Zero-Click Search Statistics 2026: What Publishers Lost

For twenty years the deal was simple. Google took your content, indexed it, ranked it, and sent you a visitor. The visitor was the payment.

That deal has been quietly renegotiated. By 2026 the majority of searches resolve inside the results page, and a large share of the clicks that do happen never reach an independent website at all.

What follows is the verified data as it stands in August 2026: how many searches end without a click, how that has moved year on year, how it splits by device and query type, and what publishers have actually lost. Where the sources contradict each other, I have said so rather than picked a favourite.

Key zero-click search statistics for 2026

• Only 30.04% of US Google searches produced at least one click in January to April 2026, down from 39.55% in 2024, a relative fall of 22.9% (SparkToro, using Similarweb clickstream data) [1]
• Almost 30% of all Google clicks go to platforms Google itself owns, and roughly 1% go to Google ads, so "a click happened" and "a website got a visit" are very different measurements (SparkToro) [2]
• On desktop, 48% of searches end with no activity at all and only 20.4% produce an external click to a non-Google site (Similarweb) [3]
• On mobile, 45.2% of searches result in an external click and only 11.5% end in outright session abandonment (Similarweb) [3]
• When an AI summary appeared, 8% of Google visits ended in a click to a link, against 15% when no summary appeared, across 68,879 searches by 900 US adults (Pew Research Center, March 2025 data) [4]
• Organic click-through rate on AI Overview queries was 2.36% in February 2026 against 3.82% on queries without one, across 5.47 million tracked queries (Seer Interactive) [5]
• Paid click-through rate held at 16.21% with an AI Overview present versus 21.85% without, a far smaller relative gap than organic (Seer Interactive) [5]
• Global publisher traffic from Google organic search fell 33% year on year to November 2025 across more than 2,500 news sites, and 38% in the US (Reuters Institute / Chartbeat data) [6]
• Google search referrals to small publishers (1,000 to 10,000 daily pageviews) fell 60% between December 2024 and December 2025, against 22% for the largest sites (Chartbeat) [7]
• Informational queries fell from 91.3% to 57.1% of all AI Overview triggers between January and October 2025 as the feature spread into commercial and navigational searches (Semrush) [8]
• Median whole-site organic click-through rate sat between 1% and 2% across 422,421 websites with Search Console data in June 2026 (Ahrefs) [9]

What share of Google searches now end without a click?

On the most widely cited measure, roughly two thirds. SparkToro's analysis of Similarweb clickstream data found that fewer than one in three US Google searches produced a click of any kind in the first four months of 2026, with 30.04% generating at least one click [1].

That is the headline everyone quotes. The more interesting number is the direction of travel: click-producing searches fell 9.51 percentage points from 2024, which is a 22.9% relative decline in two years [1].

That is the figure that actually matters to a business. A percentage-point drop sounds modest. A near-quarter reduction in the share of searches that send anyone anywhere is a structural change to a channel most companies still budget for as though it were 2019.

SparkToro's own caveat is worth repeating. The panel excludes Google's mobile app and treats mobile sessions as ended after ten seconds of inactivity, so the true zero-click rate is probably higher than published, not lower [1].

What is the difference between "no click anywhere" and "no click to an external site"?

These are two separate metrics and conflating them is the most common error in zero-click commentary. A search can produce a click that never leaves Google's own estate, which counts as a click but delivers nothing to a publisher.

SparkToro's earlier study found that almost 30% of all Google clicks land on platforms Google owns, with roughly 1% going to Google ads [2]. Strip those out and the picture for independent websites is materially worse than the zero-click rate alone suggests.

The same study put the pure zero-click rate at 58.5% in the US and 59.7% in the EU, and reported that for every 1,000 US Google searches only 374 clicks reached the open web, against 360 in the EU [2]. Those figures cover September 2022 to May 2024, so treat them as the pre-AI Overviews baseline rather than current.

The practical upshot: when a vendor quotes you a zero-click rate, ask what happened to the remainder. If a third of the surviving clicks go to YouTube, Maps, Flights or a Shopping panel, the share of searches reaching an ordinary website is far below the arithmetic complement of the headline figure.

How has the zero-click rate changed year on year?

Slowly for a decade, then sharply from 2024. SparkToro's series puts zero-click at around 45% in 2016, 49% in 2019, and 60.45% in 2024, before the jump to 68.01% in early 2026 [1].

Roughly 23 percentage points have been added across the decade, and a third of that movement arrived in the last two years [1]. AI Overviews are the obvious suspect, though as the next sections show, the evidence for direct causation is messier than the timeline implies.

One structural point is often missed. Part of the growth in zero-click is query refinement rather than abandonment: SparkToro's 2024 breakdown had searchers ending the session about 37% of the time and changing the query about 22% of the time [2]. Only the first is genuinely a lost opportunity.

Do zero-click rates differ by device?

Yes, and the direction may surprise you. Similarweb's 2026 analysis reports that on desktop 48% of searches end with no activity at all and only 20.4% produce an external click to a non-Google site, whilst on mobile 45.2% of searches result in an external click and just 11.5% end in outright abandonment [3].

That inverts the usual assumption that mobile is where clicks go to die. Similarweb's reading is that desktop skews towards research behaviour, including the B2B vendor comparisons and specification checks that AI summaries answer well, whilst mobile skews towards navigational and local intent where a tap on a specific destination is the point of the search [3].

Treat the split with caution. SparkToro's 2026 study published no device-level breakdown and instead assumed a two thirds mobile, one third desktop weighting drawn from external benchmarks [1], which suggests device-level zero-click measurement is not yet settled.

If you sell to businesses, the desktop figure is the one to plan around. Roughly four in five desktop searches producing no external click is a harsher environment than any blended average will show you.

Which query types still send clicks?

Commercial and transactional queries, mostly. Google has been noticeably more conservative about AI Overview coverage on searches with explicit purchase intent, because the user still has to reach a shop to complete the transaction [3].

The Semrush AI Overviews study shows the feature migrating outward from its informational home. Informational queries made up 91.3% of AI Overview triggers in January 2025 but only 57.1% by October 2025, as commercial triggers rose to 18.57%, transactional to 13.94% and navigational to 10.33% [8].

Source: Semrush AI Overviews study [8]

The commercial safe harbour is therefore shrinking, not fixed. If your defence against zero-click is that your money keywords have transactional intent, that defence had a shelf life of about eighteen months on this evidence.

One more group still clicks reliably: people who have already decided which brand they want. Ahrefs' Search Console benchmark across 422,421 sites found median whole-site organic CTR of 1% to 2%, with Domain Rating 90 to 100 sites at 2.91% against 0.39% for DR 0 to 10 [9]. Brand recognition now does much of the work ranking position used to.

How much of the results page do AI features actually occupy?

Nobody agrees, and this is the single largest source of confusion in the field. Semrush data reported by Press Gazette put AI Overviews on almost half, 47%, of Google results pages across all query types in 2026 [10]. BrightEdge's guidance, updated March 2026, states the opposite: that AI Overviews appear on less than 15% of all queries for logged-in US users [11].

Both can be technically true. Keyword-set measurement over a large tracked corpus overweights the long, informational queries that trigger AI Overviews most often, whilst logged-in user measurement reflects the real mix of navigational and branded searches that dominates actual usage.

Feature real estate is also being reshuffled rather than simply added. Newzdash analysis reported by Press Gazette found that 15.5% of Google results for trending news topics now show a Top Stories carousel inside the AI Overview instead of in its own standalone slot, rising to 17.5% in the UK, across more than 17 million Top Stories tracked in the 30 days to 28 July 2026 [12].

That is a reminder that "AI Overview present" is not one thing. A summary with a news carousel embedded in it is a different clicking environment from a plain three-paragraph answer, and studies treating them identically produce averages that describe neither.

How much traffic have publishers actually lost?

More than the zero-click rate on its own would predict. Chartbeat data cited in the Reuters Institute's 2026 trends report found global publisher traffic from Google organic search down 33% year on year to November 2025 across more than 2,500 news websites, with the US down 38% and Google Discover referrals down 21% [6].

Chartbeat's own analysis breaks that down by publisher size, and the gradient is brutal.

Source: Chartbeat [7]

Total network pageviews fell only about 6%, from roughly 7.64 billion weekly in 2024 to 7.19 billion in 2025, as internal navigation rose from around 38% to about 41% of pageviews and dark social grew from roughly 7% to 10% [7]. Publishers did not lose a third of their audience. They lost a third of the audience Google was handing them, and partly replaced it by working homepages, apps and newsletters harder.

Nor is the audience simply moving to chatbots at speed. The Reuters Institute's 2026 Digital News Report, surveying 48 markets and 97,520 respondents, found weekly use of AI chatbots for news at 10%, up from 7% in 2025, and 16% among under-35s [16].

AI referrals are not filling the gap. Chartbeat puts AI chatbot referrals at under 1% of total network pageviews despite ChatGPT referrals growing more than 200% year on year [7], and Previsible's analysis of 166 GA4 properties found AI penetration in the publisher category at just 0.08% in May 2026 [13].

What still gets clicked in 2026?

Paid results, branded destinations, and anything transactional. Seer Interactive's tracking of 53 brands across 5.47 million queries found paid CTR of 16.21% when an AI Overview was present against 21.85% when it was not, a much shallower decline than organic, which sat at 2.36% versus 3.82% in the same month [5].

Source: Seer Interactive, January 2025 to February 2026 [5]

There is also tentative evidence of a floor. Seer recorded organic CTR on AI Overview queries bottoming at 1.31% in December 2025 and recovering to 2.36% by February 2026 [5]. One quarter is not a trend, and December is a seasonally odd month, so nobody should build a plan on it.

Ahrefs' earlier work remains the cleanest causal estimate of the damage: position one CTR on AI Overview keywords fell 34.5%, from 0.073 in March 2024 to 0.026 in March 2025, across 300,000 keywords [14]. The ranking did not change. The value of the ranking did.

How to read these numbers

Panel data and server-side data measure different things. SparkToro and Similarweb infer behaviour from a clickstream panel, which captures what happens between sites but relies on projection to national totals. Seer and Ahrefs use aggregated Search Console data, which is accurate for the sites in the sample but blind to searches those sites never appeared in.

Definitions move the answer more than methodology does. Semrush's before-and-after analysis of the same keywords found zero-click rates falling from 33.75% to 31.53% once AI Overviews appeared [8], which looks like a flat contradiction of the 68% headline. It is not: Semrush is measuring per-keyword zero-click on a tracked keyword set, SparkToro is measuring per-search behaviour across a user panel. Different denominators, different numbers, both defensible.

Google disputes the whole framing. In August 2025 the company stated that total organic click volume to websites had been "relatively stable year-over-year" and that average click quality had increased, whilst characterising third-party traffic-decline reports as based on flawed methodologies [15]. Google has never published the underlying figures, so this remains an assertion rather than evidence.

Watch the dates. The Pew analysis, still the best-designed study of AI summary click behaviour, is built on browsing data from March 2025 [4]. It tells you what happened, not what is happening.

And be sceptical of any single number offered as "the" zero-click rate, including 68.01%. It is a US figure, from one panel, for a four-month window, on a definition that excludes Google's mobile app.

What this means going into 2027

Stop reporting organic sessions as your primary search metric. Impressions, share of AI Overview citations and branded search volume together describe your visibility more honestly than a sessions line that is falling for reasons unrelated to your performance.

Assume the commercial safe harbour keeps shrinking. Transactional queries went from 1.98% to 13.94% of AI Overview triggers in ten months [8], and there is no evident reason for that curve to stop where it is.

Budget for the paid gap. Paid CTR is holding up far better than organic under AI Overviews [5], which means the effective cost of the same demand is shifting from an SEO line item to a media one whether or not anyone plans for it.

Small sites should plan for asymmetric pain. A 60% fall in Google referrals for sites under 10,000 daily pageviews against 22% for the largest [7] tells you that whatever Google is now rewarding, it correlates with scale and established brand.

Finally, treat AI referral traffic as a signal rather than a channel. At under 1% of publisher pageviews [7] and 0.08% penetration in that category [13], it is not a revenue line yet. It is evidence that your content is being read by the systems that decide whether anyone is told your name. If you want help working out how your own search visibility is really tracking under AI Overviews, you can book a call with me.