Advertising cost of sales, almost always shortened to ACoS, is one of the first numbers people look at in an Amazon advertising report. It answers a narrow question: for every unit of currency that the ads were credited with selling, how much was spent on those ads? Because it is a ratio, it puts a small campaign and a large one on the same footing. Because it is narrow, it is also easy to read too much into. This guide covers the calculation, what a “good” value depends on, and where the number stops being useful.
The formula
ACoS = ad spend ÷ attributed ad sales × 100%
Both inputs come from advertising reports. Ad spend is the total cost of the clicks or other billable ad events in the period. Attributed ad sales are the sales that the advertising report credits to those ads within its attribution window. Neither input comes from the seller sales report, and that distinction matters later.
A worked example
Suppose a fictional campaign spent 25 in one week, and the ad report credits it with 100 in attributed sales. Then 25 ÷ 100 = 0.25, or 25%. In plain words, a quarter of the attributed sales value went back out as ad cost.
The same arithmetic applied to a few illustrative campaigns:
| Campaign (fictional) | Ad spend | Attributed sales | ACoS |
|---|---|---|---|
| Campaign A | 25.00 | 100.00 | 25.0% |
| Campaign B | 60.00 | 150.00 | 40.0% |
| Campaign C | 18.00 | 120.00 | 15.0% |
| Campaign D | 12.00 | 0.00 | N/A |
| Total | 115.00 | 370.00 | 31.1% |
Campaign D shows why the denominator matters. With zero attributed sales, the division is undefined. Showing 0% would suggest perfect efficiency, which is the opposite of what happened. Showing N/A next to the raw spend tells the true story: money went out and no sales were credited in this window.
The total row is calculated from total spend and total attributed sales (115 ÷ 370), not by averaging the campaign percentages. A campaign that spent 12 should not carry the same weight as one that spent 60.
What makes an ACoS “good”?
There is no universal target. Whether 25% is healthy depends on what the product earns before advertising and on what the advertising is meant to achieve.
Margin sets the reference point
A useful anchor is break-even ACoS: the ACoS at which ad cost would use up all of the money left from a sale after per-unit costs.
break-even ACoS = contribution margin before ad cost ÷ selling price × 100%
Contribution margin before ad cost is the selling price minus every cost that scales with each unit sold, such as product cost, referral fee, fulfillment fee and an allowance for refunds, but before any advertising.
| Per-unit item (illustrative) | Amount |
|---|---|
| Selling price | 30.00 |
| Refund allowance (2%) | −0.60 |
| Referral fee (15%) | −4.50 |
| Product and inbound cost | −8.40 |
| Fulfillment fee | −6.00 |
| Contribution before ad cost | 10.50 |
| Break-even ACoS | 10.50 ÷ 30.00 = 35% |
With a 35% break-even, a 25% ACoS leaves roughly 10% of the attributed sales value as contribution after ad cost, assuming the attributed sales are of this product at this price. Above 35%, each attributed sale loses money at the contribution level. The fee rates in the table are fictional; real fees vary by category, size and marketplace.
Goals change the target
Break-even is a reference, not a goal. A seller launching a new product may accept an ACoS above break-even for a limited time to learn which search terms convert, which is a deliberate investment with an end date. A mature product with steady organic demand might be run well below break-even to protect profit. Both choices are reasonable when the goal is written down and reviewed on a schedule; neither is reasonable when it happens by accident.
Cautions
Attribution windows and late sales
Attributed sales are credited within a defined window after an ad interaction, and windows can differ between ad products and reports. Sales also tend to arrive in reports days after the click. The most recent days are therefore provisional: yesterday’s ACoS often looks worse than it will once late attributions land. Compare periods that have had the same amount of time to mature, and compare reports that use the same window.
Small numbers swing hard
Zero attributed sales gives N/A. Very small sales give unstable values: one extra order can move a campaign from 80% to 30%. Look at the raw spend, clicks and orders before reacting to a ratio built on a handful of events.
It is not store profitability
ACoS describes advertising efficiency on attributed sales. It leaves out sales that were not attributed to ads, storage costs, overhead and returns processing, and it says nothing about whether the business as a whole is profitable. A low ACoS on a thin-margin product can still lose money, while a higher ACoS on a product that earns repeat orders may be worth it.
A different denominator is a different metric
Dividing ad spend by total seller sales produces another ratio, often called TACoS. It answers a different question, draws on a different data source and should never be labeled as ACoS. The guide on seller sales versus attributed sales explains why.
Common mistakes
- Averaging daily or campaign-level ACoS percentages instead of dividing total spend by total attributed sales.
- Judging the latest few days before late attributions have arrived.
- Reporting 0% when the honest answer is N/A.
- Using one company-wide ACoS target for products with very different margins.
- Comparing ACoS across reports that use different attribution windows or currencies.
Try it
Enter your own spend and sales in the Ads Metrics Calculator, then work out a break-even figure in the Break-Even ACoS Explorer, both at the free tools. To see ACoS on synthetic campaigns, open the advertising workspace of the demo and use “Explain metric” on any value. For the inverse view of the same numbers, read What Is ROAS? and ACoS vs ROAS. Short definitions live in the glossary.