A lower ACOS can look like progress while the business earns less. Before cutting bids or increasing the budget, check what changed in the product’s economics and the traffic reaching the listing.
Start with the calculation
ACOS is advertising spend divided by advertising-attributed sales, multiplied by 100. ROAS presents the same relationship the other way around: attributed sales divided by advertising spend. Amazon’s own guide also warns against treating ACOS as the only campaign metric.
For an illustrative example, $300 in advertising spend and $1,200 in attributed sales produces a 25% ACOS and a ROAS of 4. Those numbers describe the advertising relationship. They do not tell you how much profit remains.
Build a product-level cost picture
Suppose a product sells for $40. After product cost, marketplace fees, fulfilment, and expected returns, assume $12 remains before advertising. An average advertising cost of $10 per order would leave $2 before any fixed overhead or other costs not included in that calculation. If your assumptions change, the amount you can afford to spend changes too. This is a worked example, not a ScaleBSR client result.
Separate a traffic problem from a conversion problem
Begin with a consistent reporting period. Compare spend, clicks, attributed orders, and average selling price with the previous comparable period. Then ask which part moved.
If clicks became more expensive while conversion stayed similar, examine targeting and bidding. If click costs stayed similar but fewer shoppers purchased, examine the listing and offer. A stock issue, a price change, unclear product information, or a stronger competing offer can matter more than a bid adjustment.
Review the search terms before making a broad cut
A campaign average can conceal very different buying intentions. Review the searches and products that spent money, the orders attributed to them, and whether they fit what you sell. Set an explicit decision rule for the review period and the amount of evidence you need. A small sample does not justify a confident conclusion.
Choose the next test
Write down one diagnosis and one action. For example: a group of searches is attracting shoppers looking for a different product size. The proposed test could be to tighten relevance while checking whether useful sales are lost. Record the date and the intended outcome. Review the result after an appropriate period without changing several unrelated variables at once.
Use ACOS alongside total sales, product margin, availability, and the campaign’s purpose. A target chosen without those details is a number without a business case.
Source and next step
Amazon Ads: ACOS calculation and interpretation. Reviewed 14 October 2026. The diagnostic sequence and worked example above are ScaleBSR’s editorial explanation.