PPC Fundamentals
How to Actually Read an Amazon Search Term Report
The search term report is the only place Amazon shows what shoppers really typed. How to read it by intent rather than by ACOS, and what most sellers miss.
Almost everyone opens the search term report the same way: sort by spend descending, look at the top twenty rows, negate anything with a bad ACOS, close the file. That takes four minutes and it is the single most expensive four minutes in Amazon PPC, because sorting by spend answers a question nobody asked. Spend tells you where the money went. It does not tell you whether the money was well spent, and it definitely does not tell you what to do next.
The search term report is the only place in the entire Amazon reporting stack where you see the actual string a shopper typed, joined to the actual targeting that matched it. Everything else — Search Query Performance, keyword-level reporting, the recommendations panel — is aggregated, modelled, or Amazon’s opinion. This report is the raw join. Read properly, it will tell you which of your assumptions about the product are wrong, which keywords deserve to be broken out into their own ad group, where you are paying twice for the same shopper, and which terms are converting well enough that they should never have been left inside an Auto campaign in the first place.
The shift that makes it useful is reading by intent group rather than by row. A single row is almost always statistically meaningless — twelve clicks and one order tells you very little on its own. But forty rows that all contain the phrase “for sensitive skin”, aggregated, might represent 600 clicks at a 12% conversion rate against your account average of 7%, and that is a campaign, not a keyword. The report rewards people who read it as a corpus of language and punishes people who read it as a list of line items.
This article covers the pass I actually run on a new account: how to pull the report at the right date range and match type granularity, the four buckets every term gets sorted into, why “customer search term” and “targeting” being identical does not mean what you think it means on broad match, and the specific traps — variant-level attribution, the 7-day/14-day window mismatch, and the sub-1% impression share rows that quietly eat 8% of a budget.
More coming soon. This piece is being written out in full. In the meantime, campaign structure and naming conventions covers the layer underneath it — the filing system this report feeds into.