A high ACOS is only a problem if it is above your book's break-even ACOS after you count page reads and series sales. First check that number; then cut targets that spend without orders and lower bids on targets that sell above break-even, in that order.
Last reviewed by the InteliAds team.
First: is your ACOS actually too high?
ACOS is ad spend divided by ad-attributed sales. Amazon calls it advertising cost of sales. Whether a given ACOS is good depends on how much of the price you keep. A 70% ebook royalty can survive an ACOS near 65%; a colour paperback might break even at 25%.
Two things Amazon Ads does not count make ACOS look worse than reality: Kindle Unlimited page reads, and later books in a series that readers buy without another ad click. If your book is in KU or is book one of a series, put those back before you decide.
Work through these in order. Each one points at a different fix.
Break-even: what is your break-even ACOS for this book and format?
KU and series: what do page reads and later books add per ad sale?
Concentration: do 10 to 20% of targets cause most of the spend without orders?
Match type: are broad or auto targets matching loosely related searches?
Bids: are winning targets bid well above what they need to keep their position?
Placements: is a top-of-search adjustment raising costs without better conversion?
Product page: is conversion low for every target? Then the problem may be the page, not the ads.
Fixes, in the order that saves most
Cut pure waste first. Targets and search terms with plenty of clicks and no orders are 100% loss. Pause the keywords and product targets, and add the search terms as negatives in the campaign they came from.
Then lower bids on targets that sell but above break-even. If a keyword runs at 90% ACOS against a 65% break-even, lowering its bid by roughly the same proportion (about a quarter) usually brings it close, at the cost of some clicks. Do it in steps of 10 to 20% and wait for data between steps.
Then look at placements. If you added a high top-of-search adjustment, check whether top-of-search orders are worth the extra cost per click. Finally, look at the product page if conversion is weak everywhere: cover, description, price, reviews and Look Inside.
Where InteliAds automates the fix
InteliAds Rules are IF/THEN rules that run every hour. Conditions can use impressions, clicks, CTR, CPC, spend, orders, sales, ACOS, ROAS and conversion rate over a date range. Actions include pause, enable, set, raise or lower a bid, set a daily budget, change a placement adjustment, add a search term as a negative, or add it as a new target. Dry Run shows what a rule would change before it touches Amazon, and every change can be reverted.
Two rules cover most high-ACOS cases: pause any keyword with 15 or more clicks and 0 orders in the last 30 days, and lower the bid by 15% on any keyword with ACOS above your break-even and at least 2 orders. Run both through Dry Run first.
BidBot is the automated bidding in InteliAds. You pick a preset (Suggest only, Safe, Balanced, Aggressive or Aggressive+) and a target ACOS between 10% and 80%. Safe waits for 20 clicks before it raises a bid and leaves 48 hours between changes to the same target; Balanced waits for 10 clicks and 24 hours. You can also set it to lower bids only, and mark promo windows when it should not raise bids.
Set BidBot's target ACOS to your break-even (adjusted for KU and series if you have the data), and it will move bids down on targets above it and up on targets well below it, within its cooldown and click limits.
Calculate break-even ACOS for each book and format.
Add KU page-read earnings and series read-through if they apply.
Sort targets by spend over the last 30 days and mark those with no orders.
Pause those targets and negate their search terms.
Lower bids 10 to 20% on targets selling above break-even.
Check placement adjustments; reduce any that cost more without converting better.
Wait one to two weeks, then repeat or hand the routine to Rules or BidBot.
Worked example (illustrative numbers)
Invented numbers. An ebook at $4.99 has a $3.40 royalty: break-even ACOS 68%. Last 30 days: $200 spend, $220 sales, ACOS 91%.
Sorted by spend, 6 of 48 keywords spent $70 with no orders. Pausing them would have left $130 spend for the same $220 sales: ACOS 59%, below break-even. Three other keywords sold at 85% to 95% ACOS; the author lowers their bids by 15%. No working keyword was paused.
Frequently asked questions
What is a good ACOS for KDP ads?
Any ACOS below your book's break-even ACOS. For many 70% ebooks that is around 60% to 68%; for print books it is usually lower.
Should I pause a campaign with high ACOS?
Usually not the whole campaign. Pause the targets with clicks and no orders, and lower bids on the ones selling above break-even.
Does KU affect ACOS?
Yes. Page reads are not counted as ad sales, so books in KU often look worse in ACOS than they really are.
How fast can I lower ACOS?
Cutting waste works within days. Bid changes need one to two weeks of data to judge.
Amazon and KDP rules change; check the linked help pages before making decisions. Worked examples use illustrative numbers, not results from real accounts.