Skip to content

Insights

Cited But Not Clicked: What the Latest AI Overview Data Means for Publisher Traffic

A new DCN study confirms what analytics have been hinting at for months: Google AI Overviews reduce outbound clicks by roughly 40%, even when your content is the source.

The session report looks fine on one slide and soft on the next. Rankings are holding, the content team is producing, and yet Google referral traffic is quietly down. The two reflexes — treat it as a one-month blip, or declare search dead and pivot wholesale — are both ways of avoiding the same conversation. The data now makes that conversation unavoidable.

On September 14, 2026, Digital Content Next published a member study tracking Google Search referral traffic across publishers over an eight-week window. Median year-over-year traffic was down 10% overall: 7% for news brands, 14% for non-news brands. A separate experimental component, run with 1,065 U.S. desktop users, isolated the AI Overview effect directly. When an Overview appeared on the results page, outbound organic clicks dropped roughly 40% and zero-click searches increased roughly 35%. DCN CEO Jason Kint said the losses are a direct consequence of AI Overviews, noting that citations in the summaries recover only a small fraction of the lost clicks. Penske Media’s antitrust filing, a separate action, cites a 58% click decline figure.

Google has since added “Further Exploration” links and subscription labels to AI Overviews. Publisher groups have described those changes as cosmetic. The structural condition — a search result that answers the question on the page — remains.

Being cited in an AI Overview is no longer the same as earning a visit. The click is the gap between those two things, and the gap is now measurable.

For the institutional clients we work with — public media organizations, universities, health systems, nonprofits — this matters in a specific way. Their content strategies were built around Google organic traffic, and most of their current KPIs are still session-based. The DCN study gives a clear, dated reference for why sessions are soft even when rankings look stable: high-traffic informational queries are increasingly resolved inside the Overview, and a citation there does not reliably produce a visit.

The practical response has three parts. First, audit which query types are now triggering Overviews. Not all traffic is equally exposed; navigational, transactional, and deeply specific queries behave differently than broad informational ones. Second, restructure content that earns citations so it also earns direct action — subscription prompts, tool use, downloads, deeper exploration — rather than assuming the click will follow the citation. Third, reset how success is measured. Session counts were never a complete picture; they are now an especially incomplete one. Reach, branded search volume, and downstream conversion give a more honest read of whether the content is working.

The DCN data does not mean search is finished — it means the rules changed, and the earlier metrics no longer describe the game. That is a workable problem, and it starts with knowing what your traffic is actually exposed to.

More dispatches