Broad Match Made Everyone's Keywords Identical. The Edge Moved to Structure.

Google's AI Max for Search has been live long enough that a pattern is starting to emerge: the advertisers who reorganized around it are pulling ahead, and the ones still running keyword-heavy account structures are watching CPAs drift upward without a clear explanation.

John Williams argued on Search Engine Land this week that paid search has become strategy-driven rather than keyword-driven. The thesis is right, but I think it undersells the urgency. This isn't a gradual evolution you can plan for next quarter. For a growing number of accounts, the keyword already stopped being the competitive input. Most teams are still structured like it is.

The keyword became a shared starting point

The mechanics are pretty simple. Google's match type definitions have been loosening for years. Exact match hasn't been exact since roughly 2018. Phrase match barely resembles what it was in 2021. And broad match, paired with Smart Bidding, reads user history, location, device, and behavioral context to decide where your ad shows. The keyword you typed into the interface is, at this point, a suggestion.

AI Max takes that further. Google's official position is that keywords now function as "inputs rather than instructions" for the system. In practice, that means two advertisers bidding on the same broad match keyword in the same vertical end up showing for roughly the same set of queries. The algorithmic interpretation dominates. Your keyword list, which used to be your competitive edge, became a shared commodity.

It's a bit like competing on who has the best GPS route when everyone is running the same navigation app. The directions are identical. The vehicle and the driver are what determine who gets there first.

Navah Hopkins at Optmyzr described the shift well: keywords are "morphing into signals versus syntax." Anne Bui at Puffer Digital went further, predicting keywords will become "extinct and replaced with clusters of search themes." Whether or not you think that timeline is aggressive, the directional trend is getting hard to argue with at this point.

So when everyone has the same keywords and Google decides the matching, what separates your campaign from the next advertiser's?

Three inputs that now do the differentiating

None of these are new. All of them just got significantly more important.

Conversion data quality. This is the big one. If you're sending Google the same generic conversion events as everyone else (form submit, purchase confirmation), the algorithm treats all of you more or less the same. The teams pulling ahead right now are feeding value hierarchies into their bidding. High-value leads scored differently from newsletter signups. Offline conversion imports that close the loop between ad click and actual revenue. Server-side tracking that doesn't lose 30% of events to browser privacy changes.

Williams calls this "the new keyword research," and that framing seems right to me. The effort that used to go into finding the right long-tail keyword now goes into making sure Google's bidding system understands which conversions actually matter to your business. If your conversion tracking is lazy, no amount of keyword optimization compensates for it. We wrote about this problem recently: feed the algorithm too many low-value signals and it optimizes for the easiest one, not the most valuable.

Creative as targeting mechanism. This one catches people off guard. Google's system reads your ad copy, landing page headlines, and images to determine who should see your ad. AI Max generates ad variations based on your landing page H1s and H2s. Your creative isn't just an ad anymore. It's a targeting signal the algorithm uses to find the right audience.

That changes the math on how teams should spend their time. I've been watching this play out for a while now, and it seems to hold up in practice: four creative variants tested against each other will usually outperform four audience segments targeted with identical creative. Search Engine Land's 2026 tactics piece makes a related point, recommending advertisers stop over-pinning RSAs and let the system test creative combinations more freely.

Account structure. When keywords did the heavy lifting, granular structures made sense. Hundreds of ad groups organized by keyword theme, each with custom bids. That overhead is now mostly dead weight. The emerging consensus is 7 to 10 ad groups per search campaign, organized by conversion intent rather than keyword taxonomy.

Simplified doesn't mean careless, though. The structure decisions that matter now are which campaigns receive which conversion actions, how budget flows between branded and non-branded terms, and whether your Performance Max campaigns are cannibalizing your search campaigns or actually complementing them. These are architecture decisions. Our Google Ads strategy guide covers the structural foundations, but the specifics keep shifting as Google rolls out features like AI Max.

Google's own data has a telling gap

Google claims AI Max delivers 14% more conversions at similar CPA. For campaigns heavily using exact and phrase match, the uplift is supposedly 27%. Those are Google's numbers, and Google is obviously motivated to make the feature look good.

Independent testing paints a more complicated picture. A 250-campaign study covered by Search Engine Land found a median revenue uplift of 13%, which is genuinely solid. But CPA increased by 16%. And roughly 84% of advertisers in the study saw neutral or negative results.

That gap between Google's claims and the independent data is where the real insight lives. The 16% of advertisers who benefited from AI Max didn't just toggle the feature on and walk away. They had clean conversion tracking, intentional account structures, and creative that gave the algorithm useful signals to work with. Everyone else plugged the same generic inputs into a more powerful system and got back more of the same, just at higher cost.

From what I've seen, that's the uncomfortable part of this whole transition. When the platform handles more tactical execution, the quality of your strategic inputs gets amplified. Good data and structure compound. Bad data and structure compound too, just in the wrong direction. And honestly, that second group doesn't always realize what's happening until the quarterly review.

A 30-minute structural checkup

If your paid search account was built more than 18 months ago and nobody has restructured it since AI Max started rolling out, you almost certainly have some structural debt. Here's a quick way to find out.

Open your account and count ad groups per campaign. If you're running more than 15 in a single campaign, you're probably over-segmented for how Google's system works now. Consolidate by conversion intent, not keyword theme.

Then look at your conversion actions. If every campaign optimizes for the same single event, you're giving the bidding algorithm a flat signal with no hierarchy. Set up value-based bidding with at least two tiers of conversion value. A demo request should not count the same as a whitepaper download.

Last, pull your search terms report for the past 90 days. If broad match is matching you to queries that your ad copy doesn't address, the problem is probably in your creative, not your keyword list. Write ad variants that speak directly to those query clusters. The algorithm matches better when your creative aligns with the queries it's already finding for you.

The whole thing takes about 30 minutes if you know your account. You'll come out knowing whether you're in the 16% that benefits from this shift or the 84% still running a 2022 account structure against a 2026 matching system.

I'd guess that by the end of this year, more than half of mid-market Google Ads accounts will have fewer than 10 ad groups per campaign. Down from 30 to 50 in the keyword era. The restructuring is already happening quietly. The question is whether your account adapts on your terms or on your CPA's.

By Notice Me Senpai Editorial