Before AI Overviews launched in May 2024, Define Media Group’s portfolio of major U.S. publishers averaged 1.7 billion organic search clicks per quarter. Steady. Predictable. The kind of number you build a business model around and then stop thinking about it because why would you?
After the launch, traffic dropped by 16% and never recovered. When Google expanded AI Overviews in May 2025, the decline accelerated. By Q4 2025, organic search traffic in this portfolio fell 42% from its pre-AIO baseline.
Nearly half of organic traffic has disappeared from a portfolio large enough to set the tone for the entire publishing industry.
The traffic deal (you produce content, Google sends clicks, advertising revenue finances the next round of production) has been the economic engine of the open web for 20 years. This engine is clearly stalling, and the industry is responding by arguing about which dashboard to stare at in this case.
New interface, same deception
The first camp did what the SEO industry always does when the situation changes: They developed new tools to measure the shocks.
Prompt follow-up. LLM visibility dashboards. Share of response metrics. In less than 18 months, an entire category of vendors has emerged that sells you a number that tells you how often your brand appears in AI-generated responses. It’s Search Console for the chatbot age, and it brings with it the same reassuring conclusion: if the numbers go up, you win. If it goes down, buy more of what makes it go up.
I’ve written about this before, and I’ll say it again clearly: These tools are selling you nonsense with a confidence interval drawn on it in crayon. When a dashboard tells you that your brand “appeared in 73% of relevant AI responses,” this is what was actually measured: We sent some prompts to an API, received some output, and counted mentions. This is not a ranking. This is a lottery ticket.
The engineers who built these models cannot fully explain why a particular edition appeared. But sure, a SaaS tool on Mount Dunning-Kruger with a trend line has it all figured out.
The industry continues to buy because the alternative is to admit we are flying blind. Challenging the data means telling the room that the “directional charts” in the client deck are noise disguised as insight. Nobody wants to be that person. So the providers keep selling, the dashboards keep flickering and the number doesn’t necessarily correlate with sales. It just needs to fluctuate enough to sustain a subscription.
Jono Alderson laid out a more comprehensive version of this argument in a recent article: Clicks don’t count (and they never did). His point: SEO has always measured the interface and not the forces beneath it. Rankings, traffic, visibility values. None of these were a measure of competitiveness. These were measurements of a presentation layer. We spent two decades optimizing what we could see and calling it strategy.
He’s right. And prompt tracking is the latest iteration of the same mistake. Old visibility in a trench coat that pretends to practice two disciplines.
The second camp is intellectually more serious. Jono’s post is the best version of this argument, and I agree with more of it than I’m about to make it sound.
His framework: Stop measuring interface, start measuring competitiveness. Six structural dimensions derived from marketing science and validated over decades: experience integrity, physical availability, mental availability, uniqueness, reputation, commercial proof. AI systems collect signals about brands across the web, not isolated pages. The companies that are truly competitive are recommended and displayed. Visibility is the output, not the input.
I think that’s broadly true. I also think there is a timing problem the size of a crater.
These six dimensions operate on time scales of years. Building intellectual availability is a sustainable brand investment. Gaining reputation signals is the compound interest that comes from consistently not being bad. Strengthening its special virtues requires the approval of people who have never heard of Ehrenberg-Bass and won’t read a blog post to find out.
The gridlock happens quarterly.
Tell it to a publisher that just lost 42% of its search traffic to “strengthen structural competitiveness” and watch their face. It’s like telling someone whose house is flooding to invest in better drainage. You’re not wrong. You’re just not helping.
To his credit, Jono knows this. When someone asked in his comments how to implement the framework, his answer was honest: Redefine SEO to dominate these areas, or manage the organizational policy of collaboration with the teams that take on these areas. “A lot of organizational politics, one way or another.” That’s the kind of understatement only someone who’s actually tried it would make.
What actually broke
The measurement debate is a sideshow. The transport agreement was not a benchmark. It was the economic basis of content production on the open web.
Google needed content to crawl. Publishers needed distribution to make money. Produce something worth indexing, Google sends traffic, you convert it into revenue, that revenue funds more content. The loop ran for 20 years. Everyone acted like it was a partnership rather than a dependency, and that pretense continued because the numbers worked.
AI overviews break the vicious circle. Google synthesizes the answer from your content and provides it directly. The user gets what he needs. Your content is consumed on the Google interface with Google ads and thus generates Google’s interaction metrics. You get a citation link that virtually no one clicks on and a positive feeling about “brand visibility.”
Google’s own VP of Product for Search, Robby Stein, recently described how they had to “teach the model how to link.” Linking to publishers was not standard behavior. It had to be redesigned. The natural state of the system is to absorb your content and answer the question. Directing traffic your way is the afterthought they set out to do, so the extraction doesn’t look like what it actually is: taking your stuff and serving it as theirs.
The break is not uniform. Define’s data shows that traffic for breaking news across all Google surfaces increased by 103%, while always breaking content fell by 40%. The Top Stories carousel has been largely protected from AI Overview intrusion. This is not the case with evergreen content. The how-to guides, the explanations, the reference material, and the content categories that built the SEO industry are the very categories that AI Overviews are designed to absorb and replace.
Google chooses which content survives the transition. Time-sensitive content still gets clicks because you can’t summarize something that’s still in development. Everything else is increasingly raw material for the answering machine, and the machine does not pay for raw materials.
If “competitiveness” replaces traffic as an operational metric, the scope of SEO must change. Jono’s six dimensions mainly include product, brand and marketing. Experience integrity is product and UX. Mental availability is a brand investment. Reputation means not cutting corners for years. Commercial proof depends on whether the thing you are selling is actually good. SEO teams control technical discoverability, content strategy and website architecture. This is one level of the competitiveness framework, not the entire building.
So the discipline either expands into a cross-functional strategic role (good luck telling the CMO that SEO now owns the brand strategy because retrieval models have changed), or it honestly comes together and positions itself as the technical infrastructure that makes competitiveness machine-readable. Both options are better than “We’ll get you more organic traffic,” a promise that gets worse every quarter.
Clicks may not have been the right metric. Jono makes a convincing argument. We measured the interface and called it a system.
But Clicks paid the bills. They funded editorial teams, justified content investments, and supported the publishing ecosystem that both search engines and AI systems rely on for training data and retrieval sources. Without content to crawl, there is nothing to index. Without content to train, there is nothing to synthesize. The irony is apparently lost when the company uses AI overviews.
Nobody develops a transition strategy. The prompt tracking providers sell the new dashboard. The strategists sell the long-term view. Google doesn’t help. They’ve broken the deal, and their Discover push suggests they’d rather build a distribution surface that they completely control than fix one that shares value with publishers. The AI companies need content to exist, but haven’t yet figured out how to finance its production.
Everyone has a framework. Nobody has an answer.
The clicks didn’t count. But something has to. Soon.
Additional resources:
This post was originally published on The Inference.
Featured Image: Accogliente Design/Shutterstock
Follow us on Facebook | Twitter | YouTube
WPAP (907)
