How AI Overview Citations Impact Your Traffic

What changed with AI Overview CTR in 2026?

AI Overview CTR did not move in a straight line through 2025 and into 2026: it fell hard for most of the year, then bent back upward. Ahrefs and Seer Interactive each dated one half of that swing, with real numbers attached to both the fall and the rebound.

France's AI Overviews launch on July 22, 2026 delivered the shock: Ahrefs tracked 963 of the country's highest-click domains for the 28 days before launch and the 9 days after. Across the full panel, CTR fell 3.4%. Domains with heavy exposure, meaning AI Overviews appeared on more than a fifth of their tracked queries, lost 23.1% of their CTR almost immediately, or 24.8% once adjusted against the control group. A low-exposure control group, under 2% AI Overview coverage, gained 2.3% over the same window.

The relationship was close to linear: roughly one point of CTR lost for every point of AI Overview exposure, and 82.2% of highly exposed domains lost CTR outright. Nearly half of them, 48%, lost more than 10% of their daily clicks, one in five lost more than 30%, and one in twenty lost more than half.

Exposure itself keeps expanding, too: the median AI Overview rate across Ahrefs' panel sat at 5.4%, but 29% of domains had already crossed 10% exposure before the window closed, putting the group facing double-digit CTR risk closer to the near-term majority than the outlier.

Then came the recovery: Seer Interactive tracked 53 brands across 5.47 million queries and 2.43 billion organic impressions from January 2025 through February 2026. Organic CTR on AI-Overview queries bottomed at 1.3% in December 2025, then climbed to 2.4% by February 2026, an 85% jump in two months. Part of that is a rising tide: CTR on queries with no Overview at all grew too, from 2.93% in January 2025 to 3.97% by December.

Who does the AI Overview CTR drop hit hardest?

Exposure is not evenly spread across query types, and that alone predicts who feels the drop. Comparison queries trigger an AI Overview about 95.4% of the time, review queries 86.3%, and question-format queries close behind at 85.9%, while transactional queries trigger one far less often. By intent the gap runs wider still: informational queries show an Overview 36% of the time against 8% for commercial searches and 5% for transactional ones.

Citation status matters more than ranking position once an Overview appears on a query. A page the Overview actually cites earns 120% more organic clicks per impression than a page it leaves out, even though cited pages still underperform a non-Overview result by 38% on informational searches. Ranking near the answer without being named inside it now captures a fraction of what that ranking used to be worth.

The pain is not evenly split by content status either. Brand pages the Overview cited saw impressions surge through the back half of 2025, which dropped their measured CTR 61% in the third and fourth quarters, but that decline traces to more impressions chasing a similar click count, not to clicks disappearing. Pages that went unnamed fared worse in a way that matters more: organic CTR on not-cited pages fell 67% across informational queries over the same period.

Sector matters too: in the Ahrefs France panel, health queries carried a 22.2% AI Overview rate and a 20.4% CTR reduction, among the steepest drops recorded. A team in a heavily summarized vertical such as health or finance should treat the exposure numbers above as a floor for its own traffic, not a ceiling.

How should you respond to the AI Overview CTR shift?

The old ranking playbook will not fix a citation problem. What moves CTR now is whether the model quotes a site inside the answer, a function of content quality and freshness rather than position on a results page. That 120% click advantage for cited pages is the whole argument for treating citation, not rank, as the target metric.

Practically, that means every piece needs something a model can lift whole: a sourced statistic, a comparison table, an original data point, or a dated case result. Pages built only to rank, with nothing citable inside them, are exactly the ones that lost two-thirds of their CTR on informational queries once an Overview appeared, per Seer Interactive's tracking above.

The response looks slightly different depending on who is asking. A B2B SaaS team worried about a 'best [category] tool' prompt cares about citation share on comparison prompts specifically, the query type sitting at 95.4% Overview coverage above. A local or multi-location business cares more about '[service] in [city]' prompts and AI Overviews presence tied to calls and bookings. The data above applies to both, just at different query formats.

Omnicite tracks this directly as Citation Share, the percentage of relevant AI answers in a category that cite a given site, measured across ChatGPT, Perplexity, Gemini and AI Overviews rather than Google alone. That breadth matters because the France and Seer Interactive numbers above cover Google's Overview specifically. A site optimizing only for that one surface stays exposed everywhere else a buyer might ask.

LeadHaste, the company behind Omnicite, ran its own site through this approach and went from 0 to 1 million impressions with 200-plus AI citations, while lifting Domain Rating from 1 to 24, in four months. The mechanism was publishing volume and freshness at a scale an in-house team rarely has time for, not a trick aimed at the ranking algorithm. Omnicite does not promise a specific citation count or ranking position to any client; the guardrail is the same one that makes citation-grade content work in the first place: real sourced claims, published often enough that a model has something current to quote.

Two actions apply regardless of vertical. Audit which existing pages are cited versus merely mentioned on your top query set, since that split now predicts CTR better than position does. Then route new content toward the query formats sitting above 85% Overview coverage first, comparison and question-format pages, because citation, not rank, decides who gets the click there.

Will AI Overview CTR keep recovering, or is this a plateau?

Neither study promises a straight line forward. Seer Interactive's own figures run through February 2026 with projections only to March, and the recovery so far is modest in absolute terms: a move from 1.3% to 2.4% CTR, not a return to pre-Overview norms. France's numbers describe nine days after a launch, not a settled state.

What the two studies agree on is the mechanism, not just the direction. Exposure predicts loss in something close to a straight line, and citation predicts recovery almost as cleanly. A site that publishes little, or publishes without sourced facts a model can quote, has no lever to pull once an Overview lands in its category. A site publishing citation-grade material at volume has the one lever both datasets confirm: getting named inside the answer.

The practical read for a moment like this one is simple. Do not wait for Google to settle into a final CTR floor before acting. Exposure keeps climbing (comparison and question queries already sit above 85%), and the sites showing up inside the Overview today are the ones that will hold citation share when the next market gets the feature, the way France just did.

Sources

Source: Seer Interactive

Organic CTR on AI-Overview queries rose 85%, from a 1.3% floor in December 2025 to 2.4% in February 2026, across 53 brands and 5.47 million tracked queries Seer Interactive, 2026-05-21

Source: Seer Interactive

Pages cited inside an AI Overview get 120% more organic clicks per impression than pages the Overview does not cite Seer Interactive, 2026-05-21

Source: Ahrefs

France's AI Overviews launch on July 22, 2026 cut CTR 23.1% at highly exposed domains within 9 days, while low-exposure domains gained 2.3% Ahrefs, 2026-09-16