long as the product is still available to buy.

Low stock can hit the scale's ads twice. Sponsored Products only appear when the item is in stock, and before you get that far, Amazon's Featured Offer help says it may withhold the Featured Offer for insufficient inventory, even when you're the only seller. Your best seller can lose its ads before it sells out.

Account health can do the same. That Featured Offer page still says your order defect rate must stay below 1% to keep Featured Offer eligibility, though in July 2026 Amazon told sellers it's removing that separate eligibility check by the end of the year, with seller performance counting toward which offer gets featured instead. Either way, a bad stretch of negative feedback, A-to-z claims, or chargebacks can cost you the Featured Offer, and the ads come off the page while whoever runs advertising is still looking at a normal ACoS.

Competition can move the Featured Offer too. When a more competitive offer takes it, Amazon says impressions for the previous holder "stop immediately". A price that fails Amazon's low-price standards can cost you the Featured Offer even when nobody else is selling.

Next to ACoS, then, track each ASIN's impressions against the day before, its Featured Offer share, and its days of stock cover. Our guide to getting alerted when you lose the Buy Box sets up the Featured Offer alert. The scale sells about 25 units a day, so at 225 units it has nine days of cover left. At that point even a healthy 20% ACoS is a reason to ease bids and hold off on raising them.

To reduce ACoS on Amazon without losing profit, cut or negate the search terms and targets whose own ACoS is above break-even, and leave the profitable ones running. A campaign average hides both. Inside the scale's 20% campaign, a search term running at 50% loses $6 on every sale it brings. Another at 25% keeps $4 a sale, even though it pulls the campaign average up, and a rule aimed only at the average could cut it.

1. Negate search terms that run above break-even once they've had enough clicks to judge, counting only days whose attribution window has closed. If about one click in ten buys the scale, a term with ten clicks and no sale hasn't proved much yet.
2. If a target sits between your target ACoS and break-even, lower its bid. It's still making money, just less than you want, and a negative would throw that away.
3. Raise conversion, starting with the listing checks in our guide to Amazon PPC clicks but no sales. ACoS works out to cost per click divided by sales per click, so when a better main image or a clearer title gets more of the same clicks to buy, ACoS falls while sales go up. Had the scale's $1,500 of spend brought $10,000 of sales instead of $7,500, its ACoS would have been 15%, the same as Tight. Profit from ad sales would have been $2,000, against Tight's $600.

How to judge Amazon PPC automation on profit, not just ACoS

Judge Amazon PPC automation on profit after ads per ASIN, with ACoS as one input. It matters less whether that's Amazon's own rules, Amazon PPC tools, a freelancer, or an agency. For profit, the better choice is whichever one can answer the six questions below about your bids with something you can check.

You can answer the first question inside Seller Central. Its Profit Analytics can hold each product's cost, which for the scale is the $12 landed cost behind its 35% break-even. To keep ACoS and TACoS beside the numbers that drive them, use our Amazon seller metrics scorecard.

Where Lumian fits

Who fixes the cause outside the ads is the one question software can't answer on its own. Lumian's managed service runs Amazon advertising together with inventory, listings, and account health, with a Brand Manager, specialists, and AI agents working across the account.

The published STREGA case-study result card reports ACoS falling from 69% to 17% and TACoS from 21% to 7%. That came from a campaign rebuild. Branded searches were split from non-branded ones so each had its own budget, spend was focused and scheduled by time of day, and a strategist approved changes before they went live.

Judged on profit after ads, that's half the story. Both figures measure ad efficiency, and the card doesn't give a date range. Of the two, TACoS is closer to what we've asked you to watch. To have a team run your ads this way, see Lumian's advertising service.

If you'd rather do the checking yourself, Lumian Agents can keep an eye on the Featured Offer, stock cover, ASIN margins, and ACoS for Amazon.com (US) accounts. Buy Box Tracker reports Featured Offer ownership by product and checks the offers after a drop. Daily Restock Alert flags SKUs under 30 days of supply and estimates how much to reorder, so at 25 units a day it would flag the scale about three weeks before it got down to nine days of cover.

On the margin side, Profitability & Fee Analyzer works out ASIN margins after fees, refunds, and ad spend. Its page says product cost isn't included, so subtract yours before you call the result profit; leave out the scale's $12 landed cost and its margin reads $12 a unit too high. The fourth, Daily ACoS Watchdog, flags Sponsored Products and Sponsored Brands campaigns over your target and recommends a response. Read its flags on the last few days with the attribution lag in mind. None of the four is described as changing bids, budgets, prices, or stock, so what changes in your account stays your call.

Before you change a bid, run the scale's sums on your own best seller this week. Price minus fees, landed cost, and allowances is your margin before ads, and that margin as a share of price is your break-even ACoS. For the scale, that was $14 and 35%. Last month's total sales times that margin, minus ad spend, is your profit after ads, which came to $9,000 for the scale. If an ASIN is running within a few points of break-even, or has less than a month of stock, look there first, and you can Start Free to point the checks at it.