Should you opt out of AI Overviews? A small test by SEO Cyrus Shepard, reported on 30 October 2026, suggests caution. Blocking one site from Google's AI features cut its AI impressions to zero, and its regular organic clicks fell too. The test covered a single site and published no raw numbers, so enterprise teams should run their own controlled test before deciding.
Key facts: Shepard blocked one site from AI Overviews and AI Mode using Google's opt-out controls; AI impressions made up about 8% of the site's impressions, with an estimated 0.9% CTR; SE Ranking reports AI Overviews on 82.5% of US brand searches; the test's click drop, dates and raw figures were not published.
What did the opt-out test actually find?
The site lost more than its AI impressions: Shepard says ordinary organic clicks dropped as well. He used Google's generative AI opt-out controls to exclude one site from AI Overviews and AI Mode, then compared impressions, clicks and CTR across three periods, before, during and after the opt-out.
According to Search Engine Roundtable's report on the opt-out study, Shepard says AI impressions went to zero as soon as the opt-out took effect. His most surprising finding, in his words: "regular organic clicks ALSO DROPPED when AI features were disabled."
Other details he shared:
- AI impressions were about 8% of the site's total impressions.
- Overall impressions dropped during the opt-out.
- The estimated CTR of the site's AI impressions was about 0.9%.
- On the URLs with the most AI visibility, average position dropped during the opt-out, then rose when AI visibility returned.
- Sitewide average position increased, which Shepard attributes to a couple of new posts and brand search.
Shepard confirmed to Barry Schwartz that AI Overviews appeared on the results pages throughout. The site was cited in them before and after the test, but not during.
Why might losing AI citations reduce regular clicks?
Shepard's explanation is a "Search SERP Halo Effect": the more places a brand appears on a results page, the more likely a user clicks one of them. He suggests an AI citation may lift clicks anywhere on the page, not only on the cited link.
Greg Sterling of Near Media added that his firm's research has found AI Overviews create a form of brand lift that feeds later searches and clicks, though not consistently. Treat both as hypotheses. The sources do not rule out other causes such as seasonality, new content or algorithm volatility during the test window.
How much weight should you put on one site's result?
Very little on its own. Schwartz calls it a super limited study with a small sample, and Shepard himself calls it small. The published reports do not give the size of the click drop, the test length, the site's identity or the underlying data.
SERP Mentor's analysis of the study describes it as one site over a short period with no raw numbers published. Its advice is blunt: "don't opt out on a hunch."
Three more limits matter:
- The 0.9% is derived, not measured. Shepard reached it by subtracting known AI impressions from overall site figures.
- Confounders were present. New posts and brand search moved sitewide position during the window.
- AI Overviews never left the page. As Schwartz points out, the test shows what happens when you remove yourself from AI answers, not what happens when AI answers do not exist.
Are AI Overviews really on 80% of searches?
Some tools report coverage at or above that level, but the figures vary by tool and query set. Search Engine Roundtable's November 2026 webmaster report says AI Overviews are now appearing on 80% of Google searches. SERP Mentor breaks down the October 2026 tool data:
- Tool - Reported coverage - Scope
- SE Ranking - 82.5% - US brand searches
- SeoClarity - 62%+ - All tracked queries, about a month earlier
- SeoClarity by industry - 55% to 84% - Varies by sector
- Semrush - Chart only, rising - Not specified
Only SE Ranking's figure clears 80%, and it covers brand searches. Your exposure depends on your query mix, so measure it for your own keywords.
Doesn't other research show clicks rise without AI Overviews?
Yes, but those studies test a different change, so they do not contradict Shepard. They hide AI Overviews from users entirely. Shepard removed one site's citations while the AI answers stayed on the page.
In a randomized field experiment by Saharsh Agarwal and Ananya Sen, covered by Search Engine Journal's write-up of the 38% study, 1,065 US participants searched from January to February 2026. AI Overviews appeared on 42% of queries. Hiding them raised outbound clicks from 0.38 to 0.61 per search, and outbound organic clicks fell 38% on triggered queries. Top-of-page placements drove the effect. Satisfaction and ease of finding information were nearly identical with and without them. The sample was active desktop Chrome users, and the paper is an unreviewed draft.
A second preregistered experiment, the Wang et al. paper on publisher referrals, found that removing AI Overviews and AI Mode increased click-through to publishers, while an AI Mode-only experience reduced it. The authors note Google's position that AI features are complements that keep referral traffic steady.
Ahrefs' correlational study found an AI Overview correlates with a 58% lower average CTR for the top-ranking page, using a forecast baseline. Pew Research, cited by Search Engine Journal, found users click 8% of the time with AI Overviews versus 15% without. These figures measure different things and should not be read as one range. Google VP Liz Reid argues AI Overviews cut "bounce clicks," a claim Search Engine Journal says lacks supporting data.
The combined picture: AI Overviews may reduce clicks overall, yet being cited inside them may still beat being absent.
How can you run a controlled opt-out test of your own?
The following design is our recommendation, not guidance from the sources above.
- Use a holdout. Opt out a defined section of the site and keep a matched section as a control, rather than switching off the whole domain.
- Fix the window. Run before, during and after periods of equal length, and avoid launches or known algorithm updates.
- Split brand and non-brand queries. A halo effect should show up differently in each.
- Track citations, not just clicks. Record which queries cite you in AI answers and how often, using AI visibility tracking across engines.
- Watch rivals. If your citations vanish, someone else takes the slot. Competitor monitoring shows who gains.
- Plan for blind spots. A Search Engine Roundtable headline says Google "won't show click data" for links in AI Mode and AI Overviews, so AI-specific clicks may only be estimated.
What does this mean for enterprise SEO and brand teams?
For most enterprises, the evidence does not support opting out without proof. Clicks alone tell only part of the story; you also need to know whether you are cited, how you are described, and who replaces you when you step aside.
MarkGrid builds measurement for that gap. Model Share tracks how often your brand is mentioned or recommended by ChatGPT, Gemini, Perplexity, Claude and Microsoft Copilot compared with competitors, at query level, and flags visibility gaps. Competitive Intel tracks competitors' SEO and AI-search visibility with movement alerts, useful during a holdout test.
Citation counts matter less if answers describe you wrongly. Brand Research tracks how AI models describe your products, services and regions, which matters for accuracy in regulated sectors such as fintech and healthcare. And because brand lift may surface in conversation before it shows in Search Console, Community Signals monitors Reddit, Discord, Quora and niche forums for brand mentions and buyer intent. Together, these give leadership a fuller record than clicks alone.
FAQ
Can I opt out of AI Overviews?
Yes. Google added a Search Console control that lets a site exclude its content from AI Overviews, AI Mode and Discover's AI features, according to SERP Mentor. Shepard used Google's opt-out controls in his test.
Did opting out increase organic clicks?
Not in Shepard's test. He says regular organic clicks dropped while the site was excluded, though the size of the drop was not published.
What does the 0.9% CTR mean?
It is Shepard's estimate for his site's AI impressions, calculated by subtraction. It is not a measured CTR and not a general figure for Google's AI features.
Why do randomized studies show clicks rising without AI Overviews?
They hide AI Overviews from users entirely, which is a different change from removing one site's citations. Both findings can hold at once.
Is the 80% coverage figure reliable?
It varies by tool. SE Ranking reports 82.5% of US brand searches, while seoClarity reported 62%+ of all tracked queries about a month earlier.
Should my company opt out?
Not without your own controlled test. Use a holdout section, separate brand from non-brand queries, and track citations and competitor visibility alongside clicks.
Last updated 6 November 2026.
How often MarkGrid is named when AI models discuss this topic. About Model Share
Parteek chauhan
Growth Strategist
Parteek Chauhan is a Growth Strategist at Markgrid, turning market insights, audience behaviour and performance data into focused strategies that help brands identify and scale meaningful growth opportunities.
See it on your brand
Thirty minutes on where marketing is leaking for you.