Google Search Grew 12%, Meta Grew 26%: The Search Click Stops Being the Engine
Google Search booked 63.4 billion dollars in the third quarter and grew 12%. Meta booked 63.06 billion and grew 26%. The two businesses are level on size, not on speed: the search click has stopped growing at the pace of the ad market.
The Two Numbers, Side by Side
Both results landed a week apart and, read together, tell a story neither one tells alone. Meta reported on October 21, Alphabet on October 28, and the comparison that puts the two figures next to each other is from November 1. The point is not size, because size is nearly identical: it is acceleration.
Meta: $63.06 Billion, Up 26%
Meta's advertising brought in 63.06 billion in the quarter ended September 30, up 26% year over year. The growth did not come from one lever: ad impressions rose 10% and the average price per ad climbed 14%. Both moved at once, which is the healthy way to grow when inventory is not infinite.
Google Search & Other: $63.4 Billion, Up 12%
Alphabet's Search & Other unit booked 63.4 billion, up 12%. It is still ahead by a margin of 340 million, which is to say a dead heat. Twelve percent is not a decline: it is half the pace of the direct competitor chasing the same ad budgets.
Read also
The Widest Gap Since 2021
Fourteen points between the two growth rates, Meta's widest advantage since 2021 according to CNBC's analysis. In recent years that gap had swung both ways, which is why it matters: this is not an odd quarter, it is the sharpest point of a trend.
Why AI Search Earns Less per Answer
Less Ad Surface per Query, Not Fewer Queries
A classic results page offers ten links and several ad slots. A generated answer offers a paragraph and, with luck, a couple of ads. People keep searching; what changed is how much commercial space fits into each query. Google admitted as much in its own quarter: AI search monetises less per query than link search.
Google's Answer: More Ads Inside Generated Responses
The company did not sit still. Its fix is to place more ads inside generated answers, which lifts revenue per query at the cost of looking more and more like the format it claimed to be replacing. A reasonable answer for the quarter, a trust problem for the year.
Read also
The 12% Stock Drop and the $500 Billion of Context
On October 29 Alphabet shares fell more than 12%, from 3.97 to 3.46 trillion dollars in market value: roughly 500 billion gone in a day, per the BBC. The punishment came from the full quarter, with revenue of 127.6 billion (up 25%) but profit below expectations. We already covered that session in detail: nearly a trillion dollars in two days. Here the drop is only context: what the market punished was the relative slowdown in search, not the spending.
Why Meta Is Actually Turning AI Into Money
A Feed Doesn't Hand Out Attention, It Keeps It
A search engine sends the user away. A feed keeps them inside. Meta's entire ad economy depends on the next piece of content sitting on the same screen, and generative AI slots in without friction: it fills gaps, produces variants and holds attention. When the same company decides which third-party content it uses and on what terms, it also sets the floor of the split; that is what our post on Meta's crawling clause is about.
Impressions Up 10%, Average Price per Ad Up 14%
When both numbers rise together, advertisers are paying more for more inventory and not complaining. That signals the targeting system improved enough to justify the price. Meta's AI does not appear in the headline of the earnings release: it appears in that average price per ad line.
A Revenue Model That Doesn't Need the Outbound Click
Here is the structural difference. Meta charges per impression and per result inside its own house. Google built twenty years of business charging to send people somewhere else. The moment a generated answer replaces the link, that model loses surface while the other gains it.
The Nuance Almost Nobody Gives: AI Still Discovers Very Little
The 3.3% of US Online Discovery Time
Analyst Nate Elliott of EMARKETER put the number on the table: AI assistants account for roughly 3.3% of the time people spend discovering things online in the United States. Small in volume, and precisely the part that is growing. Dismissing the shift on size is as reckless as acting as if search were already dead.
AI Ad Spend: $32 Billion in 2026, and 80% of It Outside Chatbots
EMARKETER forecasts about 32.03 billion dollars of AI-related ad spend in the US in 2026, nearly triple the year before, and more than 80% of it sits alongside AI-generated content rather than inside chatbots. In plain terms: the money is moving to the format, not yet to the conversational assistant.
What to Measure on Your Site This Week
An industry figure says nothing about your own site. These four readings do, and none of them needs new tooling.
Impressions vs. Clicks in Search Console, by Query Type
If impressions rise and clicks do not, a generated answer is taking your slot. Split branded and non-branded queries before drawing conclusions. For the technical side of measurement, the post on Core Web Vitals with real data is still the starting point.
Referral Traffic From ChatGPT, Perplexity and Gemini
These referrers already show up in analytics. Group them as their own channel instead of leaving them scattered among referrals: that way you can see whether they grow, which pages they land on and whether they convert.
Branded vs. Non-Branded: Which Query Still Brings New People
Branded traffic survives any change of intermediary better, because the person was already looking for you. If your share of branded traffic grows while non-branded stalls, you are not growing: you are harvesting what you planted earlier.
The Click-Through Ratio, Not Just the Total
Total clicks mixes two different things: how many people see you and what share clicks. Measure the ratio. It is the indicator that warns first when the results page format changes.
What to Do, and What Not To
Don't: Rewrite the Whole Strategy Because One Quarter Came in Flat
One quarter is not a trend, and reacting to every report with a full change of course is the most expensive way to get nowhere. The gap has been building for years; it deserves decisions that hold up over several quarters.
Do: Build the Channel That Doesn't Depend on a Middleman
Email, community, direct search, brand. Any channel where the relationship is with you and not with the algorithm that places you. It is slower and it does not produce pretty charts in the first month.
Do: Treat Assistant Traffic as Its Own Channel, With Its Own Measurement
Assistants cite, summarise and sometimes link. If you are going to get visits from there, you need to know how many, from which pages and with what intent. An unmeasured channel is a channel you cannot defend in a meeting.
Don't: Measure Success Only in Visits
A visit from a generated answer is worth something different from a visit from a query with clear intent. If the goal was that someone read, the visit still counts; if it was that they subscribed, measure that.
Conclusion
Search did not vanish: it booked 63.4 billion and grew 12%. What changed is that it stopped outgrowing the rest of the ad market, and that detail decides the split for the coming years. A site that lives on one channel is betting that channel wins the race; this quarter's numbers say it is not winning it. If you want to see how this is measured from the inside, we keep publishing every week with both industry and first-party data.
