AI answers are cutting search traffic in a measurable way: Pew Research found that Google users clicked a result link on just 8% of visits where an AI summary appeared, against 15% where none did, and only 1% of searchers clicked a link inside the summary itself. That is roughly half the click rate on results pages the answer sits above. The measurement, not the hype, is the story.
Search marketing trends now pivot on one question: what happens to clicks when the answer sits on the results page? Eighteen percent of the searches in Pew's March 2025 sample produced an AI summary, so the exposure is no longer an edge case. This guide sets out what the data shows and which content plays hold up. We covered a connected angle in Best Marketing Conferences — Marketing Conferences Calendar: Which Are Worth the Badge.
What Did the Pew Study Measure?
Pew's analysis covered 900 U.S. adults who agreed to share their browsing activity, and about six in ten of them ran at least one Google search in March 2025 that returned an AI-generated summary. The comparison is clean: search pages with a summary sit against search pages without one, same users, same month.
Users who saw a summary clicked a traditional result in 8% of visits. Those who did not see one clicked nearly twice as often, at 15%. The summary's own links earned clicks in just 1% of visits. Pew also logged what happened next: 26% of sessions ended entirely after a page with an AI summary, against 16% on pages with plain results.
How Much Screen Do the Answers Take?
The feature behind the numbers is AI Overviews, which Google launched in the United States in May 2024 and expanded globally by October 2024. As of May 2025, the feature runs in more than 200 countries and territories and more than 40 languages. In Pew's sample the median summary ran 67 words, and 88% cited three or more sources.
The space math is blunt. Research from Botify and Demandsphere, from December 2024, found an AI Overview stacked with a featured snippet fills about 67.1% of a desktop results screen and 75.7% on mobile. A page can rank first and still sit below the fold, because the answer block carries the real estate.
Which Content Strategies Still Earn Clicks?
The citations point one way. Summaries cite their sources, so being one of the three or more cited pages beats ranking third. Original data, named methodology, and dated facts give a model something solid to quote, and verifiable pages are the ones that get named inside the answer. Readers following this should also see Journalist Source Verification: How Journalists Actually Verify Sources in 2026.
The traffic that stays is traffic an answer cannot compress. Queries that need a booking, a price, a tool, or a local result still end in a click, and brand searches still route to the brand. Publishers who measure position trackers alone will miss the shift; measuring clicks by landing page shows what AI answers actually took.
What Does This Cost a Publisher?
For a page that lives on informational queries, halving the click rate is a revenue line, not a talking point. Ad impressions on search-referred sessions fall first. The honest adjustment is budget, not panic: content that only restates facts is the cheapest to replace and the first to lose its traffic.
The spend that still pays is the spend a summary cannot copy. Original surveys, proprietary data sets, calculators, and reporting with a named human behind it cost more to produce, and they keep earning citations. The 88% multi-source citation rate means there is a slot for cited pages even in a zero-click world.
Conclusion: Count Clicks, Not Rankings
Ranking stopped being the score the day the answer moved above the results. Pew's 8% against 15% is the first solid loss number, and the 26% session-end rate shows readers leaving the search loop entirely. The strategies that adapt treat AI answers as a citation network: feed them facts worth quoting, keep the pages that need a human, and judge the work by clicks, not position.
