Search Terms

The search terms report is the record of what people actually typed into Google that caused your ad to show, pulled from live queries rather than from the keyword list you configured. It exists as a separate report specifically because match types are approximations of intent, not exact predictions - the only way to know precisely what a broad, phrase, or exact match keyword bought is to look at the queries themselves, which is what this report shows, mapped back to the keyword and ad group that triggered each one.

You mine it for three distinct purposes that pull in different directions: adding negatives to block queries that are technically matching but not valuable, promoting strong-performing queries into their own dedicated keywords (often exact match) so you can control their bid and budget independently, and sanity-checking a match type decision by confirming the traffic it is actually buying looks like what you intended when you set it up. Broad match campaigns lean on this report the hardest since their query variance is widest, but phrase and exact match keywords should still be checked periodically since close variant matching means even a tight setup can surface unexpected terms.

The detail that catches people off guard is that Google withholds search terms it judges to be low search volume or potentially privacy-sensitive - these show up as unlabeled or aggregated rows, or do not appear at all, rather than being listed individually. This means the report never reconciles perfectly against total account spend or click volume; there is always some remainder of traffic you cannot see broken out by exact query. The threshold and exact withholding logic are not fully published by Google and can shift, so treat the gap as a structural feature of the report rather than something to troubleshoot as an error. This is also why the report should be read as a sample of your traffic's composition rather than a complete ledger of every dollar spent.

The report interacts with match type strategy directly: it is the input you use to decide whether a keyword's current match type is buying what you want, and it is the source list for negative keyword additions and for keyword promotion. It also interacts with Performance Max and other automated campaign types more loosely - those campaign types expose a smaller, less granular version of search term insight (search category insights rather than a full report), which is a real limitation worth knowing before assuming you can audit a Performance Max campaign's traffic the same way you audit a Search campaign's.

The mistake is judging keyword-level performance from the search terms report alone while implicitly assuming the withheld remainder is negligible - on accounts with a lot of low-volume long-tail traffic, the hidden portion can be a meaningful share of total clicks, and a query mix that looks clean in the visible rows can still be masking waste in the withheld ones. Cross-check the report's totals against the keyword report's totals periodically; a large, persistent gap between the two is worth noting even though you cannot break it down further.