Search Now Gets 41% of Google Ads Clicks (and Still Eats 71% of the Budget)

Search campaigns fell from 62% to 41% of Google Ads clicks between January 2024 and August 2026 across 31 Brainlabs-managed accounts, while still taking 71% of spend. That gap means a Search click now costs roughly 3.5 times a click from Performance Max or Demand Gen. Blended CPC is falling mostly because of mix, not because anything got more efficient.

The numbers come from a Brainlabs post by Alessio Messaggi, published October 7 and picked apart in detail by PPC Land on October 11. The post itself is mostly about AI referral traffic and ChatGPT ads. The campaign-mix chart is almost a side note. I think it's the most useful thing in there for anyone setting a Q4 budget this month.

The math hiding inside two percentages

Here are the four data points. In January 2024, Search had 86% of spend and 62% of clicks across those 31 accounts. By August 2026, it had 71% of spend and 41% of clicks. The rest went to Performance Max (with Shopping) and Demand Gen (with video), and Demand Gen's slice of clicks seems to have widened sharply in spring 2026, getting close to a quarter of clicks around May.

Now do the division. In August 2026, Search delivered 41% of clicks for 71% of the money, so its cost per click ran about 1.7 times the account average. Everything else delivered 59% of clicks for 29% of the money, about half the account average. Put those two side by side and a Search click costs roughly 3.5 times a non-Search click.

Run the same math on January 2024 and you get about 3.8 times. So the relative price of a Search click barely moved in two and a half years. What moved was how many cheap clicks got added on top.

That's the part I'd sit with for a minute.

If your monthly report shows blended CPC down 20% year over year, it's tempting to put that on a slide as an efficiency win. In most accounts I'd bet a good chunk of it is just more Demand Gen video and PMax inventory diluting the average. Those clicks may be fine. They may even be great. But they're a different product, and averaging them with high-intent Search clicks is a bit like averaging the price of a steak dinner with the price of the free bread and calling the restaurant cheaper.

Caveats worth taking seriously (because Brainlabs didn't separate them)

This is one agency's data, from 31 accounts, and the mix shift partly reflects Brainlabs' own allocation choices. PPC Land points out that the post doesn't separate "the market moved" from "we moved our clients." That matters. An agency that has spent two years pitching PMax and Demand Gen will, unsurprisingly, have clients running more PMax and Demand Gen. For context, a 2023 Brainlabs study of 18 clients found they were putting an average of 17% of their Google Ads budget into Performance Max. Non-Search spend in this newer dataset is 29%, so the direction lines up with how the agency has been advising people.

There are also seasonal dips that tell their own story. Search click share bottomed at 31% in December 2024 and hit 33% in May 2026, and Search spend share dropped to 64% in December 2024 and 56% in November 2025 before bouncing back. That pattern looks a lot like Q4 budgets getting pushed into Shopping-heavy PMax, which is exactly the season we're heading into.

And honestly, the label "Search" is getting slippery anyway. A Search campaign running AI Max for Search can match queries your keywords never covered, so even the 41% isn't all classic keyword-driven traffic. Google also now says that for AI Mode, AI Overviews, Lens and auto-complete searches, the term in your report is "the best approximation of the user's intent" rather than what the person typed. So the campaign-type split is a rough proxy for intent, not a clean measure of it.

Why the cheap clicks are probably not the problem

None of this means PMax and Demand Gen clicks are junk. That would be the lazy read. Demand Gen is built to catch people earlier, and a video click that starts a consideration path shouldn't be judged on last-click CPA against a branded Search term.

The problem is reporting, mostly. Budget conversations still tend to happen at the campaign-type level ("we're spending X on Search, Y on PMax"), and that view no longer tells you where search-intent money actually goes. PMax buys Search inventory too. AI Max Search buys queries you didn't pick. Meanwhile the paid Search clicks that do happen are getting scarcer on informational queries. Seer Interactive found paid CTR on informational queries with AI Overviews fell 68%, from 19.7% to 6.34%, between June 2024 and September 2025.

Put those together and you get a slightly awkward situation: the expensive, high-intent clicks are a shrinking share of volume, and the reports most teams use to set budgets blur them together with everything else. Brainlabs' own organic data runs in a similar direction. Across 54 advertisers, organic sessions fell 10.5% after AI Overviews became common, from 140.1 million to 125.4 million. Fewer clicks, everywhere, with the remaining ones arguably worth more.

A click-mix audit to run before the Q4 budget locks

This takes maybe 45 minutes in most accounts. Pull the last 12 months and build a table with one row per traffic source. Not per campaign type. Per source.

1. Split PMax by channel. The PMax channel performance report breaks out Search, YouTube, Display, Discover, Gmail and Maps with cost, clicks, conversions and value. If you want it at scale, we covered the API channel segmentation that landed in v24.2. Move PMax's Search row next to your Search campaigns.
2. Add four columns: share of clicks, share of spend, share of conversions, share of conversion value. Do it for "all search-intent inventory" (Search campaigns plus PMax Search and Shopping) versus "everything else."
3. Apply one benchmark. If a source's share of clicks is more than double its share of conversion value, it's buying reach more than capturing demand. That can be fine, but then it belongs in a different budget line with a different success metric (incrementality or a holdout, not ROAS).
4. Re-plot CPC trends by source, not blended. If Search CPC is up 15% while blended CPC is down, you want that on the slide, not buried under the average.

One more thing on step 3. If you're running AI Max, it's worth checking whether its expanded queries are quietly inflating the "Search" bucket with less valuable traffic. Our piece on AI Max holdouts and CPCs has a simple way to think about testing that without blowing up the account.

Where I think this goes by next October

My guess, and it's only a guess: in accounts like the ones in Brainlabs' sample, Search campaigns drop below a third of Google Ads clicks by Q4 2027, while still holding more than 60% of spend. Demand Gen is the obvious grower, and AI Max blurs whatever is left. If that happens, campaign-type budget splits become close to meaningless as a planning tool.

Anyway, I'm not sure the right move is to fight the mix. Google is clearly steering money toward inventory it can price more flexibly, and some of that inventory works. I'd just want to know which clicks I'm actually paying for before somebody decides next year's number based on an average that hides a 3.5x price gap.