Key takeaways: Most high-ACoS problems are not bidding problems. Check conversion rate first, wasted search terms second, placements third and bids last. Cutting bids before fixing the listing lowers spend and sales together.
Start with the question ACoS cannot answer
ACoS tells you spend divided by ad sales. It does not tell you why. The same 45% ACoS can come from a listing that does not convert, a campaign spending on irrelevant searches, a bid war on top-of-search, or simply a product whose margin cannot support advertising at all. Each has a different fix, and the wrong fix makes things worse.
The order we check
Read this flowchart as text
- ACoS above target for 2+ weeks
- Conversion rate vs category norm?
- Listing problem: images, price, reviews, offer
- Search-term report: spend with zero sales?
- Negatives + match-type tightening
- Placement report: top-of-search ACoS vs rest?
- Cut placement modifiers, keep base bids
- Bid-level optimisation by keyword
1. Conversion rate
Pull the unit session percentage for the ASIN and compare it with the category. If the listing converts below the norm, ads are paying for clicks the page cannot close. Lowering bids here reduces spend but keeps the ratio roughly the same. Fix the main image, price position, review count or the offer before touching bids.
2. Wasted search terms
Open the search-term report for the last 60 days and sort by spend with zero orders. If more than a fifth of spend sits there, the campaign is buying irrelevant traffic. Add exact and phrase negatives, and move broad-match keywords that leak into a tighter campaign.
3. Placements
Top-of-search converts best but can cost two to three times more per click. If the placement report shows top-of-search ACoS far above the rest, the placement modifier is too aggressive. Reduce the modifier before reducing base bids, so you keep the cheaper placements that are working.
4. Bids
Only now do bids come into it. Adjust at keyword level, on evidence from at least two weeks of data, and in steps rather than cuts. A 10 to 15% change with a week to read the result beats a 40% cut that loses the placement entirely.
The fix that is not a fix
Pausing the campaign. Paused campaigns lose rank on their keywords within days, and the organic sales that came with that rank go too. If the product cannot support advertising at any ACoS, that is a margin or pricing decision, not a PPC one.
Frequently asked questions
What is a good ACoS on Amazon?
The one below your break-even ACoS, which is your margin before advertising. A 30% ACoS is strong for a product with 45% margin and a loss for a product with 25% margin. Launch campaigns run higher on purpose.
How fast should ACoS come down after changes?
Wasted-spend fixes show within one to two weeks. Listing fixes take longer because conversion rate needs traffic to prove itself, typically three to four weeks. Judge changes on two weeks of data minimum.
Should I lower bids across the whole account?
No. Account-wide cuts lower spend and sales together and lose placements that were working. Change bids at keyword level, based on that keyword's conversion and ACoS.