Roughly 68% of US Google searches now end without a click. That is the headline from SparkToro’s June 2026 study, and it is the cleanest public read we have on how common zero click searches have become. Two years earlier the same measure sat at 60.45%.
But that number is contested. Semrush’s clickstream data points the other way. Pew Research measured something different again. The studies disagree because they are not measuring the same thing.
Below is what the evidence actually supports. Who is really losing traffic. Who is not. And what still earns attention and revenue when the click never happens.
What are zero click searches, exactly?
A zero click search is a search that ends without the user visiting a website. The answer is already on the results page. An AI Overview, a featured snippet, a knowledge panel, a weather card, a sports score, a currency conversion, a local pack with the phone number right there.
The user got what they came for. They just did not visit anyone to get it.
What the research actually counts
This is where it gets slippery. Most researchers file three quite different behaviours under zero click searches:
- True no-click. The user reads the answer and stops. Session over.
- Refinement. The user runs another Google query instead of clicking. SparkToro found this grew by 7.2 percentage points between 2024 and early 2026.
- Clicks that stay inside Google. YouTube, Maps, Flights, Images, Shopping. A click happened. It just never reached the open web.
That third bucket does a lot of quiet work. SparkToro’s 2026 figures put open-web clicks at about 27.6% of US Google searches. Out of every 1,000 searches, roughly 276 land on a site Google does not own.
What it is not
A zero click search is not a failed search, and it is not automatically a lost customer. Someone who reads your brand name inside an AI Overview, then searches your name directly, has generated two searches and one very warm visit. Last-click reporting hides the first half of that story completely.
The real numbers behind zero click searches in 2026
Five credible studies. Five different answers. Here they are side by side, with the caveat that matters most for each.
| Study | Method and sample | Period | Headline finding | Main caveat |
|---|---|---|---|---|
| SparkToro + Similarweb | Desktop and mobile web clickstream panel, US | Jan-Apr 2026 | 68.01% of searches end with no click; 27.6% reach the open web | Excludes the Google mobile app, where zero click features are more aggressive |
| Pew Research Center | 900 US adults on KnowledgePanel sharing real browsing data; 68,879 searches | Mar-Apr 2025 | 8% clicked a result when an AI summary appeared, vs 15% without | Small panel, single month, pre-AI Mode |
| Ahrefs | 300,000 keywords (150k with AI Overviews, 150k without) matched to Search Console CTR | Dec 2023 vs Dec 2025 | Position-one CTR on AI Overview keywords fell from 7.3% to 1.6% | Correlation, not proof of cause; desktop-weighted |
| Semrush + Datos | 10M+ keywords for prevalence, 200k+ for click behaviour | Jan-Nov 2025 | For the same keywords before and after AI Overviews, no-click rate fell from 33.75% to 31.53% | Measures keywords, not searches; a completely different unit |
| Seer Interactive | 53 brands, 5.47M queries, 2.43bn organic impressions from client Search Console data | Jan 2025-Feb 2026 | Organic CTR with an AI Overview dropped to 1.31% by Dec 2025, then recovered to 2.36% | Client portfolio, not a representative sample of the web |
Read that table twice. SparkToro says two thirds of searches are zero click searches and rising fast. Semrush says the no-click rate on AI Overview keywords went slightly down. Both are reporting their data honestly.
The country picture adds more texture. Using the same method, SparkToro put the UK at 69.5%, the US at 68.0%, France at 65.3%, Canada at 63.8%, Italy at 63.4% and Germany at 62.1%. The spread is real but modest. This is not a US-only phenomenon.
Why the studies on zero click searches disagree
The gap between 68% and “basically flat” is not incompetence. It is four methodology differences stacked on top of each other.
The unit of analysis changes everything
SparkToro counts searches. Every query a panelist runs is one observation. Semrush counts keywords and asks what share of searches for that keyword ended without a click. If AI Overviews expand into shorter, more transactional queries that always converted well, the per-keyword average can improve while the total share of zero click searches still climbs. Simpson’s paradox, live on a SERP.
Panels are not the population
Clickstream panels run on browser extensions and desktop software. That skews toward desktop users, toward people comfortable installing tracking software, and historically toward heavier ad-blocker use. SparkToro’s own write-up flags this. They modelled a two-thirds mobile, one-third desktop split rather than measuring it directly, and they borrowed paid-click benchmarks because mobile paid clicks were not cleanly observable.
Search Console sees only half the picture
Ahrefs and Seer both use Google Search Console. That gives you enormous sample sizes and real clicks, but only for sites you own or manage, and only for queries where you already had an impression. You cannot see the searches where nobody clicked anything. So GSC-based studies measure CTR change, not zero click rate.
Coverage gaps are enormous
The SparkToro numbers exclude the Google app on iOS and Android. Google has said AI Mode passed a billion monthly users after I/O 2026, yet SparkToro’s browser panel saw AI Mode on only 0.34% of searches from January to April 2026. That gap is the app. The true zero click rate is almost certainly higher than 68%, not lower.
None of this makes the trend fake. Every method agrees the direction is up. It means the precise share of zero click searches is unknowable to two decimal places, and anyone quoting one to two decimal places is overselling.
Practical takeaway: stop quoting a single percentage as if it were a fact about your site. Use the studies to understand direction and mechanism, then measure your own properties.
How AI Overviews and AI Mode changed click-through rate
This is where the evidence lines up rather than conflicts. Every independent dataset says the same thing: when an AI answer appears, fewer people click. AI Overviews are the single biggest engine converting ordinary queries into zero click searches.
Pew Research put 900 US adults on a panel that shared real browsing data, then looked at 68,879 Google searches from March 2025. When an AI summary appeared, users clicked a traditional result on 8% of visits, against 15% without one. They clicked a link inside the AI summary on just 1% of visits. And 26% of AI summary pages ended the browsing session entirely, versus 16% for ordinary results.
Ahrefs matched 300,000 keywords to Search Console CTR and compared December 2023 with December 2025. For informational keywords at position one, CTR fell from 7.6% to 3.9% where no AI Overview appeared. Where one did appear, it fell from 7.3% to 1.6%. Their framing is blunt: for every 100 clicks a top-ranking page used to earn, Google now keeps roughly 58.
AI Mode is a different animal
AI Overviews sit above ten blue links. AI Mode replaces them. Google announced that AI Mode passed a billion monthly users around I/O 2026, and its referral behaviour is far stingier than classic search. Similarweb’s analysis puts AI Mode referrals at a low single-digit percentage of queries, against roughly 17-19% for traditional Google results.
The good news, such as it is, sits in the Seer Interactive data. Across 53 brands and 5.47 million queries, informational queries where the brand was cited in the AI Overview averaged 2.07% organic CTR. Where the brand was not cited, 0.94%. Being the cited source delivers roughly 120% more clicks per impression. It still trails a clean no-AI-Overview SERP at 3.35%, but citation is clearly worth chasing. That is the whole argument for generative engine optimization in one statistic.
Which query types and industries take the hit
Averages hide the story. Zero click searches are not distributed evenly, and knowing where they concentrate tells you which pages to stop investing in.
Seer’s 2026 analysis found AI Overviews on 95.4% of comparison queries and 85.9% of question-format queries, but only 36% of informational, 8% of commercial and 5% of transactional ones. Question-shaped content is the blast radius. “Best X for Y” comparisons are right in it too.
Semrush’s tracking of 10 million-plus keywords through 2025 showed AI Overview prevalence swinging hard: 6.49% of queries in January, a 24.61% peak in July, then back to 15.69% in November. Google keeps adjusting the dial. Intent mix shifted too. Informational queries made up 91.3% of AI Overview triggers in January 2025 and 57.1% by October, as commercial, transactional and navigational triggers all expanded.
By industry, Semrush found Science (25.96%), Computers and Electronics (17.92%) and People and Society (17.29%) most saturated by November 2025. Real Estate, Shopping and Arts and Entertainment sat under 3% each.
The pattern across sources is consistent, and it is about query shape rather than industry label. Zero click searches cluster wherever a short, general, factual answer will satisfy the user:
- Hit hardest: definitions, health explainers, how-to content, recipe and travel blogs, glossary pages, generic comparison posts.
- Largely intact: branded queries, local intent, transactional and bottom-funnel queries, anything needing a login, a booking, a price or a real product.
If your traffic model rests on informational top-of-funnel content that any model can summarise in four sentences, you are the target. If it rests on people who already know your name, you are mostly fine. Check which side you are on with an AI Overview checker before you rewrite anything.
Why some publishers report catastrophe and others barely notice
In July 2026 the Wall Street Journal reported that several large publishers were weighing whether to pull out of Google Search entirely. The Semrush figures behind that story, corrected the following day, showed trailing twelve-month search traffic declines of 44% for the Washington Post, 43% for Business Insider, 35% for Reuters, 28% for USA Today and 27% for Time.
In the same dataset, The Guardian was up 16% and the BBC up 15%.
Same search engine. Same year. Opposite outcomes. Zero click searches did not hit these publishers evenly, and three things explain most of the variance.
- Content mix. Sites built on evergreen explainers and service journalism get summarised. Sites built on original reporting, live events and distinctive voices do not summarise cleanly, so the citation still needs a click.
- Brand demand. If people type your name into the box, AI Overviews barely touch you. If people found you through “how long to boil an egg”, you were always renting that traffic.
- Concentration risk. A publisher taking 70% of sessions from Google organic feels a 20% CTR drop as an existential event. One taking 25% from search, with the rest from apps, newsletters and direct, feels it as a bad quarter.
The small-publisher stories are grimmer than the big ones. Trade coverage through 2025 and 2026 documented independent sites losing 50% to 90% of search traffic, with music site Stereogum reporting roughly a 70% fall in ad revenue. Big brands have portfolios. Niche informational sites have one channel.
Google’s own position, stated by Search head Liz Reid in August 2025, is that total organic click volume has been “relatively stable year-over-year” and that click quality has improved. Google published no numbers to support it. Both things can be true: aggregate clicks holding roughly steady while the distribution shifts violently, with Reddit, YouTube, retailers and a handful of big brands absorbing what smaller sites lose.
The impressions versus clicks reframe
Open Search Console. If impressions are climbing while clicks fall, you have met what Ahrefs named the Great Decoupling.
Ahrefs measured this on their own property. Across six months of late 2024, the correlation between clicks and impressions was +0.425. In the following six months it was -0.352. The two lines that moved together for twenty years pulled apart.
Two mechanisms drive it. First, a single AI Overview can generate impressions for several cited URLs at once, so impression counts inflate. Second, the answer sits above the fold, so the click rate on those impressions collapses. Your ranking did not change. The SERP did.
Zero click searches are the reason the two metrics came apart. This changes what an impression means. It used to be a near-guaranteed shot at a click. Now it is closer to an ad impression: a unit of brand exposure with a low, uncertain conversion rate. That is uncomfortable if you are paid on sessions. It is fine if you are paid on revenue.
Google has partly acknowledged the shift. In June 2026 it launched Search generative AI performance reports in Search Console, showing impressions from AI Overviews, AI Mode and AI features in Discover, with data starting 18 May 2026. The rollout began with a subset of UK site owners, driven by a Competition and Markets Authority requirement, with global expansion promised. The catch: it reports impressions only. No clicks, no CTR, no queries.
So Google now confirms you were shown inside an AI answer, and still will not tell you what it was worth. Plan your measurement accordingly.
Eight moves that work when zero click searches dominate
None of this is a reason to stop doing SEO. It is a reason to change what SEO is for. The goal shifts from harvesting clicks to being the source that zero click searches quote. Here is the order I would work in.
- Chase citation, not just ranking. Seer’s data says being cited inside an AI Overview roughly doubles clicks per impression versus ranking on the same SERP uncited. Track which of your pages get cited and which get skipped. Our AI visibility checker and LLM rank tracker exist for exactly this.
- Answer in the first 60 words. Models extract. Lead with a direct, self-contained answer, then expand. Buried conclusions do not get quoted.
- Publish what cannot be summarised. Original data, proprietary benchmarks, calculators, tools, interactive comparisons, real teardowns with screenshots. A four-sentence summary of your survey still makes people want the survey.
- Build brand demand deliberately. Branded search is the one query class AI answers barely intercept. Podcasts, YouTube, conference talks, an opinionated newsletter, a recognisable point of view. This is slow and it compounds.
- Get onto the third-party sources engines trust. Semrush’s June 2026 AI Visibility Index, built on 126 million US AI prompts, found brands sustaining visibility through independent reviewers, retailers and community platforms rather than their own sites alone. Reddit threads, review sites, YouTube, industry roundups.
- Own the audience directly. Email list, community, app, first-party data. Every zero click search is a reminder that borrowed distribution can be withdrawn without notice.
- Make the crawlers’ job trivial. Clean HTML, server-rendered content, sensible headings, schema markup, a maintained llms.txt, fast pages. None of it is magic. All of it removes friction between your facts and a model’s context window.
- Rebalance the content portfolio. Move budget from definition and question posts toward transactional, comparison-with-real-data and bottom-funnel pages. Keep some informational content for citation value, but stop expecting it to pay for itself in sessions.
The uncomfortable version: your content’s new job is often to be the source the machine quotes, not the page the human lands on. Price your content strategy accordingly.
If you want the longer treatment of how citation actually works across engines, start with what AI visibility means and then check whether ChatGPT already cites your site.
Measure visibility, not just sessions
The reporting problem is real. Semrush’s 2026 AI Visibility Index found 45% of marketing leaders cannot accurately measure brand visibility inside AI answers, and only 9% have tools covering every platform they care about.
You will not get clean attribution back. Zero click searches broke the last-click model permanently. Build a dashboard that triangulates instead:
- Citation share. How often you appear in AI Overviews, AI Mode, ChatGPT, Gemini and Perplexity for your priority prompts. Track it weekly.
- Branded search volume. The cleanest proxy for demand that AI answers cannot intercept.
- Direct and dark traffic. Rising direct sessions alongside falling organic often means the funnel moved, not that it broke.
- Conversion rate by source. Fewer, better-qualified visits is a real outcome, not a consolation prize. Watch whether organic conversion rate rises as volume falls.
- Search Console AI impressions. Imperfect and click-free, but it is the only first-party confirmation Google gives you.
One caution on platform differences. Semrush found ChatGPT cites around 15 sources per response while Gemini averages three, and that on Gemini the overlap between brands mentioned and domains cited can be as low as 30%. Being talked about and being linked are separate problems. Measure both.
Set the baseline now. Zero click searches will keep rising, and the teams that adapt fastest are the ones already reporting on visibility instead of arguing about sessions.
Frequently asked questions
What are zero click searches?
A zero click search is a Google or AI search that ends without the user clicking through to any website. The answer appears directly in the results, usually through an AI Overview, featured snippet, knowledge panel or local pack. Most research also counts searches that end in a refinement, or in a click to a Google-owned property like YouTube or Maps, as zero click searches.
What percentage of Google searches are zero click in 2026?
SparkToro’s June 2026 study, using Similarweb clickstream data from January to April 2026, put the US figure at 68.01%. Only about 27.6% of searches sent a click to the open web. The study excludes the Google mobile app, so the real rate is likely higher.
Why are my Search Console impressions up but my clicks down?
That is the Great Decoupling, and it is normal now. A single AI Overview can create impressions for several cited URLs at once, inflating impression counts, while the answer above the results absorbs the click. Your rankings probably did not change. The layout did.
Do AI Overviews really cut click-through rate by 60%?
Roughly, for informational queries at position one. Ahrefs found position-one CTR on AI Overview keywords fell from 7.3% in December 2023 to 1.6% in December 2025, a 58% reduction versus keywords without one. Pew Research measured 8% versus 15% overall. Numbers differ by method, but the direction is consistent.
Is SEO dead because of zero click searches?
No, but the 2019 playbook is. Ranking still matters because citation in AI answers correlates strongly with ranking well. What has changed is the payoff: in a world of zero click searches, visibility and citation often matter more than sessions. Budget shifts toward brand, transactional pages and being quotable.
Can I stop Google using my content in AI Overviews?
Partly, and only recently. Google added AI feature controls in Search Console that took effect on 17 June 2026, letting site owners opt out of appearing in AI Overviews and AI Mode. Google states the opt-out is not used as a ranking signal elsewhere. Most publishers still consider the trade-off unattractive.
Which searches still send clicks?
Branded, local and transactional queries hold up best. Seer Interactive found AI Overviews on just 5% of transactional and 8% of commercial queries, versus 95.4% of comparison queries. Anything requiring a booking, a login, a live price or a purchase still needs a destination.
Does getting cited in an AI Overview actually help?
Yes, measurably. Seer Interactive’s analysis of 53 brands found informational queries where the brand was cited averaged 2.07% organic CTR against 0.94% when not cited. That is roughly 120% more clicks per impression. It still trails a SERP with no AI Overview at 3.35%.
Why do different studies report such different numbers?
Because they measure different units on different populations. SparkToro counts individual searches in a clickstream panel; Semrush counts keywords; Ahrefs and Seer use Search Console, which only sees queries where you already had an impression. Panel composition, mobile app coverage and date ranges do the rest.
How do I explain a traffic drop to leadership?
Show impressions alongside clicks, not clicks alone. Pair that with branded search volume, conversion rate by channel and citation share in AI answers. If conversion rate is holding or rising while sessions fall, the business impact is far smaller than the traffic chart suggests.
Should I still publish informational content?
Yes, but for a different reason. Informational pages are now largely a citation and brand-exposure play rather than a traffic play. Keep the ones that establish expertise or get you quoted, cut the thin definition posts, and move the freed budget to bottom-funnel pages.
