--- title: "Zero-Click Search: What the Data Shows and What to Measure" description: "Pew, SparkToro and Ahrefs measure zero-click behaviour in different ways. Here is what each shows, what none can prove and what to put on a leadership slide." author: "Team" category: "Search & AI Visibility" date: 2026-11-06T06:45:00.593Z canonical: "https://contentagents.dev/blog/zero-click-search-what-the-data-shows-and-what-to-measure" --- # Zero-Click Search: What the Data Shows and What to Measure ![peacock butterfly, flowers visit, nectar search](https://cdn.pixabay.com/photo/2018/08/09/12/53/peacock-butterfly-3594447_1280.jpg?w=1200&q=75) > Pew, SparkToro and Ahrefs measure zero-click behaviour in different ways. Here is what each shows, what none can prove and what to put on a leadership slide. Open Search Console, see clicks sliding and the board asks whether AI is eating your traffic. Three different studies give three different answers, each measured a different way, and a report that mixes them overstates or understates the loss. The aim here is to separate the evidence into what each study counted, then say what to put on a leadership slide instead of a single falling line. The most cited measure comes from [Pew Research Center, whose 2025 data put regular-result clicks at 8% of visits with a Google AI summary and 15% without](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/). The other numbers in circulation use different panels, years and definitions. ## How the evidence was read We read the Pew short read, the SparkToro zero-click study from 2024, the Ahrefs click-through analysis from 2025 and Google's own announcements. We could not open the Gartner press release because its site blocks automated readers, so we cite it as a forecast only. We ran no study of our own and used no paid data. ## Study one: Pew, 8% against 15% The Pew team followed 900 US adults through March 2025 and logged 68,879 unique Google searches, [about 18% of which produced an AI summary](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/). A summary page gave a regular-result click in eight visits out of a hundred, where a page without one gave fifteen. Links inside the summary were used in roughly one visit in a hundred, and the browsing session ended on 26% of summary pages against 16% of the others. Pew's design compares visits with and without a summary. Those two groups are not the same kinds of search, as the next section explains, so the gap is a description of behaviour, not a measured effect of the summary itself. ## Study two: most searches never produced a click, even before summaries SparkToro, working with the Datos clickstream panel for September 2022 to May 2024, reported that [58.5% of US Google searches and 59.7% of EU searches ended with no click](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/). Roughly 37% of searches ended the session and about 22% led straight to another query. The study counts clicks on Google's own properties as clicks, and its panel excludes iPhone users. The post itself now carries a note that it uses data from an earlier study and may be out of date, and it predates the broad rollout of AI Overviews. The takeaway is that zero-click behaviour is old. Summaries add to it, they did not create it. ## Study three: position one lost clicks where an overview appeared Ahrefs compared 150,000 keywords with an AI Overview against 150,000 informational keywords without one, using aggregated Search Console data for March 2024 and March 2025. Position-one click-through on the overview keywords fell from about 7.3% to 2.6%, and Ahrefs reported [a 34.5% lower position-one CTR than a forecast based on the trend for keywords without overviews](https://ahrefs.com/blog/ai-overviews-reduce-clicks/). Ahrefs calls the finding correlational, and notes that its own figure does not exactly reproduce from the rounded inputs it shows. The three studies agree on direction and disagree on size, because they count different things: visits, searches and keywords, over different months. ## A forecast is not a measurement In February 2024 [Gartner predicted a 25% fall in traditional search engine volume by 2026](https://www.gartner.com/en/newsroom/press-releases/2024-02-19-gartner-predicts-search-engine-volume-will-drop-25-percent-by-2026-due-to-ai-chatbots-and-other-virtual-agents) as chatbots and virtual agents take share; we could not read the release, so this rests on its title. At its May 2026 event Google said [AI Mode had passed one billion monthly users](https://blog.google/products-and-platforms/products/search/search-io-2026/) and that search activity had reached a record. Both can hold at once: more searching, fewer clicks per search. ## What these studies cannot support **Time.** Pew covers one month in 2025, SparkToro ends in May 2024 and Ahrefs compares March of two years. Products change faster than the studies. **Place.** Pew covers US adults. SparkToro adds the EU. Nothing here covers other markets or other engines. **Query type.** In the same [Pew data](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/), summaries appeared on 8% of one- or two-word searches, 53% of searches of ten or more words and 60% of question-style searches. The effect concentrates in long questions. See [which searches trigger AI Overviews](/which-searches-trigger-ai-overviews-what-pew-data-shows). **Value.** Someone who spots your brand in a summary may look you up next week. Awareness survives even when the click does not, and none of these studies tried to capture it. ## Our position: stop reporting one click line The usual move is to chart organic clicks and add an AI caption. We think that is the weakest report you can send. A single line cannot tell a collapse from a shift, and the Pew gap mixes two kinds of search, so a leader who reads [Pew's 8% against 15%](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/) as "AI halves clicks" is making a causal claim the data does not make. We would report three rows instead. The cost is real: it takes more setup, and impressions data can look discouraging beside a clicks line that has held. The open question is causation. Match searches by query and intent in a controlled test; if summaries turn out to be the cause of most of the gap, this reading fails. Until someone publishes that, hold the claim loosely. ## What to put in front of leadership Row one is organic clicks, six-month trend. Row two is impressions and branded search over the same period. Row three shows how often assistants name or cite you, averaged over four weeks on a fixed prompt list. Google's Search Console now helps with row two: since June 2026 it has offered dedicated [generative AI performance reports showing impressions, pages, countries, devices and dates](https://developers.google.com/search/blog/2026/06/gen-ai-performance-reports) for AI Overviews, AI Mode and Discover, and they reached all websites by 31 August 2026. Earlier analyses, including Ahrefs', noted that Search Console could not separate these, so older advice that you cannot isolate AI traffic is out of date. Buyers themselves are a fourth source, so put a source question on forms and calls. Under the slide, add one sentence of caveat per row: the Pew sample is American and a single month, assistant presence is sampled, and an impression is not a visit. Falling clicks beside steady impressions and branded search tell a different story from three falling lines. The [Search Console reporting guide](/how-to-report-on-ai-overviews-in-search-console) and the [AI visibility metrics guide](/ai-visibility-metrics-complete-measurement-guide) show how to build the rows, [share of voice](/share-of-voice-in-ai-search-definition-formula-and-a-worked-example) covers the third, and [buyers decide before they click](/b2b-buyers-decide-before-they-click-plan-for-that-1mmv) explains why the click was never the whole story. ## How to run the three-row view in a week Pull organic clicks, impressions and branded queries from Search Console for the last six months and chart them on one page. Open the generative AI report and note impressions in AI features by page, so you can see which pages appear in summaries at all. Set up the fixed prompt list and run it weekly. Collect self-reported sources for a month. At the end, write two sentences on what moved, and name one page you will change in response. ## Conclusions we would not draw We would not conclude that content marketing has stopped working, since Pew counted clicks on Google results and not trust. Nor that every query loses clicks alike, because summaries depend on query length and type. Nor that the March 2025 figures still hold, as the product has changed since. ## FAQ ### What is a zero-click search? A search where the user gets what they need on the results page, or gives up, without clicking any result. ### Do AI summaries reduce clicks? In Pew's March 2025 study of US adults, users clicked a regular result in 8% of visits with an AI summary and 15% without. Ahrefs also links overviews to lower position-one click-through. Both are correlational. ### Is search traffic going to fall 25%? Gartner forecast a 25% drop in traditional search volume by 2026. It is a forecast, and Google reports record search activity. ### How should I report zero-click effects? Report impressions, branded search, citations and self-reported attribution beside clicks, and state the limits of each. --- Source: https://contentagents.dev/blog/zero-click-search-what-the-data-shows-and-what-to-measure