The Generative SERP Shift: What AI Overviews CTR Impact Data Actually Measures

The AI overviews CTR impact data describes a change in SERP real estate, not a collapse in demand. The summary took the space where the click used to happen, and the click moved rather than disappeared. Every published figure measures clicks divided by impressions on one surface, over a period when that surface was rebuilt twice.

• The AI Overview consumes above-the-fold space, pushing the organic result out of view rather than out of the index.
• Google redistributes clicks toward the sources it cites, and those sources overlap with the organic top 10 only about 17% of the time.
• Published decline estimates range from 15% to 61%, and the variable driving the spread is query intent rather than measurement error.
• Demand created inside a zero-click answer returns later through channels organic reporting never claims.

The AI Overviews CTR Impact Data Measures a Surface, Not a Market

Every published study measures clicks divided by impressions on a single surface, and that surface was rebuilt twice during the measurement window. Read the numbers as geometry, not as demand.

The strongest behavioural evidence comes from Pew Research Center, which observed 68,879 real Google searches made by 900 US adults in July 2025. Users clicked a traditional result 8% of the time when an AI Overview was present, against 15% without one. Only 1% clicked a link inside the summary, and 26% ended their browsing session entirely.

Published CTR decline estimates range from roughly 15% to 61% depending on the keyword set. That spread is not contradiction. Studies focused on position one informational queries sit at the high end, blended commercial portfolios at the low end. The variable doing the work is intent, not measurement error.

The denominator moved as well, and three Google-side changes explain why impressions vs clicks 2026 no longer describe the same page:

• AI Mode clicks, impressions and position have been merged into the “Web” search type since 17 June 2025. Two distinct surfaces report as one line.
• The Gen AI performance report ships impressions only. No clicks, no position, no CTR, no queries.
• Google confirmed an impression logging error covering roughly May 2025 to April 2026, affecting impressions only. Clicks were not affected.

When the denominator over-counts and the numerator stays accurate, CTR falls without a single user changing behaviour. A report reading “impressions up, clicks flat, CTR down” across that window is describing an inflated denominator, not a performance problem. Treat any impressions vs clicks 2026 comparison as contaminated until the benchmark is rebuilt on clicks alone.

The AI Overview Occupies the Fold, Which Is Why Zero Click SEO Looks Like a Content Failure

The click disappears because the click target left the screen. Zero click SEO is a layout outcome before it is anything else, and the mechanism is measurable in pixels.

BrightEdge measured the average AI Overview at more than 1,200 pixels of vertical height, against a visible desktop viewport of roughly 600 to 768 pixels. Sitebulb‘s 2026 pixel study found the median number one desktop result now sitting around 1,000 pixels down the page, with more than half of all position one results falling below the fold. Backlinko puts the share of users who scroll to the bottom of page one at 9%.

A page can hold position one, keep its rankings, keep its topical authority, and never enter the field of view. Nothing about the content changed. The page it sits on got taller.

Exposure is also volatile, which makes single-month readings unusable. Semrush, tracking more than 10 million terms, recorded AI Overview prevalence at 6.49% of keywords in January 2025, peaking near 24.61% in July 2025, then settling around 15.69% by November 2025. A portfolio whose CTR moved between those months may have seen no performance change at all, only a change in how often the feature fired.

The Click Was Redistributed, and the AI Overviews CTR Impact Data Shows Where It Went

Google did not delete the click. It moved it toward the sources the summary cites, which are largely not the pages that rank.

BrightEdge measured the overlap between AI Overview cited sources and the organic top 10 at roughly 17%. Around five in six citations come from pages a conventional rank tracker would never surface. Ranking first and being cited have become two different achievements, optimised through two different mechanisms.

That single figure reorganises the whole reporting question. A rank tracker reporting stable positions while clicks fall is not malfunctioning. It is measuring a layer that no longer determines traffic on the affected queries.

Branded Queries Move Against the Trend, Which Is Why a Blended CTR Line Misleads

Two branded metrics move in opposite directions, and averaging them into one organic CTR line destroys the only information the pair contains. Click-through rate on branded queries rises. Branded click volume falls. Both are true at once, which is why a single blended figure describes neither.

A 2025 industry study covering 700,000 keywords found that fewer than 5% of branded keywords trigger an AI Overview at all, and that when one does appear, branded click-through rate rises by roughly 18% on average. Non-branded queries in the same dataset lost close to 20%. The widely quoted blended figure of around 15% is the average of two opposite movements.

The reason is intent. Someone typing a brand name has already made the navigational decision, so the summary confirms the choice rather than substituting for it. Someone typing an informational question has not, so the summary can answer it outright.

The reporting consequence is the point here: a single organic CTR figure averages a segment that improved with a segment that collapsed, and reports the midpoint as the state of the channel. The four-way split that makes the line readable is branded against non-branded, AI Overview present against absent, informational against transactional, cited against uncited. Run those cuts and one falling line usually resolves into two lines going different directions.

The Demand Still Arrives, It Just Stops Looking Like Organic Traffic

Zero-click does not mean zero outcome. The visit happens later, through a channel your organic report does not own.

Loamly‘s 2026 analysis of 446,405 visits found that 70.6% of AI-driven visits arrive with no referrer header, which sends them into the GA4 direct bucket alongside people typing a URL from memory. Google compounds this: AI Overviews and AI Mode pass no distinct referral signal and are bundled into google / organic with conventional search clicks, so there is no supported way to isolate them.

The pattern that results is recognisable once you know to look for it:

• Organic CTR falls on non-branded informational queries, where the summary answers the question outright.
• Branded search volume and direct sessions rise, because the brand was seen inside an answer that produced no click.
• Conversion rate on direct improves, because those users arrive having already done their comparison inside the AI conversation.

A dashboard showing falling branded organic clicks alongside rising direct traffic and improving direct conversion is not describing a brand in decline. It is describing discovery that happened where your tag could not follow. The traffic did not leave. It stopped passing through the door you count.

Reading the AI Overviews CTR Impact Data Correctly Changes What You Fund Next Quarter

The AI overviews CTR impact data documents a redistribution of SERP real estate, not an evaporation of demand. The summary took the fold, clicks moved toward the sources it cites, branded queries held or improved, and a measurable share of the resulting demand now arrives with no referrer attached.

The mistake worth avoiding is rearbitrating a budget on a blended CTR line drawn from a contaminated measurement window. Before that decision, the last twelve months need re-reading segmented: branded against non-branded, AI Overview present against absent, cited against uncited. That reading usually changes the conclusion, and it is the diagnostic the Primelis organic growth team runs before recommending a single structural change, drawing on 15 years of work with more than 300 global brands and over $1B in annual media managed across 200+ specialists.