SellerTacos

Insights: are you marketing the right product for each keyword?

Reports → Advertising → Insights looks for keywords where your ad spend goes to a product that shoppers rarely add to cart from that search, while another of your products gets added to cart far more often. It works on one Brand Analytics period (week, month or quarter), not a date range, because cart adds per keyword per product exist only in the Search Query Performance report.

Where the spend per product comes from

Amazon reports a search term's spend per ad group, and the products in each ad group, but not which product a click went to. So:

  • In an ad group advertising one product, the spend is that product's exactly.
  • In an ad group advertising several, it is split in proportion to each product's own spend there.

Each result shows how much of its spend was exact. In practice only about a quarter to a third of spend is exact, so treat shared ad groups as estimates.

What a suggestion means

A switch pair shows a keyword's marketed product and a better one:

  • The marketed product takes at least 40% of the keyword's attributed spend, with at least 10 ad clicks and 10 search clicks.
  • The better product has at least 10 search clicks, at least 3 cart adds and a cart-add rate at least 1.5× the marketed one's.
  • Spend at stake = marketed spend × (1 − marketed rate ÷ better rate).

Only variations in the same product group are suggested: an ad for one variation can sell another, so the question is which variation to market, never an unrelated product. Set product groups in the ASIN Library, or let Group variations fill them from Amazon's variation families. A banner counts keywords left out because the marketed product has no group.

High-volume keywords (the top 5% by searches) can list up to two more products that also convert better. A product Brand Analytics does not report on a keyword is never judged there: unknown is not zero.

Best and worst combinations

  • Best: keyword × product combinations with spend, at least 10 ad clicks, 10 search clicks and 5 cart adds, lowest ad cost per cart add first.
  • Worst: spend with no cart add first, then the highest cost per cart add.

About the click counts

Ad clicks and search clicks differ by a lot (Sponsored Products counts about twice what Brand Analytics credits), so the page labels them apart: "N of M search clicks · ₹X ad spend on K ad clicks". The spend is right; Amazon counts the two clicks differently.

Small click counts swing from week to week, so the page shows the clicks and cart adds behind every rate. Compare opens a side-by-side of the two products.

Still stuck? Ask us from the chat bubble in the corner of the page. Tell us the store, the page and the date range.