Google AI Overviews correlate with a 58 percent reduction in click-through rate on top-ranking organic pages, according to a February 2026 Ahrefs study, nearly double the 34.5 percent decline recorded less than a year earlier. But that decline isn't evenly distributed. Branded queries with an AI Overview present actually saw an 18 percent increase in click-through rate over the same period, which is the clearest evidence yet that AI Overviews are redistributing clicks toward already-recognized brands rather than suppressing traffic uniformly across the board.
That distinction, between "traffic is disappearing" and "traffic is being redistributed toward brands people already recognize," changes what a content team should actually do about it. Chasing the old strategy (rank first, hope for a click) is fighting a trend that a growing body of evidence says is structural, not temporary. The more useful question is where the surviving traffic and citation opportunity is actually going.
How bad is the traffic decline, really?
Bad enough that it's showing up at the level of overall publisher traffic, not just individual query CTR. Google search traffic to publishers fell 33 percent globally in the year to November 2025, with U.S. publishers hit harder at 38 percent. A 2026 randomized field experiment, the first to establish causal (not just correlational) evidence on this, found a 39.8 percent drop in outbound clicks specifically attributable to the presence of an AI Overview. Pew Research Center data adds another angle: only 8 percent of users click a traditional search result when an AI Overview is present, compared to 15 percent when no overview appears, essentially cutting click-through roughly in half at the point of presentation.
Is the decline the same for every kind of site?
No, and the size gap is significant. Chartbeat's data shows smaller publishers lost 60 percent of search referral traffic over two years, compared to a 22 percent decline for large publishers over the same period. Larger, more established domains are weathering this shift meaningfully better than smaller ones, which tracks with the branded-query finding above: recognition and existing authority appear to buffer against the AI Overview effect, while smaller or newer sites absorb more of the loss.
What does the branded-query data actually tell us?
That being the source an AI Overview cites, and being a brand the user already recognizes by name, is worth more now than it used to be. An 18 percent CTR increase on branded queries with an AI Overview present means that when a user searches a brand name directly and Google generates a summary, that summary appears to be driving more clicks to the source, not fewer. The mechanism is plausible: an AI Overview citing a known brand functions as a confirmation and a convenient link, rather than a full substitute for visiting the site, in a way a generic informational overview doesn't.
So what should a content strategy actually optimize for now?
A shift from "rank for the query" to two more specific goals that better match where the surviving traffic is:
| Old goal | Better-suited 2026 goal | Why |
|---|---|---|
| Rank #1 for a broad informational query | Be the cited source inside the AI Overview for that query | The click on the ranked result is worth less than it was; the citation itself carries brand-visibility value even without a click |
| Generic content optimized for search volume | Content and brand-building that makes people search your name directly | Branded query traffic is the category actually growing in this data |
| One-time SEO push | Ongoing presence across AI Overviews, ChatGPT, and Perplexity, tracked over time | Citation, not ranking alone, is now a meaningful chunk of the actual visibility a page delivers |
Does this mean SEO doesn't matter anymore?
No, but its payoff has changed shape. Ranking well is still often a prerequisite for being cited in an AI Overview, since Google draws heavily on already-ranking pages to build its summaries. What's changed is that ranking first no longer reliably converts to a click the way it used to, so the value of that ranking increasingly comes from citation-driven brand visibility (being named, even without a click) rather than direct traffic alone. Teams that only measure success in clicks and sessions are now under-counting a real, if harder to measure, part of what their content is doing.
What should a team actually change about how they measure success?
Three practical adjustments:
- Start tracking citation appearances separately from click-through traffic. A page can be doing real work (getting cited, building brand recognition) while showing a declining CTR in Google Search Console, and treating that CTR decline as pure failure misses what's actually happening.
- Invest more deliberately in brand-recognition content, not just keyword-targeted informational content, given the branded-query CTR advantage. This includes things that build name recognition beyond your own site: PR, guest content on recognized publications, and consistent presence in industry conversations.
- Prioritize pages where a citation without a click is still valuable, and be more selective about ad-supported or pure pageview-dependent content, since that business model is under the most direct pressure from this shift.
Frequently asked questions
Is the 58 percent CTR drop the same across every industry? No, the study's headline number is an aggregate. Informational, non-branded queries in competitive categories tend to see the steepest declines, while branded and highly specific long-tail queries fare better, consistent with the redistribution pattern described above.
Should I stop investing in top-of-funnel informational content? Not entirely, but weigh it against the citation opportunity it offers rather than expected click volume alone. Informational content that's well-structured enough to become the cited source inside an AI Overview or an AI chatbot answer still delivers brand value, even if the raw click number looks worse than it did two years ago.
Does blocking Google-Extended stop my content from appearing in AI Overviews? No. AI Overviews draw on Google's standard search index, not the separate Google-Extended training crawler. Blocking Google-Extended affects Gemini training data, not AI Overview citation eligibility.
Is this trend expected to get worse? Publishers surveyed by Press Gazette expect traffic declines to continue, averaging 43 percent over the next three years, with about a fifth anticipating losses exceeding 75 percent. Whether that plays out exactly as projected, the direction of the trend is consistent across every data source cited here.
This traffic shift is also the clearest evidence for why citation, not just ranking, deserves its own tracking. See GEO vs. SEO: what's actually different for how that changes what a page is actually optimizing for, and how AI answer engines choose who to cite for the mechanism behind the citations that are replacing some of that lost traffic.
Next step
Pull your Search Console data for your ten highest-impression, lowest-CTR pages. Those are very likely the pages Google is already summarizing in an AI Overview instead of sending clicks to. Check whether your content is actually the one being cited in that summary, since that's now a meaningful measure of success even on a page whose click numbers look like they're failing.