Google’s AI Mode Is Growing Fast. The Attribution Data Around It Is Shrinking.

Google just told investors that AI Mode earns less money per search than classic results do. That is a real admission. What it didn’t mention: this year Google also narrowed the AI Mode attribution data advertisers can see, right as AI Mode scales. The two trends moving together are the real story here.

• Google’s Q3 2026 Search revenue grew 12 percent year over year, down from 19 percent in Q1 and 17 percent in Q2, even as search queries hit a five year high.
• Chief Business Officer Philipp Schindler named “lower monetization in AI Mode” as a factor in that slowdown, and Alphabet’s stock fell roughly 10 percent after the call.
• The same year, Google quietly narrowed the Branded Searches reporting window from 30 days to 7 and dropped Performance Max from the campaign types that get Branded Searches data.
• Google Ads help documentation was also updated in 2026 to say search term reports may show only “the best approximation” of what someone typed, not the literal query, whenever AI Mode, AI Overviews, Lens, or autocomplete are involved.

Why Did Google’s Search Revenue Growth Slow to 12 Percent?

Search queries hit a five year high in the third quarter of 2026, but Search revenue barely moved, up just 12 percent year over year to $63.35 billion, versus 19 percent growth in Q1 and 17 percent in Q2. Schindler told analysts that AI Mode earns less per query than the classic results page it is replacing, and that gap is one reason growth decelerated so fast.

Here’s what I’d ask if I were sitting on that earnings call: if AI Mode is pulling in record query volume and still not paying for itself, how exactly is Google going to close that gap? Schindler didn’t say. Google’s answer so far is more ads inside AI Mode itself, including multiturn ads rolled out to more markets in late October.

What Did Google Change About Attribution Reporting This Year?

Three separate, unannounced changes landed in 2026, and each one makes it harder to independently check Google’s own monetization story. None of them got a press release.

None of these three came with a changelog entry, a rollout date, or an explanation. Advertisers found all of them the same way: by noticing the help documentation had quietly changed.

Is Performance Max Quietly Absorbing Brand Search Too?

Not absorbing it. Filtering it out, according to updated Google Ads help pages Search Engine Roundtable flagged on November 3, 2026. The documentation now says Performance Max may filter out navigational or brand specific queries so dedicated Search campaigns can serve them instead. The page wasn’t there in a June 17, 2026 snapshot of the same help article, so the change landed sometime in between. Google hasn’t said when, or why.

That detail matters more than it looks. Brand search used to be the metric advertisers pointed to when they wanted to show Performance Max was actually building brand demand, not just harvesting it. Pull that signal out of PMax and you’ve pulled out one of the few numbers that let an advertiser argue with the algorithm.

What Should Attribution Teams Do About It?

You can’t out-dashboard a reporting change you can’t see coming. A few things do help:

1. Export and archive search term and branded search data monthly, not quarterly, so a 7 day window change doesn’t quietly erase your trend line.
2. Track Performance Max brand query volume outside Google Ads too, through server side analytics or a CRM source field, so a filtering change doesn’t leave you blind.
3. Treat every “approximate intent” search term row as a label, not a transcript, when building negative keyword lists.
4. Re-run mix model and incrementality checks on any account where AI Mode traffic is material. Our own look at which channels are getting new ad budget that mix models can’t see is worth revisiting here.

The Pattern Nobody at Google Is Connecting Out Loud

Put the two halves together and you get a conclusion none of these sources states on its own: Google’s asking advertisers to trust that AI Mode will eventually monetize as well as classic search, while it’s shrinking the independent evidence advertisers would need to check that claim. That’s not a conspiracy. It’s just what happens when the product growing fastest is also the one with the thinnest paper trail. We already flagged this direction when Google cut attribution choice back in September; the Q3 earnings call just gave it a dollar figure.

FAQ: AI Mode and Attribution Reporting

Does Google explain why AI Mode earns less than classic search?
Not in detail. Schindler called it lower monetization per query and said the gap should close as more ads run inside AI Mode’s answers, but Google hasn’t published a number for the gap itself.

What exactly changed in the Branded Searches report?
The default lookback window dropped from 30 days to 7, and Performance Max was removed from the list of campaign types that generate Branded Searches data, a change Google never announced or dated.

Can I still see exact search terms for AI Mode traffic?
Sometimes. Google’s own documentation now says the report may show an approximation of intent rather than the literal query whenever AI Mode, AI Overviews, Lens, or autocomplete are part of the path.

Should I turn off Performance Max for brand terms?
Not automatically. Google’s help pages now say PMax may filter brand and navigational queries on its own, so test before you build a manual exclusion on top of a filter that might already exist.

Sources: Alphabet’s Q3 2026 earnings commentary as reported by RelevantAudience; the Branded Searches window and Performance Max change, also via RelevantAudience; the Performance Max brand and navigational query filter reported by Search Engine Roundtable; and the May 2026 search term approximation update documented by TechWyse. All figures are Google’s own or as reported by these outlets in 2026; none of this is independently audited by a third party, which is itself part of the point.