Google AI Opt-Out Test: Why Organic Clicks Dropped
A small experiment suggests that leaving Google’s AI answers may cost you regular clicks too. Here is what the test found, what it does not prove, and what you should do next.
By Snehal Singh | Published: 2 November, 2026
AI Overview Summary
SEO consultant Cyrus Shepard ran a small test in which a website opted out of Google’s AI search features, which include AI Overviews and AI Mode. Search Engine Land reported three results. AI impressions fell to zero, the estimated click-through rate from AI features was only 0.9%, and clicks from regular organic results also dropped. Shepard calls this a possible “Search SERP Halo Effect,” where appearing in both AI answers and normal listings builds trust and earns more clicks. The test was small and key details were not shared, so it is a clue and not proof. The safe move is to measure total clicks, not only AI referrals, before you change any settings.
Table of Contents
- What Happened in the Test
- The “Halo Effect” Explained Simply
- Why AI Referral Traffic Alone Can Mislead You
- What the Test Does Not Prove
- What Other Studies Say
- A Simple Check Before You Decide
- What You Should Do Next
- Mistakes That Look Like Strategy
- How Ad2Connect Approaches AI Search Measurement
- Key Takeaways
- FAQ: Google AI Opt-Out
What Happened in the Test
Google gives website owners a way to opt out of its AI search features, such as AI Overviews and AI Mode, and many site owners have wondered what would actually happen to their traffic if they used it.
Cyrus Shepard shared a small case study on LinkedIn, which Search Engine Land later covered, in which a website opted out and he then compared the numbers from before and after the change.
He shared three findings: first, AI impressions fell to zero straight away; second, the estimated click-through rate from AI features was 0.9%, based on how overall clicks changed; and third, which was the most surprising, clicks from regular organic results also went down once the AI features were turned off.
The “Halo Effect” Explained Simply
Shepard called this a possible “Search SERP Halo Effect,” which in plain words means that being seen in more than one place on the results page may build trust with the people searching.
Think of a shop that appears on a busy street and also in a trusted local guide, because people who notice both places are more likely to walk in.
In the same way, a brand that appears in the AI answer and in the normal list of links may look more reliable, and some people may then click the normal listing instead of the AI answer.
The site in the test also saw that AI impressions made up about 8% of its total impressions, which means the AI answers were only a small part of its overall visibility.
Why AI Referral Traffic Alone Can Mislead You
This is the most useful lesson from the test, and it is worth reading slowly and carefully.
Many marketers judge AI search by looking only at the visits that arrive directly from AI answers, and because that number is usually small, such as the 0.9% estimate in this test, they decide that AI visibility is not worth the effort, but the test hints that this view may be too narrow, since being cited in an AI answer could also lift clicks on your regular listings further down the page, which means the value of an AI citation may not show up under the label of AI traffic at all and may instead be hidden inside your ordinary organic numbers, so a team that measures only direct AI visits could easily undercount the benefit, and might even remove itself from AI features, as the test site did, only to see its normal clicks fall as well.
So the real question is not how many people clicked the AI answer, but what happened to all of our clicks when the AI features were switched on, compared with when they were switched off.
What the Test Does Not Prove
It is important to stay honest here, because Shepard himself described the test as small and warned that results may vary from one website to another.
He did not share how much traffic the site receives, when the test took place, how large the drop in organic clicks was, or which method the site used to opt out.
Without those details, it is difficult to draw firm conclusions, since other factors could have changed during the same period, such as seasonal demand or ranking movements.
So treat this as an interesting clue and a reason to measure carefully, rather than a rule to follow blindly.
What Other Studies Say
Other research points in a different direction, because earlier studies covered by Search Engine Land found that AI Overviews can reduce clicks to regular listings.
For example, Ahrefs found a 34.5% drop in clicks to the top organic result when an AI Overview appeared, based on an analysis of 300,000 keywords.
These findings do not cancel each other out, but they do show that the effect of AI answers can change depending on the website, the topic and the type of search.
That is exactly why one small test should not make your decision for you.
A Simple Check Before You Decide
| What to Check | Where to Look | What It Tells You |
|---|---|---|
| How much of your visibility comes from AI features | AI impressions compared with total impressions | Whether AI is a small or large part of your reach |
| Total organic clicks over time | Google Search Console, compared before and after any change | Whether your overall clicks rise or fall |
| Clicks on your regular listings | Search Console performance report for web results | Whether a possible halo effect exists for your site |
| Brand searches and direct visits | Branded queries and homepage visits | Whether people remember and look for your name |
What You Should Do Next
You do not have to guess, because a few calm and practical steps will help you protect your traffic while you learn more.
Do Not Rush to Opt Out
One small test is not enough reason to change a setting that affects your visibility, and in this case the opt-out was linked to fewer clicks and not more.
If you are thinking about leaving AI features, first gather your own data so that you clearly understand what you might lose.
Measure Total Clicks, Not Only AI Clicks
Look at your overall organic clicks in Google Search Console, and compare comparable periods carefully, because a benefit that hides inside normal organic numbers is surprisingly easy to overlook.
This approach also protects you from judging AI search only by its relatively small amount of direct traffic.
Aim to Appear in Both Places
The idea behind the halo effect is simple: be present in the AI answer and in the regular results, by writing clear, trustworthy pages that deserve to be cited and also deserve to rank.
Helpful content, original facts and a clearly recognisable brand name make both outcomes more likely over time.
Test Carefully and Write Everything Down
If you do run a test, note the dates, the pages involved, the traffic level and the method you used, which are exactly the details missing from this case study, because good records turn a rough guess into evidence that you and your team can trust.
Mistakes That Look Like Strategy
These habits seem sensible in a quick meeting, but they can cost you real traffic.
- Judging AI search only by the small number of direct clicks it sends.
- Opting out of AI features because of anxiety, without investigating the potential impact first.
- Treating one small case study as a rule for every website.
- Changing settings without recording dates, pages and starting numbers.
- Ignoring brand searches and direct visits when you review results.
- Assuming that AI features and regular listings always compete, when they may actually support each other.
Good decisions come from your own numbers, read carefully over a long enough period of time.
How Ad2Connect Approaches AI Search Measurement
At Ad2Connect, a boutique digital marketing agency in Malad, we look at the whole picture before recommending any change. We never judge AI search by one small number.
First, we compare total clicks, impressions and brand searches before and after any change, so that a hidden drop does not surprise you. This is the careful approach you should expect from the best SEO agency in Mumbai, because good SEO depends on clear measurement.
Second, we connect search data to real business results, such as enquiries, leads and sales. As a leading performance marketing agency in Mumbai, we make sure that a rise or fall in clicks is judged by what it does for revenue, and not by clicks alone.
The aim is steady and sustainable growth that you can clearly explain to your team and repeat in the future.
Key Takeaways
- A small test found that opting out of AI features lowered regular clicks too. AI impressions fell to zero.
- The estimated click-through rate from AI features was 0.9%. Yet the value may be larger than direct AI visits show.
- The halo effect is a possible explanation. Being in both AI answers and normal listings may build trust.
- The test was small and missing key details. Treat it as a clue, not a rule.
- Measure total clicks before you change anything. Do not rely on AI referral traffic alone.
Ad2Connect is a digital marketing agency helping SaaS, eCommerce, D2C and local businesses grow through data-driven SEO, performance marketing and content strategy. If you are unsure how AI search is affecting your clicks, talk to our team about a search performance review.
