Performance Max spends your budget across Search, Shopping, YouTube, Display, and Gmail, and tells you almost nothing about where. The search terms are the part every advertiser wants back, so here's the honest version.

Why Performance Max hides the search terms

A normal Search campaign gives you a search terms report. Performance Max doesn't.

Google's line is that PMax is goal-based and automated, so raw queries aren't actionable. The practical effect: you can't see most of what triggered your ads, and you can't easily add negatives.

That's the black box people complain about, and it's deliberate. It also means your brand traffic and your prospecting traffic sit in the same campaign with no clean way to separate them.

What Google does show you

The campaign's Insights tab has a Search terms insights report. It groups queries into categories and subcategories with clicks, impressions, and conversions. Themes, not raw terms. Good for direction, useless for precise negatives.

Nearby you get Search themes, which are inputs you provide rather than a report, plus asset group and audience signal performance, and a placement report under Where ads showed for the YouTube and Display side.

None of it is the raw query list, but together it narrows the guessing. For retailers, the product and listing group reporting shows which items PMax actually pushed, another angle on where the budget went.

How to pull more into a report you control

The Google Ads API exposes campaign_search_term_insight, which returns those search categories with metrics. Query it with GAQL through a Google Ads script or the API:

SELECT
  campaign_search_term_insight.category_label,
  metrics.clicks,
  metrics.impressions,
  metrics.conversions,
  metrics.conversions_value
FROM campaign_search_term_insight
WHERE segments.date DURING LAST_30_DAYS
ORDER BY metrics.conversions DESC

Run it per PMax campaign. A Google Ads script can loop your campaigns and write rows to a Google Sheet or BigQuery.

Schedule the pull so you build history, because Google keeps only a rolling window and the labels drift. In BigQuery, join the categories to conversion value to see which ones actually pay, not just which ones get clicks.

Cutting the waste it reveals

Once you can see the categories, act on them. Account-level negative keywords are self-serve and apply everywhere.

Campaign-level negatives for PMax used to be a rep-only request and are now opening up in the interface, which lets you block a brand term or an irrelevant theme without nuking it account-wide.

Performance Max also takes brand lists: apply a brand exclusion so it stops eating your own branded search, which is the single biggest source of inflated PMax conversions. Pair that with placement exclusions for the Display and YouTube side, where a lot of quiet spend hides.

Make it a habit, not a one-off

Wire the scheduled pull into a Looker Studio page or a BigQuery table you check weekly.

Watch how category spend shifts after each asset or budget change, because PMax reallocates on its own and last month's winning theme can quietly turn into this month's waste. The advertisers who stay ahead of it treat this monitoring as routine, not as a fire drill when the cost per conversion jumps.

Where it breaks

You still get categories, not every raw query, so set that expectation with stakeholders up front. Category labels drift over time, so trend them carefully. And there's no historical backfill: if you don't schedule the pull, that data is gone.

Between Search terms insights and the API's category data, you can rebuild enough to steer budget and stop obvious waste.

Performance Max will still tell you almost nothing about where your money went on its own. A scheduled pull is how you answer that question anyway.

Want a stronger data analyst role or a raise? Grab the FREE Product Analyst Playbook and get the exact roadmap to your next offer.