The Traffic Disappeared (And Nobody Told You Why)
Google rolled out AI Overviews last year. They answered search queries directly in the results, no click required. You already knew that part. What you might not have noticed: your analytics dashboard got quieter. Fewer visits from organic search. But your rankings didn't drop. Your content is still there. So where did the traffic go?
It went to Google's answers. It never reached your site.
This isn't a measurement problem. It's an attribution problem. And it's breaking the foundation that every marketing operation in the world is built on.
Zero-Click Isn't New. But AI Made It Catastrophic.
Zero-click searches existed before. Google showed quick answers for recipes, weather, sports scores. A user searched, got their answer in the SERP, and moved on. Fair enough.
But AI Overviews aren't quick answers. They're comprehensive summaries pulled from multiple sources, sometimes entire articles condensed into a paragraph. They cite sources, sure. But the user gets the information without ever visiting those sources. They get answers about your company from your own content, displayed on Google's page, with your information attributed but no traffic sent your way.
A hotel marketer searches for "best oceanfront resorts in Maui." Google's AI pulls resort names, prices, reviews, all sourced from hotel websites. The searcher reads about your resort. They get everything they need. And your web analytics shows zero visits from that search. The query happened. Your content was used. But you registered nothing.
Your attribution model assumes: search query to click to visit. Google's AI Overviews created a new pattern: search query to AI answer (powered by your content) to no click. The query still happened. You still provided the answer. But your metrics don't see it.
This Is Happening at Scale
The data is getting harder to ignore.
Hospitality companies report click-through rates dropping 30-40% on branded searches as AI Overviews answer questions directly. Publishing platforms show 20-25% traffic reductions from organic search even though impressions haven't declined. E-commerce sites are seeing shopping searches skipped entirely, users read product comparisons from AI Overviews without ever entering the site.
Google's own internal acknowledgment came in December when they launched a "publisher partnership pilot." The pitch: if you're losing traffic to AI Overviews, Google will pay you. The language was indirect, they called it "direct publisher payments" to "offset zero-click risk." But what they were actually saying was: "We're answering your content without sending you traffic, and we know that's a problem. Here's some money."
That pilot was supposed to be temporary. Publishers are still waiting for permanent solutions.
Why Your Attribution Model Is Broken Now
Last-click attribution, the standard for organic search, assumes the final touchpoint before conversion is the only one that matters. You clicked a Google result, landed on the site, bought something. Google gets credit. Simple.
But when Google intercepts the interaction before the click ever happens, last-click attribution doesn't work anymore. The query happens. Your content gets cited. The user gets their answer. And last-click credits, nobody, because there was no click.
Some marketers think this is fine, they'll just look at branded search volume instead of clicks. But branded searches are only part of the picture. Informational queries (product research, comparisons, reviews) are where the real traffic used to come from. Those are the searches where AI Overviews do the most damage. A user searches "is [product] worth it" and gets a summary of reviews and specifications from your site without ever visiting. You can't measure that impact because they never arrived.
Others point to traffic directly from Google Search Console instead of GA4. Search Console shows impressions even without clicks. But impressions aren't conversions. They're not revenue. Knowing your content appeared in an AI Overview doesn't tell you if it drove business value.
The real problem: you've built your entire marketing operation (budget allocation, channel strategy, vendor contracts) on the assumption that search traffic equals clicks that land on your site. That model is collapsing. And you're not equipped to operate in a world where your content answers search queries without sending traffic.
The UK Just Forced Google's Hand. The US Hasn't.
In June, the UK's Competition and Markets Authority ordered Google to give publishers control over whether their content appears in AI Overviews. Publishers can now opt out entirely. They can demand proper attribution. They can require Google to show which content is being used.
Google's response was immediate, they complied. Within weeks, major UK publishers had opted out. Traffic started recovering.
The US has no equivalent requirement. The FTC is investigating, sure. But investigation isn't action. Meanwhile, Google is quietly rolling out AI Overviews in more search categories. International markets are getting versions that cite sources even less explicitly. The default is still "show the answer without the click."
This will eventually change. Either through regulation (forced opt-outs like the UK), or through market pressure (publishers and advertisers demanding change loudly enough). But until then, you're operating in an attribution environment that Google controls and that actively works against you.
What This Means for Your Attribution
If you're still measuring success by last-click search traffic, you're measuring less than half the picture now.
Brand queries that used to bring direct traffic are being intercepted. Product research queries that used to bring high-intent traffic are being answered without the visit. And you have no way to know how many sales were influenced by those unseen interactions.
Some marketers are trying to build workarounds. Tracking brand mentions in Google Search Console as a proxy for attribution. Using branded search volume as a leading indicator instead of click-through. Building custom audiences from "branded search impressions" to retarget users who searched for them but didn't click.
These are patches. They don't fix the core problem: the last-click model is dead. Google's AI killed it. And every marketing operation that depends on that model is now running on borrowed time.
The real fix comes in three parts.
First: demand transparency from Google about which content appears in AI Overviews and how often. This is happening in the UK now. Push for it everywhere else.
Second: build attribution models that account for zero-click answers. Track branded search volume, impressions in AI Overviews, and compare them to trailing conversion data. It's messier than last-click, but it's honest.
Third: diversify away from search. If Google is taking the traffic, distribution diversification matters more now. Direct channels, email, audience platforms, owned media, these become less optional.
The attribution model that made marketing predictable is gone. Google replaced it with answers that Google controls. You don't get to opt out. You don't get to set rules. You get to adjust and survive until the regulation catches up.
This changes your budget conversations.