Google Ads rolls out Search Term categories for RSAs

Google Ads is testing a new feature that brings fresh transparency to responsive search ads (RSAs), a feature previously available only for PMax campaigns.

Called “Search Term Categories,” the beta tool groups user search queries into intent-based categories and subcategories, including those that were previously hidden due to privacy thresholds.

Why we care. The update offers advertisers aggregated performance metrics – such as conversions, click-through rate (CTR), impressions, and search volume – for each category, along with month-over-month trends. This empowers marketers to better understand which themes are driving performance and adjust campaigns accordingly.

How it works. Search terms that triggered your RSAs in the past 56 days are analyzed and categorized by intent and attributes. These insights appear in the RSA asset details view and are especially valuable for Broad Match targeting, where intent often outweighs precise keyword matching.

• Low-volume or unclassified terms are grouped under “other search terms.”

Zoom in. You can click into categories to reveal related subcategories, associated metrics, and even view which campaigns surfaced under a particular theme.

• Google shows two default metrics – conversions and search volume – but others like CTR or conversion rate can be toggled on.

What’s next: With this added layer of insight, advertisers are encouraged to replace low-performing creative assets, expand coverage into untapped keyword areas, and refine their ad copy to better align with user search behavior.

• Google suggests considering budget increases if shifting strategy toward high-performing categories.

Bottom line: Google’s new Search Term Categories tool could reshape how advertisers optimize RSAs, offering data-driven clarity on user intent while surfacing actionable creative and keyword insights that were previously hidden.

First seen. Search Engine Land learned of this update from senior digital advertiser Georgi Zayakov, who commented that this feature will be “especially useful with Broad Match, where matching happens more on intent than exact queries.”