Organic traffic has stopped being the only currency of visibility: a brand now wins or loses inside the answer itself, before any click. Here is how far click-through has fallen, and how to rebuild your metrics for it.
You pay for content and SEO while more and more users get their answer in the results and never reach the site.
Classic analytics — sessions, CTR — is blind to the impressions where the brand was cited and no click followed.
"Publish more articles" fails when click losses are substantial across whole query sets — with the sampling, vertical and region caveats of each study.
Most Google searches end without a click to the open web. In the SparkToro and Datos study (2024), for every 1,000 Google searches in the US only 360 clicks reach a property that is not owned by Google; in the EU the figure is 374. Zero-click searches accounted for 58.5% in the US and 59.7% in the EU, and almost 30% of all US clicks go to platforms Google owns [83].
An AI summary roughly halves the chance of a click. Pew Research Center recorded clicks on a traditional search result in 8% of visits where an AI summary appeared, against 15% without one. Links inside the AI block itself were clicked in just 1% of visits, and users ended their browsing session on 26% of pages with an AI summary against 16% without [78].
Ahrefs compared 300,000 keywords and recorded a 58% fall in average CTR for position 1 where AI Overviews are present; position 2 fell 50.8% and position 3 fell 46.4% [92].
Claim: the value of ranking has shifted from "bring them to the site" to "be cited in the answer". Argument: if an AI summary appears on around 18% of searches [78] and cuts up to 58% of clicks from top positions [92], part of the SEO budget now buys visibility without a visit. Evidence: Pew found that 88% of AI summaries cite three or more sources — the mention becomes a distributed resource rather than a single winner's prize.
Share of visits with a click on any link — without AI: 15% · with AI: 8% — 900 US adults, KnowledgePanel, 68,879 searches, March 2025.
CTR of position 1 (desktop) — without AI: 0.076 → 0.037 (forecast) · with AI: 0.016 (observed) — 300,000 keywords, Google Search Console data, Dec 2023 vs Dec 2025.
Clicks to the open web per 1,000 searches (US): 360 — tens of millions of panelists, Sep 2022 – May 2024.
Methodological caveat: Pew sampled only US adults over a single month (March 2025), and behaviour may differ by country and audience. SparkToro/Datos runs on a panel with minimal iOS device coverage, and ad blockers understate paid clicks. Ahrefs uses aggregated Search Console data, and its without-AI CTR is a modelled forecast rather than an observation.
Claim: zero-click does not mean a lost user. Argument: Pew found roughly two thirds of searches ended in leaving or a new query with no click — yet a brand cited in the summary still shapes awareness and intent. Evidence: SparkToro notes total search volume rising while referrals to the open web sit at historic lows — the audience has not disappeared, the point of contact has moved.
The practical conclusion: the "organic traffic" metric needs a companion — share of answer, the percentage of relevant AI answers in which the brand appears as a source. Share of answer is an additional metric, not a replacement for every SEO metric: it rises where CTR falls, and it is read alongside traffic, conversion and positions rather than instead of them.
Ahrefs also notes that an earlier measurement (April 2025) put the decline at 34.5%, while independent studies from Seer Interactive, Kevin Indig and Authoritas reported falls in the 47.5–65.2% range — the estimates agree in order of magnitude and differ in method.
Per SparkToro/Datos (2024), about 360 of every 1,000 US searches produce a click to the open web, and 374 in the EU. Over 58% of searches end with no click, and almost 30% of clicks go to Google-owned properties.
Pew Research (2025): the chance of a click falls from 15% to 8% when an AI summary is present. Ahrefs (2026): for content at position 1 the click loss reaches 58%.
Yes. AI summaries cite three or more sources 88% of the time (Pew). Content works for brand presence inside the answer and for shaping intent, even without a visit.
Share of answer — the percentage of relevant AI answers in which your brand appears as a source. It rises exactly where CTR falls.
Google users are less likely to click on links when an AI summary appears in the results. 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/ — research report.
2024 Zero-Click Search Study. 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/ — industry research.
AI Overviews Reduce Clicks by 58%. https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/ — industry research.
The Enigma editorial team.
18 May 2026.
18 May 2026.
Peer-reviewed work, preprints and industry research — the type is stated next to each claim.
Every figure re-checked against the primary report, including the US/EU split, which the SparkToro URL slug transposes relative to the published title.
Preprints and industry reports are cited with their methodological limits; verify the conclusions on your own project.
The full list of sources and their reliability levels lives in the research catalogue.