Amazon ACOS, judged against your own margin

What is ACOS on Amazon?

ACOS is ad spend as a share of the sales Amazon credits to your ads. It stands for advertising cost of sales, and it measures how much each dollar of ad sales cost you, not whether those sales made a profit (Amazon Ads). The other metrics are in what each Amazon ads term means.

The sales in ACOS are attributed sales: orders Amazon credits to an ad because they came within its attribution window after a click. Sales outside that window, and organic sales, are not in it.

How do I calculate ACOS?

Divide ad spend by ad sales and multiply by 100. $50 of ad spend that brings $100 of ad sales is a 50% ACOS (Amazon Ads).

Calculate it over the same window for spend and sales, per product or per segment rather than for the whole account, and over a period that ended at least an attribution window ago (below).

What is a good ACOS on Amazon?

A good ACOS sits below your break-even ACOS, your contribution margin before ads as a share of price, by the profit you want to keep per order. Amazon says the same: there is no definitive good ACOS, and to make a profit your ACOS has to sit below your profit margin (Amazon Ads).

For reference, here is how far ACOS ran from store to store in the accounts we run. The middle half of stores sits between the 25th and 75th percentile columns.

• Branded ran lowest everywhere. Branded ACOS was below non-branded in all 20 stores; at the median store it was 0.38 times the non-branded figure.
• Non-branded varies most. A quarter of stores ran non-branded searches above 84% ACOS. Those searches buy new customers, so whether that pays depends on margin and on repeat orders you can measure, not on the ACOS alone.
• The blended figure depends on the mix. All searches together ran at 37.0% at the median store, between the two segments, and where a store lands moves with how much of its spend goes to its own brand name.

Use the table to see where you stand, not as a target. A 47.7% non-branded ACOS loses money on a product with a 30% margin and makes money on one with 60%.

What is break-even ACOS?

Break-even ACOS is the ACOS at which an ad-driven order makes no profit on its first purchase: your contribution margin before ads as a share of price. A product that keeps $12 of a $40 price, after cost of goods, Amazon’s fees and returns, breaks even at a 30% ACOS. Our bidding guide works through how to calculate break-even ACOS and the bid it implies.

Should my target ACOS equal my break-even ACOS?

No. At break-even the ads earn nothing on the first order, so set the target below it by the profit you want to keep per order: in the example above, keeping $3 of the $12 leaves $9 to spend per $40 order, a 22.5% target ACOS.

• Set it per product. One target across products with different margins overspends on the low-margin ones and underspends on the high-margin ones; profit per click is the better number to bid from, with ACOS read as the result (is a target ACOS the best strategy?).
• Above break-even only on purpose. Non-branded clicks that win new customers can justify a target above break-even, but only as far as you can measure their repeat orders.
• Expect new campaigns to run high. Amazon notes that new campaigns tend to have a high ACOS simply because they are new (Amazon Ads); judge them on clicks, not the first week’s ACOS.

Why is branded ACOS so much lower than non-branded?

Shoppers who search your brand already want your product, so their clicks convert more and cost less. Comparing within each store, the median store’s branded clicks converted 1.9 times as often as its non-branded ones and cost 23% less, and branded ACOS came out lower in all 20 stores.

That gap makes a blended ACOS easy to improve without winning a single new customer: move budget to your own brand name and the account figure falls (how blended ACOS drifts away from new sales). Report and target the two segments separately.

ACOS vs ROAS: which should I use?

Either: they are the same number inverted. ROAS, return on ad spend, is ad sales divided by ad spend, so a 25% ACOS is a ROAS of 4. Use the one your reports and team already use, and set it against margin either way.

What is TACOS, and what is a good TACOS?

TACOS, total advertising cost of sales, is ad spend divided by all your sales, ads and organic together. A good TACOS is one your margin can carry while total sales grow: ACOS tells you what the ads cost per ad sale, TACOS tells you what advertising costs the whole store. To turn a TACOS target into a monthly budget, see how to set an Amazon ads budget from a TACOS target.

In the 14 of our 20 stores with total sales for every day of the window, Sponsored Products spend came to a median 13.9% of total sales, with the middle half between 11.2% and 16.1%. Sponsored Products ads were credited with a median 48.7% of those stores’ total sales (middle half 25.4% to 55.2%). Sponsored Brands and Display spend is not in these figures, so a TACOS across all ad types runs higher.

• Watch the two together. If organic sales grow faster than ad spend, TACOS falls even while ACOS holds steady. If TACOS climbs while total sales stay flat, the extra spend is not adding sales.
• A worked case. A store at a 30% ACOS whose ads bring 40% of its sales runs at a 12% TACOS. The numbers are illustrative.

Why is my Amazon ACOS so high?

Usually because spend has moved to clicks that convert less, or click prices rose faster than conversion (what sets the price of a click). Check these in order:

• The mix. More spend on non-branded, competitor or new searches raises ACOS even when every click is fairly priced. Split branded from non-branded before reading the trend.
• Placements. Non-branded product-page clicks ran at a 64.9% ACOS at the median store, against 42.1% at the top of search (ACOS by placement).
• Bids above what clicks are worth. Placement adjustments and “up and down” dynamic bidding can push the price you pay well above your keyword bid (why your CPC is higher than your bid).
• Irrelevant searches. Broad match and automatic targeting can match searches that never convert for your product.
• The listing. Price, reviews and images change CVR for every click at once; a lower CVR at the same CPC is a higher ACOS.
• The season. Click prices rise in the weeks before Black Friday faster than shoppers buy (Amazon ad costs in Q4).
• Recent days. Orders keep arriving for days after a click, so the last week’s ACOS reads high until they are counted (below).

How do I lower ACOS without losing sales?

Cut spend only where the last dollar loses money, not across the board. A lower ACOS bought by dropping profitable clicks is a smaller business, not a better one.

• Lower bids where clicks cost more than they are worth: search terms and placements whose ACOS sits above break-even once they have enough clicks to judge (how many clicks before pausing a keyword).
• Add negative keywords for searches that do not fit your product, and for those that spent through enough clicks with no order (when to add negative keywords).
• Bid each placement on its own results, and use “down only” dynamic bidding unless your placement data shows the extra is worth paying (up and down or down only).
• Raise CVR, not just bids: a better price, images or reviews lower ACOS on every click without giving up any.
• Do not move budget to your brand name to lower the figure. The account ACOS falls, but new-customer sales do not grow.

Is a lower ACOS always better?

No. Below break-even, a higher ACOS can earn more profit, because the extra spend buys orders that still clear your margin. Judge the last dollar of spend, not the average: the worked example below shows a product earning the most at a 20% ACOS, not at its lowest.

Does ACOS differ by placement?

Yes, a lot. At the median store we run, non-branded clicks ran at a 42.1% ACOS at the top of search, 44.7% in the rest of search and 64.9% on product pages, and product pages ran a higher ACOS than the top of search in 17 of 20 stores. Our bidding guide has the full placement table and the bid adjustments it implies.

How does the attribution window change ACOS?

ACOS counts only the sales inside the window. Amazon credits a seller’s Sponsored Products ad with orders placed within 7 days of the click, and 14 days for vendors (Amazon Ads). Three things follow:

• Recent days read high. Orders from the last few days of clicks have not all arrived, so judge ACOS on periods that ended at least a window ago.
• Compare like with like. A longer window counts more orders, so the same campaign never shows a higher ACOS over 14 days than over 7. Check the window before comparing two reports; Sponsored Brands and display ads count longer still (why ad types report different results).
• Repeat orders never show. A customer’s second order weeks later is outside every window, so ACOS understates what new customers are worth.

A worked example

One product, one month, three ways of bidding it. The numbers are hypothetical. The product sells for $40 and leaves $12 per order after cost of goods, Amazon’s fees and returns, so it breaks even at a 30% ACOS.

1. 01
   The lowest ACOS earns lessMoving from low to middle bids adds 50 orders for $500, $10 each, under the $12 each order leaves. ACOS rises from 15% to 20% and profit rises by $100.
2. 02
   Past break-even, profit fallsMoving from middle to high bids adds 40 orders for $1,000, $25 each, more than they bring. That last $1,000 ran at a 62.5% ACOS on its own sales, though the month’s average is only 32.1%.
3. 03
   Read the change, not the averageThe average ACOS hides where the last dollar went. Compare the extra spend with the extra sales it brought, and stop where that ratio reaches your break-even.

How Laurence uses ACOS and TACOS

Laurence prices each click from your margins rather than a target ACOS. Each bid starts from what an order leaves after Amazon’s fees and the cost of goods you enter, times the CVR it expects, and Laurence bids to the goal you choose: net profit, revenue growth, or a target CAC or CPA. More on why Laurence prices clicks from your margins.

• ACOS by segment, as a result: once your brand terms are set, branded, competitor and non-branded searches run in separate campaigns, so each one’s results are read on their own and branded spend cannot flatter the rest.
• TACOS against your ceiling: you can give each product a maximum TACOS in Settings. The platform’s Overview charts your TACOS against it, and your budgets can be sized from it.

How we measured this. Sponsored Products search term reports (Amazon Ads API) for 20 US stores on the Laurence platform, August 6 to November 3, 2026, at Amazon’s default attribution window. Searches for a store’s own brand or its own products are branded; every other search, competitors’ brands included, is non-branded. Each store’s ACOS is computed first, then summarized as percentiles across stores; no store is named. TACOS uses each store’s total ordered sales from Amazon’s Selling Partner API, for the 14 stores with a figure for every day of the window.