# Zero-Click Searches: The Real Numbers and What Still Works in 2026

> Zero-click searches now end ~68% of US Google queries. Here is what that number really means, why rankings hold while traffic falls, and the levers that still work.

- Published: 2026-08-16
- Updated: 2026-08-16
- Author: Samy BEN SADOK
- Canonical: https://geotoolbox.ai/blog/zero-click-searches

---

Your rankings are steady and your traffic is sliding anyway. That gap has a name: zero-click searches, the queries that now end on the results page without anyone clicking through. In the US, about 68% of Google searches end that way, and AI Overviews are pushing the number higher every year.

This is the honest version. What zero-click actually is, how big it really is once you sort out the competing numbers, why ranking first no longer guarantees the click, and the specific levers that still work when people stop clicking.

<figure className="not-prose my-8">
  ![Bar chart of the US Google zero-click rate: 58.5% and 60.45% in 2024 (both on the Datos panel) rising to 68.01% in early 2026 (Similarweb panel).](/blog/zero-click-searches/zero-click-trajectory.png)
  <figcaption className="mt-3 text-center text-sm text-gray-500">The 2024 readings use SparkToro's Datos panel; the 2026 reading switched to Similarweb, so the jump is partly a change of measurement. Source: SparkToro 2024 and 2026 zero-click studies.</figcaption>
</figure>

## What Is a Zero-Click Search?

A zero-click search is a query that ends on the results page itself, with no click through to any website. The searcher reads the answer, then closes the tab or types a new query. You can see the full definition in our [zero-click search glossary entry](https://geotoolbox.ai/glossary/zero-click-search), but the mechanics matter more than the label.

Zero-click happens because the results page increasingly answers the question for you. The features doing the answering are familiar: AI Overviews at the top, featured snippets, knowledge panels, People Also Ask boxes, local packs, and instant widgets like the weather, a calculator, or a sports score. Each one resolves a certain class of query without sending anyone to a site.

One thing to be clear about early: zero-click is not a metric Google reports. You will not find it in Search Console. It is a behavioral outcome measured by clickstream panels, where a research firm watches a sample of real users and reconstructs what happened after each search: a click to the open web, a click to a Google property, or no click at all. That distinction shapes every number in the next section.

## How Big Is Zero-Click, Really? (And Why the Numbers Disagree)

The headline figure is **68% of US Google searches ended without a click** in the first four months of 2026, from [SparkToro's 2026 study](https://sparktoro.com/blog/in-2026-less-than-one-third-of-google-searches-still-send-a-click/) built on Similarweb clickstream data.

That number gets quoted as 58.5%, 60.45%, and 68% across different articles, and the gap is not sloppiness. It is what happens when the measuring instrument changes underneath the trend. Here are the figures SparkToro actually reports, with the panel each one came from.

<table>
<thead>
<tr><th>Figure</th><th>Source</th><th>Panel</th><th>Window</th><th>What it measures</th></tr>
</thead>
<tbody>
<tr><td><strong>68.01%</strong> US</td><td>SparkToro 2026</td><td>Similarweb clickstream</td><td>Jan–Apr 2026</td><td>US searches that ended with no click</td></tr>
<tr><td><strong>60.45%</strong> US</td><td>SparkToro 2026 (2024 baseline)</td><td>Datos clickstream</td><td>2024</td><td>The 2024 figure the 2026 study compares against</td></tr>
<tr><td><strong>58.5%</strong> US / 59.7% EU</td><td><a href="https://sparktoro.com/blog/2024-zero-click-search-study-for-every-1000-us-google-searches-only-374-clicks-go-to-the-open-web-in-the-eu-its-360/">SparkToro 2024 study</a></td><td>Datos clickstream</td><td>Sep 2022 – May 2024</td><td>The earlier study, same panel, different window</td></tr>
</tbody>
</table>

Here is the trap. SparkToro measured 2024 with Datos and 2026 with Similarweb, different clickstream panels with different users and devices. SparkToro says so itself, calling the comparison "a bit of apples and oranges." So the honest read is that zero-click is clearly climbing, but the exact size of the jump is fuzzy, because the ruler changed between readings. Both 2024 figures, 58.5% and 60.45%, come from Datos over slightly different windows, which is the whole reason you should never lift a number from one study, set it beside a number from another, and call the difference a trend.

What is not in doubt is the direction. AI Overviews are the clearest accelerant: [Ahrefs data](https://ahrefs.com/blog/ai-overview-triggers/) that SparkToro cites puts them on more than 20% of searches, and when one appears, [click-through to the results below drops by close to 60%](https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/). For a longer read on where these numbers come from, see our breakdown of the [state of AI search in 2026](https://geotoolbox.ai/blog/state-of-ai-search-2026).

One footnote worth keeping straight: the phrase "zero-click search" itself gets searched a little less than it did a year ago, even as the phenomenon behind it hits record highs. The term cooling is not the problem cooling.

## Why Your Rankings Hold but Your Traffic Falls

If you have watched a page sit at position one while its clicks slide toward zero, you are not imagining it. The rank did not move. The click did.

The reason is simple once you name it. Ranking first used to mean you got the click, because the answer lived on your page. Now the answer often lives on the results page. When an AI Overview or a featured snippet resolves the query, the classic blue links below it become optional, and most people never scroll. That is why click-through to those links falls so sharply, even for the pages ranking directly beneath the answer. We cover how those summaries get built in [what Google AI Overviews are](https://geotoolbox.ai/blog/what-are-google-ai-overviews).

This shows up in Search Console as a specific, recognizable pattern: impressions flat or rising, clicks falling, average position steady. Your content is being shown and read. It is just being read inside Google's answer instead of on your site. Ranking number one now can mean almost no clicks at all, and the instinct is to treat it as a penalty or an algorithm hit. It usually is neither.

The unhelpful conclusion is that you should rank harder. You cannot out-rank a layout change. When the answer is delivered above your listing, a higher position does not buy back the click. That reframes the whole problem: the job is no longer only to rank, it is to be the answer and to capture the intent that still converts. The rest of this guide is about how.

## Cited but Not Clicked: What Visibility Is Actually Worth

Here is the part most guides skip. Being named in an answer is not the same as being visited, and the gap is enormous. In [Pew Research Center's 2025 study](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/) of 900 US adults across 68,879 searches, on the visits where an AI summary appeared, users clicked one of its cited source links in about 1% of cases. A citation is real estate in the answer, not a referral to your site.

So is that visibility worth anything? The answer depends on who you ask, and both camps have a point. The skeptical read, argued well by Search Engine Land, is that a citation nobody clicks is brand awareness at best and has no direct business impact. If the reader never reaches your page, your calls to action and your funnel never load. The optimistic read, from Semrush and others, is that repeated presence in answers builds trust and pulls people back later through branded search or a direct visit.

Both can be true, and where you land depends on the query. For a definitional question the reader was never going to convert on, a citation is mostly a branding impression. For a commercial question, being the cited source shapes who the buyer trusts before they ever compare options. There is a thin upside hiding in that 1%, though: the people who do click through have already read the gist, so they arrive better informed than a cold searcher would. The defensible middle is this: treat citations as a real signal, but measure them as demand, not as traffic, and expect the value to show up downstream rather than in the same session.

## Zero-Click Isn't Just a Google Problem

Most coverage frames zero-click as a Google story. It is bigger than that. When someone asks ChatGPT, Perplexity, Gemini, or an AI browser a question, there is no results page at all. The model answers directly, and if your brand is not in that answer, you are invisible in a way that has no ranking to check and no snippet to win. That is zero-click taken to its limit: not a click lost from a SERP, but a SERP that never appears.

This matters for measurement more than for panic. If you only watch Google, you are watching one surface of a problem that now spans several. A brand can be summarized accurately by Google, misquoted by one assistant, and absent from another, all at the same time, and none of it shows up in a rank tracker. Seeing the whole picture means tracking how often you are mentioned and cited across engines, not just where you rank on one. Cross-engine mention tracking is what our tools at geotoolbox are built for. If you want the mechanics, we walk through [how to track brand mentions in AI search](https://geotoolbox.ai/blog/track-brand-mentions-in-ai-search).

## How to Measure Zero-Click Impact on Your Own Site

You can measure this with tools you already have, before paying for anything.

1. **Find the erosion in Search Console.** Filter to the last six to twelve months and look for pages where impressions held or rose while clicks and CTR fell. That divergence is the signature of zero-click: you are being shown and read, just not visited. One caveat, straight from [Google's own documentation](https://developers.google.com/search/docs/appearance/ai-features): AI Overview appearances are folded into your normal Search Console traffic under the "Web" type. There is no dedicated filter to isolate them, so treat the impressions-up-clicks-down pattern as your best available proxy, not a clean read.

2. **Watch branded search and direct traffic.** When people meet your brand inside an answer and do not click, the influence often surfaces later as a branded query or a direct visit. Rising branded impressions in Search Console are a reasonable proxy that your presence in those answers is registering.

3. **Read the AI Assistant channel in GA4.** In 2026 Google [added a native AI Assistant channel to GA4](https://www.searchenginejournal.com/google-analytics-adds-ai-assistant-as-default-channel-group/574974/) that recognizes referrals from assistants like ChatGPT, Gemini, and Claude. A couple of limits keep it from being a full count. AI Overview clicks are not in it at all, because Google reports those as ordinary organic search. And any assistant Google has not added to its recognized list (Perplexity, at the time of writing), plus the many AI answers that arrive with no referrer, still land in Referral or Direct. Treat it as a floor on AI-driven traffic, not a full count, and lean on steps 1, 2, and 4 for the rest.

4. **Track citations across engines, not just clicks.** The metric that most directly tracks whether you are present is how often you are mentioned and cited in AI answers, measured over time. That is a visibility number, not a sessions number, and it is the one closest to the outcome you now care about. Our guide on [how to measure your AI visibility](https://geotoolbox.ai/blog/how-to-track-ai-visibility) lays out the full approach.

None of this requires new schema or a rebuild. It requires reading the tools you have for a different signal.

## What Actually Still Works

The fatalist take is that zero-click means SEO is dead and there is nothing you can do. That is wrong, and the people saying it are usually treating every query the same. The shift is not from search to nothing. It is from capturing traffic to capturing demand and citations. Here is what that looks like in practice.

**Start by triaging your queries.** Some still earn the click and some never will again. Spend your effort where the click survives, and optimize the rest to be the cited answer rather than fighting a layout you cannot win.

<table>
<thead>
<tr><th>Query type</th><th>Still earns the click?</th><th>What to do</th></tr>
</thead>
<tbody>
<tr><td>Definitional, "what is", quick facts</td><td>Rarely</td><td>Concede the click; structure the passage to be the cited source</td></tr>
<tr><td>Generic how-to</td><td>Decreasingly</td><td>Answer-first for citation; do not over-invest</td></tr>
<tr><td>High-intent commercial, pricing, "best X for Y"</td><td>Yes</td><td>Fight for it: sharp titles, a page that converts the visit</td></tr>
<tr><td>Comparison, "X vs Y", buying decisions</td><td>Yes</td><td>Go deep with original data a summary cannot compress</td></tr>
<tr><td>Branded and navigational</td><td>Yes</td><td>Own it; this is where zero-click demand lands later</td></tr>
</tbody>
</table>

**Become the source worth citing.** For the queries you concede, the goal is to be the passage the answer is built from. Lead sections with a direct, self-contained answer, then support it with something a model cannot flatten: proprietary data, a real framework, first-hand results. If your content can be summarized without losing anything, it will be, and you will not get credited. Our guide on [how to optimize for AI search](https://geotoolbox.ai/blog/how-to-optimize-for-ai-search) goes deeper on the formatting.

**Capture demand off the click.** Branded search, email capture, and community are how you stay reachable when Google and the assistants keep the session. Treat every zero-click impression as the top of a funnel that ends somewhere you own.

A couple of myths to drop. Blocking AI crawlers will not save your Google traffic: AI Overviews are built from Google's own index, so short of removing a page from search entirely with noindex, you cannot opt a ranked page out of them. Blocking an assistant's crawler is a real lever but a blunt one. It can lower your odds of being cited by that engine, though the bots differ (OpenAI alone runs one crawler for training and another for search) and some assistants can still surface you from third-party pages. Mostly it costs you AI visibility without buying back a single Google click. And there is no special "AI schema" to chase. Google states plainly that no structured data is required to appear in AI Overviews or AI Mode. Standard schema that matches your visible content is still worth having, but not as an AI shortcut. If you are still asking whether any of this means the end of the discipline, we answered that directly in [is SEO dead](https://geotoolbox.ai/blog/is-seo-dead).

## The Shift Is Real, and So Is the Response

Zero-click is the search page doing more of the answering, and the trend line points one way. Waiting it out is not a plan. You will not win the old clicks back by ranking harder. What you can do is decide which queries still deserve the fight, become the source the answers are built from, and measure whether your presence in those answers is actually growing.

That last part is where most teams are flying blind. If you want to see where your brand is being cited and mentioned across Google's AI Overviews, ChatGPT, Perplexity, and the rest, [geotoolbox tracks your AI visibility across engines](https://geotoolbox.ai/features/domain-overview) so you can watch the metric the click no longer captures on its own. The traffic you lost to zero-click is not coming back in its old form. Being the answer, and knowing when you are, is the version that does.

## Frequently Asked Questions

### Is it true that 60% of searches are zero-click?

It is higher than that now. The most current figure is about 68% of US Google searches ending without a click in early 2026, per SparkToro's Similarweb-based study. The 58.5% and 60.45% figures are 2024 numbers, both from the Datos panel, while the 68% comes from a different panel (Similarweb), so the studies are not directly comparable.

### How many searches are zero-click in 2026?

Roughly 68% in the US, per SparkToro's 2026 Similarweb data. It is one of the fastest-rising numbers in search, up from a Datos-panel reading in the low 60s in 2024, and AI Overviews are the clearest driver.

### Are zero-click searches killing SEO?

They are killing one version of it: rank first, collect the click. The discipline itself is moving from winning traffic to earning demand and citations. High-intent, commercial, comparison, and branded queries still earn clicks. Generic definitional and how-to queries increasingly do not, and for those the goal becomes being the cited source.

### Should I block AI crawlers to stop zero-click searches?

No. Blocking AI crawlers will not stop Google's own AI Overview from summarizing a page you already rank for, and blocking an assistant's crawler mostly just lowers your odds of being cited there. You lose AI visibility without recovering the Google click.

### How do I measure zero-click search impact?

Look in Search Console for pages where impressions held or rose while clicks and CTR fell, which is the zero-click signature. Track branded search and direct traffic as downstream signals, use GA4's AI Assistant channel for ChatGPT, Gemini, and Claude referrals, and monitor how often you are cited across AI engines rather than only how you rank.

### Do zero-click citations actually convert?

Not directly and not often in the same session. Pew found that on visits where an AI summary appeared, users clicked one of its cited links in about 1% of cases. The value shows up downstream, as branded search and better-informed visits later, and it is stronger for commercial queries than for purely informational ones.

## Sources

- SparkToro - In 2026, Less than One Third of Google Searches Still Send a Click, 2026 - `sparktoro.com/blog/in-2026-less-than-one-third-of-google-searches-still-send-a-click`
- SparkToro - 2024 Zero-Click Search Study, 2024 - `sparktoro.com/blog/2024-zero-click-search-study-for-every-1000-us-google-searches-only-374-clicks-go-to-the-open-web-in-the-eu-its-360`
- Ahrefs - AI Overview triggers and their effect on click-through rate, 2025 - `ahrefs.com/blog/ai-overview-triggers` and `ahrefs.com/blog/ai-overviews-reduce-clicks-update`
- Pew Research Center - Google users are less likely to click on links when an AI summary appears, July 2025 - `pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results`
- Google Search Central - AI Features and Your Website, 2026 - `developers.google.com/search/docs/appearance/ai-features`
- Search Engine Journal - Google Analytics Adds AI Assistant as Default Channel Group, 2026 - `searchenginejournal.com/google-analytics-adds-ai-assistant-as-default-channel-group/574974`