Zero-click search: Google is answering the question, so fewer people click

The short answer

Zero-click search means the searcher got what they needed without visiting a site. It is now the majority of Google searches in the US: 68.01% in January to April 2026, according to SparkToro. To appear in AI Overviews, Google says a page only needs to be indexed and snippet-eligible. The commercial response is to make every remaining visit count.

What is zero-click search?

Zero-click search is a search that ends on the results page, with no click to any website. The answer might come from an AI Overview, a featured snippet, a map or a knowledge panel. It is not new, but AI summaries have made it far more common on informational questions.

SparkToro's 2026 study found fewer than one in three US Google searches now sends a click anywhere. SparkToro puts the 2024 comparison figure at 60.45%. Its original 2024 study, on a different data provider, reported 58.5% zero-click in the US and 59.7% in the EU. SparkToro itself calls the comparison "a bit of apples and oranges", because the 2026 figure uses Similarweb's panel and the 2024 figure used Datos. Both exclude the Google app, so the true rate is likely higher, and both are US and EU only.

What has Google actually shipped?

Google has shipped two different things. AI Overviews are the generated answer above the normal results. AI Mode is a separate conversational search experience, which rolled out in Australia in English in October 2025. Both draw on pages that are already in Google's index.

Google rolled AI Mode out in Australia in English on 8 October 2025 (Google Australia blog). So this is no longer a United States story that might reach us later.

Google's documentation for site owners is short and worth reading in full, because it contradicts a large amount of what is being sold. To appear as a supporting link, "a page must be indexed and eligible to be shown in Google Search with a snippet", and "there are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary" (Google Search Central).

The controls are the ordinary ones. nosnippet, data-nosnippet, max-snippet and noindex limit what appears in AI features exactly as they limit snippets. Google-Extended is separate and governs training and grounding in some of Google's other systems (same source).

That is the whole documented surface. If an agency is quoting you for AI Overview optimisation as a distinct service, ask them which documented lever they are pulling.

How do you rank in AI Overviews?

To appear in AI Overviews, a page must be indexed and eligible to show a snippet; Google says nothing more is required. Beyond that, the pages that get cited tend to answer the question directly, early and in plain sentences, and to be well regarded on the wider web. There is no separate switch to flip.

That makes it ordinary good SEO done properly. Answer the question in the first paragraph. Give each sub-question its own heading and a direct answer. Use real numbers with sources. Make sure search and answer-engine crawlers are not blocked. The practical version for marketing teams is our guide to generative engine optimisation, and the ChatGPT-specific checklist is in how to get your business cited by ChatGPT.

Being cited appears to help. Seer Interactive found brands cited in an AI Overview earned 35% more organic clicks than those not cited. Seer states it cannot prove the citation caused the difference, and its sample is its own agency clients, so treat it as a correlation.

To find out whether ChatGPT and Google's AI features mention you today, work through how to check your AI visibility, or run the ten buyer questions AI visibility check by hand in an hour.

What does independent measurement show?

Independent measurement shows that a Google search with an AI summary produces roughly half the clicks to websites of one without. The most careful study, from Pew, found 8% against 15% among US adults in March 2025. The direction is consistent across studies; the size is not.

The most careful public study is from the Pew Research Center, published 22 July 2025 (Pew Research Center).

Pew tracked real browsing for about 900 United States adults on an opt-in panel, covering 68,879 Google searches made in March 2025, of which 12,593 produced an AI summary. The findings:

• Users clicked a traditional search result in 8% of searches where an AI summary appeared, against 15% where one did not.
• Users clicked a link inside the summary itself in 1% of visits.
• Users ended their browsing session after 26% of pages with a summary, against 16% without.

What this is: a real behavioural measurement rather than a survey of opinions, with the methodology published.

What it is not: Australian, current, or generalisable to your category. It is United States adults, an opt-in panel, one month in early 2025. Pew itself notes that the summary shown for a given search changes over time, so the results reflect how search looked when it was collected. AI features have expanded considerably since.

What does Google say, and what has it not published?

Google says total organic clicks to websites have been relatively stable year on year, and that click quality has risen. It has published no data to support either claim.

In the same post Google's head of search added that "average click quality has increased and we're actually sending slightly more quality clicks to websites than a year ago" (Google blog).

Take both halves of that seriously.

The claim is not obviously false. Total volume across the whole web can hold steady while it redistributes, and Google says as much in the same post: traffic is moving between sites rather than only disappearing. A total that is flat is entirely compatible with your own site being down.

The problem is that Google published no data with it. No chart, no time period, no definition of a quality click that anyone outside Google could test. So it is a statement from an interested party that cannot be checked, which is not the same as a lie and is also not evidence.

Why do the zero-click numbers disagree?

The numbers disagree because each study measures something different: a browsing panel, aggregated Search Console data, one agency's clients, or a clickstream panel. They also compare different periods and query types. Read the method before the headline.

Ahrefs, which sells search tools and therefore has its own interest, published an updated study on 4 February 2026 finding a 58% reduction in click-through rate for the top-ranking page when an AI Overview is present (Ahrefs). It used 300,000 keywords and aggregated Search Console data, comparing December 2023 with December 2025.

Read the underlying figures rather than the headline, because they are more interesting than it is.

For keywords without an AI Overview, average position-one click-through rate fell from 0.076 to 0.039. For keywords with one, it fell from 0.073 to 0.016.

So click-through rate roughly halved on the control group too, where there was no AI Overview at all. Seer found the same thing independently: organic click-through on queries without an AI Overview fell 41%, from 2.74% to 1.62%. Whatever is depressing clicks is not only AI Overviews. The Ahrefs headline of 58% comes from comparing the observed 0.016 against a projected 0.037, which is a modelled counterfactual rather than a direct measurement. That is a reasonable method. It is not the same kind of fact as Pew's 8 against 15.

Seer also found paid click-through on the same informational queries fell from 19.7% to 6.34%. The paid sample is small, but it matters for anyone who assumed ads were immune.

What is settled: across several studies using different methods, the direction is consistent. A search that shows an AI summary produces fewer clicks to websites than one that does not.

What is contested: the size of the effect, how much of the decline is attributable to AI features rather than to other changes in search over the same years, and whether any of it holds for commercial or local queries in Australia. Anyone quoting you a single precise percentage for your market is extrapolating.

What can you not measure yet?

You cannot yet measure clicks lost to AI Overviews. Search Console now reports AI feature impressions, but not the clicks, click-through rate or prompts behind them, and clicks from AI features are folded into ordinary web search.

Google added a generative AI performance report to Search Console in 2026. It shows impressions from AI Overviews, AI Mode and generative AI features in Discover, broken out by page, country, device and date. It does not report clicks, click-through rate or the prompts people typed (Search Console Help).

Clicks arriving from AI features are folded into ordinary Search performance under the Web search type (Google Search Central).

The practical consequence: you cannot currently produce a number for "traffic we lost to AI Overviews", and you should refuse to put one in a board pack. What you can produce is your own impressions-to-clicks trend, which is a real measurement of a real thing. The Search Console pages losing clicks walkthrough shows how to pull it.

PostCliq's audit includes an AI readiness check: whether your public pages can be read by search and answer-engine crawlers, and whether they answer questions in a form those systems can quote. It does not measure AI citations or traffic. The AI search visibility page sets out exactly what it checks today.

What to do next

1. Split your own Search Console data by query type. Informational queries, the how and why and what is, against commercial ones with your service and a location in them. If clicks fell on the first group and held on the second, that is the pattern this shift produces, and it is far less alarming commercially than a flat decline.
2. Look at impressions and clicks separately, not just sessions. Flat impressions with falling clicks is a different problem from falling impressions, and analytics alone cannot tell them apart.
3. Do not buy AI Overview optimisation as a separate service. Google documents no special optimisation. Eligibility is being indexed and snippet-eligible. Anyone selling more than that is selling a theory.
4. Decide deliberately about nosnippet. Opting out of snippets removes you from AI features and from ordinary search snippets at the same time. For most service businesses that trade is bad, but it should be a decision someone made rather than a default nobody examined.
5. Reweight the pages that still get visits. If the top of your funnel leaks, the visits that still arrive carry more of the load. That is the argument in every click is worth more than it used to be, and the full sourced case is our report on the post-click squeeze. It is the part of this you actually control.

The honest summary for your director is one sentence: fewer people are clicking through from search, the exact cause is not measurable with the tools we have, and the response is to convert more of the people who still arrive.

If you want to work out whether your own drop is search, campaigns or the site, how to tell a website problem from a campaign problem is the order to check things in, and a free PostCliq audit checks the pages those remaining visitors land on.

Frequently asked questions

What is zero-click search and why is it growing?

Zero-click search is a search that ends without a click to any website, because the answer appeared on the results page. It is growing because Google shows more AI summaries, featured snippets and panels that answer questions directly. SparkToro measured 68.01% of US Google searches as zero-click in early 2026, against 60.45% in 2024, though the two figures come from different panels.

What are AI Overviews?

AI Overviews are the generated answers Google shows above its normal search results for some queries, with links to supporting pages. They draw on pages already in Google's index. AI Mode is a separate, conversational search experience that launched in Australia in October 2025. Both reduce the need to click through on simple informational questions.

How do I get my website into Google AI Overviews?

Google says a page must be indexed and eligible to show a snippet, and that no special optimisation is needed. In practice, pages that answer a question directly in their opening lines, cover related sub-questions under clear headings and cite real sources are easier to quote. Check that you have not set nosnippet or blocked crawlers by accident, because either removes you.

Do AI Overviews reduce website traffic?

For informational searches, independent studies agree they reduce clicks. Pew found US adults clicked a result 8% of the time with an AI summary against 15% without, and Seer found organic click-through fell 61% on informational queries with an AI Overview. Google says total clicks are stable but published no data. Commercial and local searches may be affected less.

How do I know if AI Overviews are affecting my site?

Search Console's generative AI report shows AI feature impressions, but not the clicks you lost. The practical test is to split your Search Console queries into informational and commercial groups and compare impressions and clicks over two matched periods. If informational clicks fell while impressions held, and commercial clicks held, you are seeing the zero-click pattern.

Should I block Google from using my content in AI Overviews?

For most businesses, no. The controls that remove you from AI Overviews, such as nosnippet, also remove your ordinary search snippets, which usually costs more than it saves. Google-Extended governs training and grounding in some other Google systems and is a separate decision. Make either choice deliberately, with someone accountable for it, rather than leaving a default in place.