The Amazon search term report is the record of what shoppers actually typed before your ad appeared: the real searches, not the keywords you bid on. Those are two different lists, and the gap between them is where most of an account's wasted spend and most of its cheapest growth both hide. I open this report before anything else in an audit, because every structural problem I find leaves fingerprints here first: zero-order bleed, broken harvest loops, campaigns splitting each other's data. This guide covers how I read it: the keyword-versus-search-term distinction most sellers blur, where to download the report and which columns matter, the four things I hunt for on every pass, and the free tools that turn the read into negatives, harvests, and bid changes.
A keyword is what you bid on. A search term is what the shopper typed
A keyword (or target) is an instruction you give Amazon; a search term is the query a real shopper typed into the search bar before clicking your ad. Exact match keeps the two nearly identical. Broad match, phrase match, and auto targeting fan one keyword out into hundreds of search terms you never chose. Bid broad on "laptop sleeve" and you will show up for sleeves, cases, covers, competitor names, and searches you could not have invented.
This is why campaign-view metrics mislead. A broad keyword's row in the console is an average of every search it matched: strong searches and junk blended into one ACOS. The keyword can look mediocre while one search term underneath it prints money and five others burn it. You cannot manage what you have averaged away, and the search term report is the only place the average comes apart.
| Keyword / target | Search term | |
|---|---|---|
| Who writes it | You, in campaign settings | The shopper, in the search bar |
| What it is | A bid plus a match type: an instruction | The actual query that triggered your ad |
| How many exist | One row per target you created | Hundreds per broad or auto target |
| Where you see it | Campaign manager and targeting tabs | The search term report, nowhere else |
| How you control it | Bids, budgets, match types | Only through negatives and harvests |
Where to get the report, and the columns that matter
The version I work from comes out of the bulk file: go to Amazon Ads > Bulk Operations > Download, and include Sponsored Products Search Term Data (plus Sponsored Brands Search Term Data if you run SB) for the last 60 days. The ads console's report builder can generate a standalone search term report too, but I use the bulk sheet because it is the same file my free tools and Audit Dashboard parse.
Pull 60 days, not 30. Thirty days misses slow-converting search terms and seasonality; sixty catches terms that genuinely do not convert without false-flagging longer purchase cycles.
Nine columns do the real work: the customer search term, the keyword or target that matched it, the match type, clicks, orders, spend, sales, ACOS, and conversion rate. If your export does not include conversion rate, compute it: orders divided by clicks. Everything else in the file is context. These nine are the verdict: what the shopper typed, which of your instructions caught it, and whether the money came back.
The four things I hunt for in every search term report
Every row in a search term report sorts into one of four buckets: winners you do not own yet, bleeders that never convert, expensive converters that need a smaller bid, and duplicates splitting data across campaigns. Reading the report well is nothing more than assigning rows to buckets fast and honestly.
1. Converting search terms to harvest into exact match
Any search term converting repeatedly under broad, phrase, or auto targeting that you do not own at exact match is a harvest candidate. Move it into an exact match keyword so you control its bid, budget, and placement directly instead of renting it through a discovery setting. This is the growth half of the report, and it is the half most sellers skip: negation feels urgent, harvesting feels optional. It is not. A broken harvest loop means you are paying Amazon, every month, to rediscover terms you have already bought, which is exactly what a bloated auto-campaign spend share looks like on my benchmark table.
2. Zero-order terms bleeding spend: negate them
Any search term that took clicks and spend without producing a single order is bleed, and the fix is a negative keyword in the campaign it came from. Some of this is tuition: discovery costs money, and a term needs enough clicks to prove itself before you judge it. What is never tuition is the same unconverting term taking clicks month after month because nobody negated it. Sort the report by spend with orders filtered to zero and the worst offenders introduce themselves.
3. High-ACOS terms that still convert: bid down, don't kill
A search term that converts but at double your target ACOS is not bleed, and negating it is the most common self-inflicted wound I see in the report. The term produces orders; kill it and you surrender that revenue plus whatever organic rank the sales were feeding. The right move is a bid cut sized to the gap: pay what the term is actually worth per click instead of what your match type happens to charge. Bleed gets negated. Expensive converters get repriced. Blurring those two buckets costs real money in both directions.
4. Search term bleed: the same term split across campaigns
Search term bleed is the same search term surfacing across multiple campaigns because broad or phrase keywords in one campaign match the searches another campaign was built to own. The conversion data splits: each campaign sees a fragment of the term's real performance, so no campaign has the statistics to optimize it, and your effective ACOS on the term inflates while every individual row still looks plausible. This is the problem you cannot see reading one campaign at a time, which is exactly why most sellers miss it. Sort the whole report by search term instead of by campaign and the duplicates line up. It is one of the nine panels in my free Audit Dashboard, because fixing it is structural: deciding which campaign owns the term and negating it everywhere else.
From reading to acting: three moves, and a free tool for each
The read produces three mechanical moves: negate the bleed, harvest the winners, reprice the expensive converters. I built a free tool for each job, and all three parse the same bulk file in your browser with no email and no account.
The Wasted Spend Finder ranks the zero-order bleed in dollars: every keyword and search term that spent money over 60 days with nothing to show for it, grouped by campaign and sorted by worst offender. Run it first: it sizes the leak before you spend a weekend fixing it. My benchmarks guide has the phase-by-phase thresholds for how much zero-order spend is tolerable, from 50–60% at launch down to 25–30% for a mature account.
The N-Gram Analyzer finds the word-level patterns hiding across your bad search terms. If the same two-word pattern shows up in fourteen different non-converting searches, you do not need fourteen negatives; you need one phrase-match negative on the pattern. Negate once, not fifty times.
The Negative Keyword Finder builds the negation list itself: zero-order and high-ACOS terms, each matched to the right negative type, exact versus phrase, in a ready-to-upload format. It closes the loop the other two tools open.
Harvesting is manual work no tool should automate blindly, and it pays twice. Exact-match spend on proven terms buys efficiency now, and the concentrated sales push organic rank, which is how ad spend turns into revenue you stop paying for. My ACOS vs TACOS guide walks through that mechanic, including a real account where concentrating budget on converting head terms helped double monthly sales.
A worked example: reading one real dataset
Take the dataset I graded in my benchmarks guide: $36,303 in ad spend producing $111,058 in ad sales, from 4,140 orders on 45,672 clicks. The toplines compute to a 32.7% ACOS, a 9.1% conversion rate, a $0.79 average CPC, and a $26.82 average order value.
Here is what those averages hand you before you read a single row. A 9.1% conversion rate means this account produces roughly one order for every eleven clicks. So when the search term report shows a term with dozens of clicks and zero orders, that is not bad luck; at the account's own averages it should have converted several times by now. At $0.79 a click, every one of those rows is real money leaving quietly, and there is never just one of them.
The same logic runs in reverse for harvest candidates. Any term converting well above 9.1% under broad or auto targeting has earned an exact match keyword; the account average is the bar the winners have to clear. And the toplines alone can never tell you which rows are responsible for the 32.7% ACOS. That verdict lives entirely in the report. If you want to see this read performed on real rows before running your own file, the free audit demo renders it on a real anonymized account: no upload needed.
Frequently asked questions
What is the Amazon search term report?
It is the log of the actual customer searches your ads appeared on, with the keyword that matched each one and the clicks, spend, sales, and orders that followed. It is the only report that shows what you actually paid for, as opposed to what you asked to pay for.
What is the difference between a keyword and a search term?
You write keywords; shoppers write search terms. A keyword is a bid plus a match type, an instruction to Amazon. A search term is the query a shopper actually typed when your ad appeared. Exact match keeps the two aligned; broad, phrase, and auto targeting let one keyword collect hundreds of different search terms.
Where do I download my search term report?
From the bulk file: Amazon Ads > Bulk Operations > Download. Tick Sponsored Products Search Term Data, add Sponsored Brands Search Term Data if you run Sponsored Brands, and set the window to the last 60 days. That one .xlsx feeds every free tool on this site.
What is search term bleed?
Search term bleed is when broad or phrase match keywords in one campaign surface the same search terms as another campaign, splitting the conversion data and inflating your effective ACOS. Each campaign sees a fragment of the term's performance, so none of them can optimize it. The fix is ownership: one campaign gets the term, the rest get negatives.
How often should I review my search term report?
Monthly for a stable account, always over a 60-day window so slow converters get a fair trial. Review more often, every week or two, when discovery is running hot: during a launch, after adding broad match, or whenever auto campaigns take a bigger share of spend, because exploration mints new search terms fastest.
Run the read on your own account
The fastest version of everything above is the free Account Health Snapshot: upload your bulk file and it grades nine metrics, wasted spend included, in about 60 seconds, parsed in your browser with no email and no account. And if your report raises a question a scorecard cannot answer, the free 30-minute diagnosis call is where I read the file with you, row by row.