Google AI Overviews cut clicks to outside sites by about 40% in randomized test

RevScope, a revenue-analytics vendor, has published a review of the evidence on Google's AI Overviews and on GEO (generative engine optimization), the practice of getting content cited in AI-written answers. Its central piece of evidence is a randomized experiment by Saharsh Agarwal and Ananya Sen. They gave US desktop Chrome users an extension that put each user at random into one of two groups: one saw Google as usual, the other had AI Overviews hidden. The researchers then watched two weeks of real searches in January and February 2026.

On searches that would show an Overview, the group that could see them averaged 0.37 clicks to outside websites per search, against 0.62 for the group that couldn't. The authors report that as a 39.8% reduction. Searches ending with no outbound click went from 54% to 73%. Clicks on ads didn't change, which RevScope reads as a sign that Google takes care not to cannibalize its own revenue. Across all searches, with and without an Overview, the authors put the loss at roughly 18.5% of the organic clicks a site gets from Google, and they expect it to grow as Overviews appear on more searches. The effect was statistically significant for informational searches; the navigational and transactional samples were too small to settle anything. It is a working paper, not yet peer reviewed, and covers desktop only. A separate preregistered experiment by Wang and colleagues, with 1,100 people, also found that removing AI Overviews and AI Mode raised clicks to publishers.

Overviews are spreading fast. In DataForSEO's monthly panel of 20 million keywords, the share of results showing one climbed from 5.93% in September 2024 to 61.43% in August 2026. Over the same period featured snippets fell from 8.96% to 0.44%. Choosing different keywords doesn't reliably avoid them: rates rose across every search intent, and keywords with 201 to 300 monthly searches showed Overviews 58.59% of the time in the third quarter of 2026. By industry, health is at 69.68% and finance at 68.80%, while real estate is at 22.27%.

Observational data points the same way. Pew found people clicked an ordinary result on 8% of visits to a results page with a summary and on 15% without one, and ended their browsing session on 26% of visits against 16%. Pew's comparison isn't a controlled experiment, because searches that draw a summary tend to be longer and phrased as questions. Pew also saw people click a link inside a summary on just 1% of visits, against 8% for ordinary results. Across more than 2,000 websites in DataForSEO's panel, the desktop click-through rate of the first organic result fell from 17.83% in September 2025 to 7.43% in August 2026, a drop of 58%; mobile fell from 25.17% to 20.80%, a drop of 17%. The report itself says its data "does not establish" whether AI answers, layout changes or changes in behavior are the cause. Desktop fell from 15.18% to 7.94% between February and March 2026, a month when Overviews were almost flat (42.79% then 43.38%), and the unique desktop queries counted fell from 1.97 million to 1.32 million over the year. DataForSEO's model also estimates that the first organic result holds 9 to 16 points less click share on pages with an Overview, though these are different searches on a single day's snapshot.

Google disputes that search traffic to websites is declining. In August 2025 its head of Search, Liz Reid, wrote that total organic clicks were "relatively stable year-over-year" and that Google was sending "slightly more quality clicks", meaning clicks where people don't quickly come back. She gave no figures to back that up.

On who gets cited, the report compares AI Overviews with ChatGPT. Reddit and Wikipedia rank near the top of both, and 16 of the 30 domains in each top 30 appear on both lists by RevScope's count. YouTube is first in AI Overviews and nineteenth in ChatGPT. ChatGPT gave no source for 27.3% of its answers, against 0.7% for AI Overviews. Referring domains correlate moderately with AI citations (Spearman 0.406 for AI Overviews, 0.377 for ChatGPT). Ahrefs found 76% of pages cited in AI Overviews ranked in the top 10 for that query in July 2025, and about 38% in March 2026, though the two studies differ in sample, parsing and Gemini model. RevScope's verdict on GEO: it looks mostly like SEO plus two additions, being present on the third-party sites AI answers draw from and writing pages that are easy to quote. It found little evidence for any other tactic, and in March AI assistants sent websites just 0.3% of their traffic.

Key facts

  • In a randomized experiment (Agarwal and Sen, US desktop Chrome users, January and February 2026), searches that would show an AI Overview averaged 0.37 outbound clicks per search with Overviews visible against 0.62 with them hidden, reported as a 39.8% reduction.
  • The 40% figure applies to searches that show an Overview. Across all searches the authors estimate a loss of roughly 18.5% of organic Google clicks to a site.
  • Searches with no outbound click rose from 54% to 73%, while clicks on ads did not change.
  • In DataForSEO's panel, the share of results with an AI Overview rose from 5.93% in September 2024 to 61.43% in August 2026; desktop CTR of the first organic result fell from 17.83% to 7.43% between September 2025 and August 2026, though the report does not establish the cause.
  • The study is an unreviewed working paper limited to desktop, and the review comes from a vendor that sells a product showing revenue per visitor from Google.

Why it matters

Most evidence on AI Overviews and traffic is observational, so cause is hard to separate from other changes. A randomized design, where some users simply don't see Overviews, is the cleaner way to ask whether the summaries take clicks. Here it points the same way as Pew's browsing data. The context is rapid spread: Overviews now appear on more than half the US keywords DataForSEO tracks, up from 6% two years ago. Google, by contrast, disputes that search traffic to websites is declining.

Who it affects

Publishers, site owners and SEO and GEO practitioners who depend on Google clicks. The effect was statistically significant for informational searches. Navigational and transactional samples were too small to say anything. Health and finance see Overviews on roughly 69% of results, while real estate sits at 22.27%, so exposure varies by industry.

How to use it

RevScope's own advice is that SEO still decides whether you can be found, but where an AI answer appears, a ranking brings about 40% fewer visits than it used to. For GEO it points to two additions to ordinary SEO: being present on the third-party sites AI answers draw from, and writing pages that are easy to quote. It also suggests checking what your Google clicks are worth, and offers its own product, which shows revenue per visitor from Google, for that. Google's stated position is that no additional requirements or special optimizations are necessary to appear in AI Overviews.

How solid is it

The experiment is randomized, which is the most reliable kind of study for separating cause from coincidence, but it is a working paper, not yet peer reviewed, and covers desktop only. The Agarwal and Sen figure of 39.8% is as reported by the authors. The DataForSEO click-through data is weaker: the report does not establish the cause of the fall, desktop CTR dropped sharply between February and March 2026 while Overview share was almost flat, and the counted desktop queries shrank from 1.97 million to 1.32 million. The article's suggestion that a change in Overview format or quality might explain part of the drop is, in its own words, speculation. Pew's comparison is not a controlled experiment. The publisher sells a related product.

Risks and caveats

The 40% figure is not a 40% drop in total Google search traffic to websites: it covers clicks to outside websites on searches that would show an Overview, desktop only, and the authors' all-search estimate is roughly 18.5%. Mobile users were not covered. No timescale is given for when the 18.5% loss might grow or by how much. Google's response to this specific study is not given. Link-profile correlations with AI citations are moderate and don't show that links cause citations, and the citation data rests on prompts whose selection the report doesn't explain. The Ahrefs fall from 76% to about 38% of cited pages ranking in the top 10 is not a clean measurement, since the studies differ in sample, parsing and Gemini model.

“relatively stable year-over-year”

— Liz Reid, Google's head of Search, August 2025, on total organic clicks