The Amazon Search Term Report: What It Shows and How to Act On It

What the Report Actually Tells You

The Amazon Search Term Report shows the exact words a shopper typed into Amazon's search bar right before your ad served and they clicked. It comes out of Sponsored Products, Sponsored Brands, and Sponsored Display reporting separately — each ad type has its own version — and it's the only place you see the gap between the keyword you bid on and what people actually searched.

That gap matters most on broad and phrase match. If you're bidding "cast iron skillet," broad match can serve on "best pan for searing steak," "skillet gift set," or a hundred other terms you never typed. The search term report is where those show up, one row per term, with impressions, clicks, spend, orders, and sales attached.

Don't confuse it with the Search Query Performance report in Brand Analytics. That one is marketplace-wide — impression share, click share, cart-add share, and purchase share for a term across every seller, not just you. The search term report is account-specific: your spend, your clicks, your orders, against terms your own ads triggered.

How to Download It

In the Amazon Ads console, go to Measurement, Reporting, then Create report. Pick the ad product — Sponsored Products, Sponsored Brands, or Sponsored Display — and select the search term report under Targeting. Set a date range, request it, and it generates as a downloadable .xlsx or .csv file. For accounts running this weekly across dozens of campaigns, the Amazon Ads API can pull the same report on a schedule instead of clicking through the console each time.

One setting trips people up: the report is scoped to a single ad product. Pull Sponsored Products and Sponsored Brands separately and you get two files, not one combined view — stitching them into a single account picture happens after export, not inside Amazon's tool.

Sales don't post instantly. A click today can convert several days later, so a report pulled the morning after a launch will understate real performance. Give it time before drawing conclusions, especially in the first week of a new campaign.

A Worked Example: Turning Rows Into Decisions

Say a Sponsored Products campaign for a kitchen brand pulls a two-week search term report. These aren't real account figures — they're illustrative, to show the read:

• "cast iron skillet 12 inch" — 40 clicks, $38 spend, 6 orders. Cheap and converting. Add it as an exact-match keyword at a bid you control, then negative-exact it under the broad campaign that surfaced it, so the two don't compete for the same auction.
• "nonstick pan gift set" — 65 clicks, $71 spend, 0 orders. High spend, zero return, off-topic for a skillet listing. Negative it before it burns another week of budget.
• "best skillet for searing" — 12 clicks, $9 spend, 1 order. Low volume but relevant and profitable. Leave it running under broad match; it hasn't earned a dedicated keyword yet.

That's the whole discipline: harvest what converts into its own keyword, cut what spends without returning, and leave undecided terms alone until they've earned a verdict either way.

Common Mistakes (Including Ones We've Made)

The most common mistake is negating too fast on broad match. A term with two clicks and no order in three days isn't proven bad — it's unproven. Cut it now and you might be removing a term that converts on click four.

The second is comparing a search term's ACOS to a blended account average instead of the actual break-even for that specific product. A skillet with thin margin and a skillet with fat margin don't share a target ACOS, even inside the same campaign.

The third — one we've made ourselves — is harvesting a converting term into exact match and forgetting to negative-exact it out of the broad or phrase campaign that found it. Both keywords then compete in the same auction, you pay twice to reach the same shopper, and the report starts looking noisier than the account actually is.

A useful bar for any search-term decision, whatever tool you use to make it: what's the evidence behind the change, how will you know if it worked, and what's the way back if it didn't. That standard holds up whether you're doing it by hand in Excel or at full account scale.

When the Numbers Don't Match What You Expected

If a term you know converted shows zero orders in the report, check the date range first — attribution can take several days to post, so a report pulled too soon will understate real performance. Check next whether orders are landing on a different child ASIN than the one you're tracking; Amazon counts the sale, but on the variation the shopper actually bought.

If spend looks wildly higher than expected, check you're not accidentally viewing a lifetime report instead of the date range you meant, or that you haven't pulled Sponsored Products and Sponsored Display into the same read as if they were one campaign type. And if a term simply isn't in the report at all, it may not have generated a click yet — impressions alone don't populate every column.

Where This Fits Into a Bigger Ad Account

Reading a search term report by hand works fine at one or two SKUs and a modest weekly spend. It gets harder at scale — a real account can carry hundreds of search terms across a dozen campaigns, refreshed weekly, and the harvest-and-negative discipline above has to run every time without someone forgetting the negative-exact step.

Dr. PPC reads that report as part of a wider account picture and proposes the negative, the harvest, or the bid change with the evidence, a measurement plan, and a rollback trigger attached to each one — the same checks described above, run at the account's actual scale. It's $300 a month plus 3% of ad spend, capped, month-to-month, with the first 30 days free, and it includes Orbit, Full Circle's own analytics suite — search-term and profitability tracking many tools sell as a separate subscription — at no extra charge. Full Circle has managed more than $500M in revenue across 100+ brands. None of that replaces reading your own search term report at least once; it just does it at a pace a spreadsheet can't keep up with.