Statistics · Click-through impact

AI search CTR statistics

How click-through behavior changes when an AI Overview or AI-generated answer appears — with the specific population each figure describes attached, not stripped away.

Ritik Namdev Ritik Namdev ·Published September 2026 ·Population-qualified throughout ·11 min read
The short version

Click-through to organic results is reported to drop meaningfully when an AI Overview appears for a query — but the specific percentage circulating varies by study and population, for exactly the reasons documented in the provenance audit. This page states what's actually measured, qualified by population, rather than restating a bare number.

Headline numbers

38%

of AI Overview citations come from Google's top-10 pages, down from ~76% a year earlier — the clearest, best-sourced trend in this category.

Ahrefs, 2026

Why CTR is hard to measure honestly here

Most CTR figures in this space blend at least two different things. One: does a page still get clicked when it's also cited inside the AI answer? Two: does the page get clicked at all when an AI answer exists and the page didn't make it into the citation? These are different outcomes with different implications. Aggregate "CTR dropped by X%" figures rarely specify which one they describe.

The four outcomes a single "CTR" number blends

Cited in AI answer?Query has AI answer?What "CTR" means here
YesYesDoes being cited still drive a click, or does the citation itself satisfy the user?
NoYesDoes a page lose clicks simply because an AI answer exists, even without being cited at all?
YesNoThe classic baseline — a normal organic click with no AI answer present.
NoNoAlso the classic baseline — not ranked, not cited, no AI answer involved.

A single headline "CTR dropped 40%" figure almost never specifies which of the top two rows it's actually describing. A site owner's response to each is completely different, though. The first is a citation-value question: is being cited still worth something, even without a click? The second is a pure-loss question: a page that used to get organic clicks now gets fewer, for a reason unrelated to its own ranking quality.

Citation position and clicks

Does being cited earlier or more prominently within an AI answer drive more clicks than being cited later? That's the AI-search analog of classic SERP position. It's an intuitive hypothesis. But no rigorous public measurement of it exists that we could locate. Hypothesis pending direct study.

The classic-SEO CTR curve, as a point of contrast

Classic SEO has a well-documented, widely replicated finding here. Click-through rate drops sharply and predictably as ranking position moves from #1 down through the results page. Many independent studies have produced broadly similar curve shapes, over many years.

That maturity is exactly what AI-search CTR measurement currently lacks. No comparably rigorous, independently replicated curve exists yet for citation position within a generative answer. Does an equivalent, similarly steep curve exist for AI citation position? Or does the underlying dynamic, users reading a synthesized answer rather than scanning a ranked list, behave fundamentally differently? That's a genuinely interesting open question. This field hasn't answered it yet with the rigor classic SEO eventually achieved for its own ranking curve.

Most AI-search CTR figures blend two different outcomes: 'clicked despite being cited' and 'clicked because not cited at all.' Different questions, routinely reported as one number.

Share on X

Full table

StatisticPopulationGrade
AIO citation-to-rank overlap declineQueries triggering AI OverviewsTraceable
"AI Overviews cut clicks by 58%"Conflated across multiple distinct studies — see the zero-click pageBroken chain

How to measure your own CTR impact

Filter Search Console data to queries you know trigger AI Overviews. Use the AI Overview Exposure Checker to find them. Then compare click-through on those specific queries against a comparable set that doesn't trigger one. A first-party comparison is more reliable for your own site than any aggregate external figure.

A worked example of the segmentation

Here's how, concretely. Export Search Console query-level data for the trailing 90 days. Tag each query as "AI-Overview-triggering" or not, using the checker tool above, or manual spot-checks for your highest volume queries. Within the AI-Overview-triggering group, split further by whether your page appears among the AI Overview's citations for that query.

You now have three groups: cited-and-triggered, not-cited-but-triggered, and a non-triggering baseline. Comparing CTR across all three gives a genuinely diagnostic picture. A meaningful gap between the first two groups suggests citation itself still carries real click value. A similar drop across both, relative to the baseline, suggests the mere presence of an AI Overview reduces clicks regardless of citation status. A single blended average across all AI-Overview-triggering queries, without this split, would hide exactly this distinction.

What to actually do about a real drop

If your own segmented data shows a genuine loss, two responses are worth trying before assuming nothing can be done. First, check whether your page is cited at all for the query. An uncited page facing an AI Overview has the weakest position of anyone in the table above, so getting cited is the first fix to try.

Second, look at what the page still offers beyond the answer already summarized in the AI Overview. A page that only restates the same short answer gives a reader no reason to click through. A page that adds a calculator, a comparison table, or a deeper worked example gives them one.

Verification status

The one strong figure here is the AIO rank-overlap trend — Traceable . Bare "CTR dropped by X%" claims are treated as Broken chain as a category until a population-qualified, disclosed-method version exists.

How to cite this
Namdev, R. (2026). AI search CTR statistics (v1). Retrieved from https://ritiknamdev.com/blog/ai-search-ctr-statistics

Published under CC BY 4.0 — reuse freely with attribution.

Related work on this site

Read alongside zero-click search statistics for the full disambiguation of the figures most commonly conflated in this space.

§ References

Sources

Figures attributed to third parties above have not been independently verified unless stated otherwise.

FAQ

Frequently asked questions

What's the single most reliable way to measure my own AI-search CTR impact?
Segment Search Console data by query. Isolate queries where an AI Overview is known to trigger. Compare click-through on those queries against comparable queries without one. A first-party comparison beats any external aggregate figure for your specific site.
Does citation position within an AI answer affect click-through the way SERP position does?
Plausible by analogy to classic search. But no rigorous public study directly measures this for AI answers specifically. It's registered as an open question, not assumed to transfer directly.
Why do so many published figures blend the different outcomes together?
Mostly because the available data, aggregate Search Console exports, generic vendor reports, doesn't naturally separate "cited and clicked less" from "not cited and clicked not at all." That takes deliberate, careful segmentation. A single blended number is also simpler to publish as a headline than the harder-to-summarize real breakdown.
Is a CTR drop always bad for a site, even when the site is the one being cited?
Not necessarily. A citation inside an AI answer can still deliver brand exposure, and sometimes referral value, even without a click. See the zero-click search statistics page for that distinction in more depth. A CTR-only view can understate the total value of being cited, even while it counts fewer clicks accurately.
Ritik Namdev
Written by

Ritik Namdev

Growth · SEO · GEO

Growth marketer documenting a brand-new site's climb into Google and the AI engines - in public, with real numbers. Every tactic here is tested on real sites before it's published.

The Lab · Weekly

One experiment. Every week.

The field notes in your inbox - one thing I tested, the raw numbers behind it, and what it means for getting cited by AI.

Free forever. Unsubscribe anytime.