Seen, clicked, bought: read the Search Query Performance report one share at a time

The Business Report tells you how many people visited your product page. It does not tell you what they searched for before they arrived, or how many people searched for the same thing and bought from someone else. The Search Query Performance report does.

For every search term that led to your product in a given week, it shows the whole market for that search, and your part of it. Every claim below points at a column in that report. We checked them on the weekly reports of 96 products, on six Amazon marketplaces, from May to September 2026: 8,203 rows in total.

Where the report is

In Seller Central it is under Brands, in Brand Analytics, as Search Query Performance. You need Brand Registry to see it. Through the API it is GET_BRAND_ANALYTICS_SEARCH_QUERY_PERFORMANCE_REPORT, one product and one week at a time.

The columns we read:

  • Search Query. What the buyer typed.
  • Search Query Volume. How many times it was searched that week.
  • Total impressions, total clicks, total purchases. What happened on that search across all products, not just yours.
  • Impression share, click share, purchase share. Your part of each total, as a percentage.

The report also has a cart stage between clicks and purchases. Everything below works the same way with it.

Impressions are not searches

Total impressions were higher than the search volume on 8,201 of the 8,203 rows. That is not an error. Every search shows many products, and every one of them counts as an impression. Do not divide your impressions by the search volume to get a reach figure. Use the impression share instead, which already compares you with everyone else shown on that search.

Empty is not zero

A share can be empty. In the data we checked, the purchase share was empty on 4,119 rows, and on every one of them the total purchases were zero: nobody bought anything on that search that week, from anyone. The same held for the 276 empty click shares.

A zero share means something else: the market did buy, and you got none of it. Sort by purchase share and the two end up next to each other, but they ask for different decisions. An empty row is a search that did not sell. A zero row is a search where someone else sold.

The three shares, compared

Each share on its own is just a number. The useful part is how they change from one stage to the next:

  • Click share higher than impression share. When buyers see you, they pick you more often than your visibility would suggest. Your image, title, price and rating work in the search results. This was the most common pattern: 5,670 of the 7,867 rows that had both shares.
  • Click share lower than impression share. You are shown, but skipped. The problem is in what the buyer sees in the search results: the main image, the title, the price, the star rating, the delivery date.
  • Purchase share lower than click share. Buyers click, look at the page, and buy somewhere else. The problem is on the product page, or in the price compared with what they saw next to you. On 2,680 rows, the market bought on a search where our products got clicks, and our purchase share was zero.
  • Purchase share well above impression share. You convert on this search, but you are rarely shown. On 229 rows, the purchase share was at least twice the impression share. These are the searches to put money behind: as a keyword in your ads, and, if it is not there yet, in your listing.

One week is too little

Most rows are small. On 3,582 of the 8,203 rows, the whole market made fewer than ten clicks that week, and on 6,851 rows fewer than ten purchases. At those sizes one buyer moves the share a lot: three clicks in total, one of them yours, is a 33% click share.

So do not judge a search on a single week. Add up the totals and your counts over at least four weeks, and calculate the shares from the sums. A search that keeps the same pattern over a month is a finding. A search that shows up once is noise.

Half an hour a month

Download the report for your main products, for the last four weeks. Drop the searches that had fewer than ten purchases in total, and the ones where the purchase share is empty.

Then make three lists. Searches where your click share is below your impression share: check your main image, title and price in the search results for that exact search. Searches where your purchase share is below your click share: open your product page on your phone, next to the competitor that took the sale. And searches where your purchase share is well above your impression share: these go into your ads first.