AI Overviews Click Studies, Oct-Nov 2026: What They Found

The studies published in October 2026 show AI Overviews on a large majority of tracked keywords, but none of them measures how many clicks your site loses. Semrush found coverage of 88 percent for 400 commercial keywords, SE Ranking found 82.5 percent on US brand searches, and a one-site opt-out test by Cyrus Shepard estimated a 0.9 percent click-through rate. Each has limits. (As of 1 November 2026, I found no newer study.)

Key Takeaways

• Semrush's 21 October study of 400 commercial keywords reports AI Overview coverage rising from 45.75 percent in September 2025 to 88 percent in August 2026, and 97.08 percent when weighted by search volume (Semrush).
• SE Ranking's 22 October study found AI Overviews on 82.5 percent of 1,000 US brand keywords and 40.4 percent of 1,000 German ones, checked three times on desktop in October (SE Visible).
• The "over 80 percent of searches" headline is Search Engine Roundtable's summary of three vendors' data, not one study, and seoClarity's separate figure was 62 percent (Roundtable, 26 October).
• Cyrus Shepard's 0.9 percent CTR figure comes from one website and is an estimate derived from before-and-after data, not a direct measurement (Roundtable, 30 October).
• The strongest causal evidence is older: a randomized field experiment with data from January to February 2026 found AI Overviews cut outbound organic clicks by 39.8 percent.
• Google's 15 October link changes came with no click data, and a Google representative said click data is not yet in AI performance reports.

Coverage studies: how often AI Overviews appear

Coverage tells you the odds a query of yours triggers an AI Overview. It says nothing about clicks.

What stands out inside the Semrush data

The modifier types that look most commercial are almost saturated: "best X for Y" queries reached 96.12 percent in August 2026, "reviews" nearly 90 percent (up 59 points), and "alternatives" over 82 percent. Brand-versus-brand queries rose the least, to over 86 percent. Ecommerce rose the most, from 32.62 percent to 85.11 percent. Three industry and modifier pairs hit 100 percent. If you run comparison or alternatives pages, the SERP you are competing on probably has an AI Overview already. I covered how to structure those in AEO for comparison pages.

What stands out inside the SE Ranking data

The citation findings matter more to me than the coverage figure. AIOs cited an average of 2.57 brand-website pages versus 1.72 in the organic top 10, and 62.2 percent of US brand-site citations pointed to URLs outside the organic top 10. Homepages were 36.8 percent of citations in AIOs but 71.1 percent of brand-site URLs in the organic top 10. This fits the pattern in AI Overview citations decoupling from rankings. The study is a snapshot of presence and citation patterns, and its authors say it does not measure accuracy, traffic or conversions.

Why I would not repeat "80 percent of searches"

Roundtable's article says the number is drawn from Semrush, SE Ranking and seoClarity, and does not describe their methods or give collection dates beyond "recent." The only figure near 80 percent is SE Ranking's 82.5 percent for US brand searches, and the author writes that he assumes non-brand is higher. The seoClarity figure is lower at 62 percent. All three sources sell SEO software. Brand and commercial keyword sets are not all searches, so I would say "most of the commercial and brand keywords these vendors track" and no more.

Click studies: what they actually measured

Sources: Roundtable on Shepard, PPC Land on Agarwal and Sen, Seer Interactive.

The October test, read carefully

Roundtable calls Shepard's work a "super limited study." He used Google's generative AI opt-out controls on his own site, then compared impressions and clicks before, during and after. AI impressions dropped to zero immediately, and made up about 8 percent of impressions. Regular organic clicks also fell during the opt-out, and the URLs with the most AI visibility dropped and then recovered when visibility returned. The 0.9 percent is his subtraction of known AI impressions from overall totals. It is interesting because it hints that opting out does not protect classic clicks. It is not a benchmark. My existing guide to the opt-out control covers the decision.

The field experiment, which remains the best causal evidence

The Agarwal and Sen experiment is from before the window of this article, but it is the study I would cite if forced to pick one. Outbound organic clicks per search on AI Overview queries rose from 0.37 to 0.62 when the overview was hidden, and zero-click searches fell from 0.73 to 0.54. The effect was concentrated at the top of the page and in informational queries. Satisfaction measures did not differ. The authors do not estimate revenue, and the AI Mode arm had high attrition, so it is descriptive only.

The Seer data, which pushes the other way

Seer's update shows organic CTR on AIO queries recovering from the December 2025 floor. It also reports that being cited went with roughly 120 percent more organic clicks per impression than not being cited, still about 38 percent below queries without an AIO. Seer states the data is correlational and that higher-authority brands are more likely to be cited, so do not read it as "citations cause clicks."

Studies about AI referral traffic

A Roundtable roundup of 22 October summarizes three other studies: Marie Haynes (100 sites, 145 million sessions, ChatGPT sessions at 0.62 percent of Google organic volume), SEO Works (organic search delivered 39 times the traffic of AI assistants), and Kevin Indig (no reliable link between AI visibility and direct or branded traffic in a 12-site sample). I read these only through Roundtable's summary, and they concern AI assistants, not AI Overviews. For the conversion angle, see I tried to verify the 4.4x claim.

What changed in the interface

On 15 October Google said it had increased inline links in AI Mode and AI Overviews, was testing an official-links section, and was expanding news carousels that highlight Preferred Sources (Google). The post makes no click or traffic claim. On 29 October Brodie Clark reported a possible test removing the citations card from AI Overviews on desktop, with a sharp drop in how often 2 to 3 domains appeared; Google had not confirmed it and Roundtable could not reproduce it (Roundtable). Treat that as an unconfirmed test. For the mobile side, see AI Overviews expanding by default.

What this means for marketers and founders

1. Stop using one site-wide click-loss number. Use Search Console to split queries into those that trigger an AI Overview and those that do not, using a fixed sample of 50 to 100 tracked queries. My method for this is in measuring AI Overview traffic loss.
2. Do not act on Shepard's result alone. I would not run an opt-out on a revenue-critical site to replicate a one-site anecdote.
3. Check your comparison, alternatives and review pages first. Semrush's modifier data says those query types are almost always showing an overview.
4. Keep ranking work. The SE Ranking citation data shows most brand-site citations point outside the top 10, but the study also reports 98.3 percent domain-level overlap with the top 10, so domain strength still appears to matter.
5. Report impressions and citations, not clicks, from AI features until Google adds click data.

Caveats and what we do not know

• All three coverage figures are vendor datasets with undisclosed or partial methods. They agree on direction, not on a number.
• No October study measured the click effect of AI Overviews with a control group. The one that did was published in April.
• Google has said nothing about click rates for AI Overviews in this period, and its representative said click data is not yet in AI performance reports.
• I did not open Shepard's LinkedIn post or the Haynes, SEO Works and Indig studies directly.

For earlier benchmarks, see CTR benchmarks under AI Overviews and zero-click statistics for 2026.

FAQ

Do AI Overviews appear on most Google searches now?

For the keywords vendors track, mostly yes. Semrush reports 88 percent of 400 commercial keywords in August 2026 and SE Ranking reports 82.5 percent of US brand searches in October. These are not samples of all searches, so "most searches" overstates it.

How much do AI Overviews reduce clicks?

The best controlled estimate I found is a 39.8 percent drop in outbound organic clicks when an AI Overview appeared, from a randomized experiment using January and February 2026 data. It is not yet peer reviewed. Observational studies give different numbers.

What is the 0.9 percent click-through rate figure?

It is Cyrus Shepard's estimate for AI impressions on one site, derived by subtracting known AI impressions from total impressions during an opt-out test. Search Engine Roundtable called the study super limited. It is not a general benchmark.

Does being cited in an AI Overview bring more clicks?

Seer Interactive found cited brands had roughly 120 percent more organic clicks per impression than uncited ones, but still about 38 percent fewer than queries with no AI Overview. Seer says the data is correlational.

Did Google publish click data for AI Overviews?

Not in this period. Its 15 October post on links makes no click claim, and a Google representative told Search Engine Roundtable that click data is not yet in AI performance reports.

Should I block AI Overviews with Google's opt-out control?

Shepard's test suggests classic clicks fell for him during an opt-out, but it is one site. I would not block on current evidence, and I would only test on a section with low revenue exposure.

Is the 80 percent figure reliable?

Not as a single number. It is Roundtable's summary of three vendors' data, and seoClarity reported 62 percent. Use it as a sign that coverage is high, not as a precise rate.

Want help with this?

I'm Younus Fardeen, an organic growth, SEO and AEO strategist with 4+ years of marketing experience across edtech and startup teams. If you want help separating AI Overview effects from ordinary ranking changes in your Search Console data, see my work and reach me through the contact form at younusfardeen.com.