Every Amazon ad account keeps a record of exactly what shoppers typed before they clicked your ads. It is called the search term report, and it is the closest thing Amazon advertising has to a list of where your money went. Most sellers know it exists. Far fewer open it every week, because the file is long, the columns are dense and it is not obvious where to start.
You do not need hours with it. With a fixed routine and two numbers decided in advance, fifteen minutes a week is enough to stop the worst waste and feed your best searches more budget.
The short answer
Once a week, pull a Sponsored Products search term report for the last 14 to 30 complete days, leaving out the most recent few days. Decide your break-even ACoS and target ACoS before you look. Then work through the file in four passes: bleeders (enough clicks, no sales), irrelevant terms (wrong product or audience), winners (sales below target ACoS) and product targets (ASINs in the search term column). Negate, lower bids or move terms to exact match, write down what you changed, and give each change time to work before you judge it.
What the search term report shows
Amazon describes the report as showing the searches shoppers entered on Amazon that led to at least one click on your ad. Searches that only produced impressions are not in it, which keeps the file focused on terms that actually cost you money.
Three columns matter more than the rest:
- Targeting is the keyword or product target you chose (or the automatic targeting group).
- Match type tells you how that target matched: broad, phrase, exact, or automatic.
- Customer search term is what the shopper actually typed. The gap between this column and the targeting column is where most of the insight sits.
Alongside these you get impressions, clicks, cost per click, spend, purchases and sales, from which you can work out conversion rate and ACoS. When a product target or an automatic campaign generates a click, the search term column shows an ASIN instead of words. Amazon notes that an ASIN is always ten characters long and is either all numbers or starts with "B".
Sponsored Brands campaigns have a search term report too, and Amazon offers a separate search term impression share report for Sponsored Products. This routine uses the Sponsored Products report, because that is usually where most spend sits.
Getting the report, and the reporting change to know about
Amazon is moving its ad reporting into one place. In June 2026 it announced that unified reporting is generally available in the advertising console under Measurement and Reporting, and that the older Sponsored Ads reports page will be sunset by 31 December 2026. If you still rely on the old page or on scheduled reports set up there, rebuild them in unified reporting and download any history you want to keep before the cut-off. Unified reporting keeps up to 15 months of daily or weekly data, according to Amazon.
When you build the report, include campaign, ad group, targeting, match type and search term, plus impressions, clicks, spend, purchases and sales. Choose a summary view for the weekly review, so each search term appears once per target, and schedule it weekly so it is waiting for you. Menus and metric names have been changing, so confirm the exact steps in your own account.
Leave out the last few days
Sales are credited to an ad for a period after the click, not just on the day. For Sponsored Products the standard window is commonly cited as 7 days for sellers and 14 days for vendors. That means the most recent days in any report will look worse than they really are, because some purchases have not been credited yet. Ending your date range a few days back, or at least a full attribution window back for low-volume products, stops you cutting terms that were about to convert.
Before you open the file: two numbers
A report without targets invites guesswork. Decide these first, per product or product group.
- Break-even ACoS. Amazon's own guidance is that your ACoS must stay below your profit margin to remain profitable. Take the price, subtract product cost, Amazon fees and shipping, and divide what is left by the price. Illustratively, a product selling at $40 that leaves $12 before advertising has a 30% margin, so a 30% ACoS breaks even.
- Target ACoS. Below break-even, set by what you want the ads to achieve. A launch might accept close to break-even; a mature product might aim for 20%, which on the $40 example leaves $4 of profit per ad-driven sale.
The 15-minute routine
Minutes 1–5: bleeders
Sort by spend, highest first. Look for terms with real clicks and no purchases. Amazon's targeting guide suggests evaluating a keyword after it has received at least 20 clicks before adding it as a negative. For a relevant term past that point, either lower the bid or add it as a negative. For a term that is close to your product but not quite right, lower the bid first; negate when it keeps spending without converting.
Minutes 5–8: irrelevant terms
Scan the remaining terms for searches that were never going to buy: the wrong product, the wrong size or use, an audience you do not serve. These do not need 20 clicks. Add them as negatives straight away. Amazon offers negative exact and negative phrase, at campaign or ad group level. Be careful with negative phrase on single words: it blocks every search containing that word, which can quietly switch off good traffic.
Minutes 8–12: winners
Now filter for terms with purchases and an ACoS at or below target, especially those found by automatic, broad or phrase targeting. Add each one as an exact match keyword in a manual campaign with its own bid. Then add the same term as a negative exact where it came from, so your own campaigns stop competing for it. This harvesting loop is the heart of the structure we describe in our guide to Amazon PPC campaign structure.
Minutes 12–15: product targets and the log
Filter the search term column for ASINs. Product pages that bring sales are candidates for a manual product targeting campaign; ones that bring clicks and no sales can be excluded as negative product targets. Finally, write down every change with the date. Amazon suggests letting negatives run for two weeks or more before changing strategy, and a simple log is how you know what to judge next week and what to leave alone.
A worked example
All figures below are illustrative. The product sells for $40, break-even ACoS is 30% and target ACoS is 20%. Last fortnight's report shows four terms worth acting on.
| Search term | Clicks | Spend | Orders | Sales | ACoS | Action |
|---|---|---|---|---|---|---|
| Term A (relevant) | 48 | $43.20 | 0 | $0 | n/a | Bleeder: lower bid or negate |
| Term B (relevant) | 60 | $66.00 | 5 | $200 | 33% | Above break-even: lower bid |
| Term C (from auto) | 30 | $24.00 | 4 | $160 | 15% | Winner: move to exact |
| Term D (wrong use) | 9 | $6.30 | 0 | $0 | n/a | Irrelevant: negate now |
Term B is worth keeping, but at the wrong price. It converts 5 times in 60 clicks, about 8.3%, so each click is worth roughly $3.33 in sales ($200 ÷ 60). At a 20% target ACoS you can afford about $0.67 per click ($3.33 × 20%), against the $1.10 you paid. Bring the bid down towards that level and review again in two weeks.
Mistakes that undo the routine
- Judging on too little data. Three clicks and no sale is noise, not a verdict.
- Reading only one campaign. The same search term can appear under several targets. Add up its clicks and sales across the account before deciding.
- Changing everything at once. If you cut bids, add negatives and launch new campaigns in the same week, you will not know which change worked.
- Skipping weeks. Waste accumulates quietly. A short weekly review beats a long quarterly clean-up.
Frequently asked questions
How often should I review the Amazon search term report?
Weekly suits most accounts. Very high-spend accounts may check bleeders more often, but give changes time to work. Amazon suggests letting negatives run for two weeks or more before changing strategy.
How many clicks should a search term get before I add it as a negative?
Amazon suggests evaluating a keyword after at least 20 clicks. Clearly irrelevant searches, such as the wrong product, can be negated straight away.
Why do recent days in the report show fewer sales than expected?
Sales are credited to an ad for several days after the click, commonly cited as 7 days for sellers and 14 days for vendors on Sponsored Products. Recent days fill in later, so leave them out of your review.
Is the old Sponsored Ads reports page going away?
Yes. Amazon has said the Sponsored Ads reports page will be sunset by 31 December 2026 in favour of unified reporting. Rebuild scheduled reports there and download any history you need, as of November 2026.
If you would rather have this review done every week without fail, Nexafern runs Amazon advertising for brands across every Amazon marketplace, with the Nf Engine flagging wasted spend and new winning searches each day. See how our advertising service works, or get in touch.