The Search Query Performance Report: Reading the Four Stages

Search Query Performance is a Brand Analytics report showing how the entire marketplace behaves on a search query, not just how your own ads performed. It sits in Seller Central under Brands, then Brand Analytics, then Search Query Performance, and it requires Brand Registry plus permission from your brand's primary administrator.

The report tracks four stages for every query: impressions, clicks, cart adds, and purchases. Each stage reports a marketplace total, your brand's count, and your brand's share of that total. Brand view covers the top 1,000 queries, pullable weekly, monthly, or quarterly.

Read it by comparing your share across consecutive stages rather than reading any number on its own. Share that survives impressions and collapses at clicks is a different problem from share that survives clicks and collapses at cart adds, and the stage where the drop happens tells you which part of the listing to go fix.

TWO REPORTS, SIMILAR NAMES, VERY DIFFERENT JOBS

The Search Term Report lives in the Ads Console and shows the queries that triggered your paid ads, with your spend and your conversions. Search Query Performance lives in Brand Analytics and shows every query in your category, with the whole market's behavior alongside yours.

One tells you what your campaigns did. The other tells you what the market did while your campaigns were doing it. Sellers conflate them constantly, and we have made the mistake in our own published work, because the names are nearly identical and both contain the words search and query.

The distinction has a practical edge. If your Search Term Report shows a term converting well, you know that term works for you. Only Search Query Performance tells you whether you captured two percent of that query's purchases or forty, and that number is the one that says whether to scale.

SHARE AND COUNT ANSWER DIFFERENT QUESTIONS

Every stage gives you both a count and a share, and they move independently. Count is your absolute volume. Share is your position against everyone else competing on that query.

The combination that catches people out is a rising count with a falling share. Your clicks went up, which looks like progress, while your share of the query's clicks went down, which means the query grew faster than you did. You gained traffic and lost ground in the same week.

Put numbers on it. A query draws 40,000 searches one week and you take 900 clicks, which is 12% of that query's clicks. Four weeks later it draws 70,000 searches and you take 1,100 clicks, which is 8%. Your click count rose by a fifth while your share fell by a third, and only one of those two numbers tells you a competitor moved in.

Reading count alone produces false confidence in a growing category and false alarm in a shrinking one. Reading share alone hides whether the query is worth anything. The honest read needs both, and the report gives you both specifically so you can tell the difference between your performance changing and your market changing.

THE STAGE THAT THREE-STEP DIAGNOSIS SKIPS

Between clicking and buying there is a cart add, and it is the most diagnostically useful column in the report. A shopper who adds to cart has decided they want the product. A shopper who clicks has only decided to look.

That distinction splits what a collapsed funnel reads as one problem. High click share with low cart-add share means your listing failed to convince someone who arrived interested, which points at content, images, bullets, and A plus modules. High cart-add share with low purchase share means you convinced them and then lost them at the final step, which points at price, review count, delivery promise, or stock.

In practice a cart-add gap sends you to the part of the page a shopper reads after the first scroll. That usually means bullets four and five, where most listings stop making an argument, the A plus modules answering objections nobody wrote down, and the later images that would establish scale or use case if anyone had shot them.

Those two failures need completely different fixes, and a three-stage read showing impressions, clicks, and conversion cannot tell them apart. Both appear as weak conversion. One is a content project and the other is a pricing or operations decision, and running the wrong one costs a quarter.

When Ray-Ban's click-through rate moved from 0.02% to 20% in our work with them, the fix was entirely upstream at the click stage. Nothing about that project would have helped if the real failure had been sitting at the cart-add line.

EVERY RATE COLUMN DIVIDES BY THE SAME NUMBER

This is the mechanic that inverts conclusions, and almost nobody explains it. The rate columns in Search Query Performance calculate against Search Query Volume, not against the preceding funnel stage.

So cart add rate is not cart adds per click, the way you would compute it anywhere else. It is cart adds divided by total searches for that query. Purchase rate is purchases divided by total searches, not purchases per cart add.

Here is the inversion. Take two queries both converting a fifth of their clicks into cart adds. The first draws 2,000 searches and posts a cart add rate near 4%. The second draws 90,000 searches and posts a cart add rate under 0.2%. Identical performance on the traffic each one sent, and the column makes the second look twenty times worse.

Read those columns as stage-to-stage conversion and you will reach backwards conclusions on high-volume queries, because a big denominator makes every rate look small regardless of how well you convert the traffic you get. If you want true stage-to-stage conversion, compute it yourself from the count columns. The rate columns are useful for comparing queries against each other, not for diagnosing a funnel.

IT WILL NOT RECONCILE WITH YOUR BUSINESS REPORTS, AND THAT IS CORRECT

Search Query Performance is query-level and marketplace-wide. Business Reports are session-level and limited to your own detail pages. The two measure different populations and will not tie out, no matter how long you spend trying.

Use them for different questions. Search Query Performance answers how you performed on a query against the field. Business Reports answer how your own listing performed once someone arrived on it, through sessions, unit session percentage, and Featured Offer percentage.

Sellers lose real time trying to make these agree. They are not supposed to agree, and a discrepancy between them is not a data problem to solve.

WHERE EACH DIAGNOSIS SENDS YOU

A drop at the click stage is a search-results problem, which means main image and title are doing the work. With titles capped at 75 characters since July 2026, the prioritization of which keywords survive in a title directly determines whether you earn the click you were shown for.

A drop at the cart-add stage is a listing content problem, and the conversion levers worth prioritizing first are the ones that hold a shopper who already arrived interested. A drop at the purchase stage sends you to price, reviews, and availability instead, and no amount of listing work will move it.

For the research side, meaning which queries belong in your account at all rather than how you are performing on the ones already there, the Amazon keyword research guide covers the four data sources worth mining. For how these diagnoses feed bid and structure decisions, the Amazon PPC management guide covers the campaign side.

Pull the report weekly rather than monthly. Share moves faster than most sellers check, and a four-week-old read of a query you are losing is a report on a decision you already failed to make.

---

FAQ

What counts as a good brand share?

There is no universal threshold, because share depends on how many sellers compete on the query and how broad it is. The useful comparison is your own share over time on the same query, and your share at one stage against your share at the next. A brand holding 30% of impressions and 4% of purchases on a query has a problem regardless of what any benchmark says.

Does the report separate paid performance from organic?

No, and this is its most significant limitation. The counts combine both, so a query where your share looks healthy may be held up entirely by ad spend. Cross-reference against your Search Term Report to see how much of your share you are renting.

Why do I have impression share on queries I never targeted?

Amazon surfaces your products for queries it considers semantically related, whether or not you bid on them or placed those words in your listing. Finding unexpected impression share is one of the more useful things the report does, because a query showing you impressions and no clicks is a relevance mismatch Amazon believes in and shoppers do not.

Can sellers without Brand Registry get this data?

No. Brand Analytics requires an active Brand Registry enrollment tied to the selling account. Without it, the closest available substitute is the Sponsored Products search term report in the Ads Console, which shows your own paid queries but nothing about the wider market.

Should I pull weekly, monthly, or quarterly?

Weekly for anything you are actively working on, because share shifts inside a month and a monthly average hides the week it moved. Quarterly views are better for deciding which queries deserve attention at all, since they smooth out promotional spikes and stockouts that make a single week misleading.