AI Overviews have made organic click-through rate harder to interpret. One study says the top-ranking page loses roughly half its clicks, while another shows organic CTR recovering after a sharp decline. A behavioral study finds that people click traditional results almost twice as often when an AI summary is absent.
Those findings look contradictory only when the methodology disappears. They do not measure the same population, ranking position, device, time period, SERP layout, or type of click. AI Overviews CTR is therefore becoming a benchmark that is easy to quote and easy to misuse.
There is no single “AI Overview CTR loss.” The number changes with ranking position, citation status, query intent, device, SERP layout, and the methodology used to measure it.
A 50% decline can accurately describe one slice of search while badly misrepresenting another. The question is no longer simply whether AI Overviews reduce clicks, but under which conditions the loss becomes meaningful.
The Headline Number Depends on the Study
The broad direction across current research is consistent: AI Overview click-through rates tend to be lower than CTR on comparable SERPs without the feature. The size of that effect varies considerably, and that variation is not necessarily a research problem. It is often a comparison problem.
| Study | Headline Finding | What It Measures |
|---|---|---|
| Ahrefs | ~58% lower CTR | Estimated effect on the top-ranking page for AIO-triggering informational keywords |
| Advanced Web Ranking | Position 1: 20.02% → 9.69% | Position-level CTR on Google US desktop in July 2026 |
| Seer Interactive | AIO organic CTR rebounded from 1.3% to 2.4% | Large Search Console query cohort tracked over time |
| Pew Research Center | 8% clicked traditional results with AI summary vs 15% without | Actual browsing behavior across real user search sessions |
Ahrefs compared 300,000 keywords and estimated that the presence of an AI Overview correlated with roughly 58% lower CTR for the top-ranking page in December 2025. Advanced Web Ranking used a different method and found a 52% gap at position one in its July 2026 US desktop sample. Both studies describe click suppression, but they arrive there through different datasets and assumptions.
Seer Interactive approached the question through Search Console data across millions of queries. Its early-2026 update showed AIO-present organic CTR rising from about 1.3% in December 2025 to about 2.4% by February 2026. Pew did not calculate rank-specific CTR at all; it observed what real users did after almost 69,000 Google searches.
All four datasets are useful, but combining them into one average “AI Overview CTR loss” would remove the conditions that produced each number. The same issue appears in other benchmark work: a current Microsoft Advertising benchmark and an older platform average can both be accurate within their original datasets while answering different questions.
AI Overviews Still Reduce Organic Clicks
Different methodologies do not erase the larger pattern. Current datasets consistently show more zero-click behavior and weaker traditional organic click-through rates when AI-generated answers appear, even though the size of the decline varies.
Advanced Web Ranking reported a 74.3% zero-click rate on AI Overview SERPs in its July 2026 desktop sample, compared with 53.6% where no AI Overview appeared. Its position-one CTR fell from 20.02% across all SERP types to 9.69% on AI Overview SERPs. Those figures describe a substantial redistribution of attention before the user reaches the traditional result.
Pew reached the same broad direction from user behavior. People clicked a traditional result on 8% of visits where an AI summary appeared, compared with 15% where it did not, while only about 1% of visits produced a click on one of the sources cited inside the summary. These are different measurements, but both suggest that part of the search need can now be satisfied before an organic listing earns the click.
Why AI Overview CTR Changes by Ranking Position
AI Overviews do not affect every ranking position equally. Advanced Web Ranking’s July 2026 US desktop data shows particularly large differences near the top of the results, where organic clicks have historically concentrated.
| Organic Position | All SERPs | With AI Overview | Relative Difference |
|---|---|---|---|
| 1 | 20.02% | 9.69% | -52% |
| 2 | 10.36% | 3.69% | -64% |
| 3 | 3.89% | 1.70% | -56% |
Source: Advanced Web Ranking’s 2026 organic CTR study.
This changes forecasting because a page ranking first cannot assume that a historical position-one CTR curve still describes an AIO-heavy SERP. A page at position eight already receives a much smaller share of available clicks, so the absolute loss can look very different even when the relative effect is substantial.
Stable ranking therefore no longer implies stable CTR. The amount and type of content occupying the space around the organic result can materially change the click opportunity, which makes rank-only reporting less informative than it used to be.
Being Cited Changes the Outcome
An AI Overview can reduce the opportunity for a traditional organic click, but citation status changes the size of the loss. Seer’s full-year informational-query averages were 2.07% when the brand was cited, 0.94% when it appeared on an AIO SERP without a citation, and 3.35% when no AIO appeared.
The uncited number needs context because Seer notes that one large account represented 47% of impressions in that segment. Excluding that account would raise the aggregate uncited CTR to roughly 1.61%. The difference is large enough to show how strongly dataset composition can affect a headline average.
Citation Helps, but It Does Not Restore a Clean SERP
Seer’s data still suggests that citations matter. Cited AIO queries performed substantially better than uncited ones in the aggregate, yet remained below queries where no AI Overview appeared. Citation therefore looks less like a return to the old SERP and more like partial protection within the new one.
A useful sequence is AIO presence → citation status → impressions → clicks → query value. Treating every AI Overview impression as the same event hides too much of what determines whether the visibility is actually useful.
Seer’s AI Overview CTR research also shows why CTR alone can mislead. The rate may decline while total clicks remain flat if impressions grow quickly enough, so the denominator matters as much as the percentage itself.
Query Intent Changes the Economic Meaning
Not every lost click has equal business value. In Advanced Web Ranking’s July 2026 position-one data, informational CTR fell from 17.69% across all SERPs to 10.88% with an AI Overview. Commercial queries showed a larger gap, from 11.91% to 3.51%, while transactional queries were more resilient at 10.38% versus 7.8%.
Seer’s dataset measured exposure differently and found AI Overviews much more common on informational searches. They appeared on about 36% of informational queries in its sample, compared with 8% of commercial and 5% of transactional queries. The studies should not be merged into one intent benchmark because their classifications and samples differ, but the business pattern is still useful.
A site may lose a large share of clicks on definitions, comparisons, or early research queries while retaining most of its revenue-producing transactional demand. Another business may depend on that early research traffic to create future customers, making the same decline far more consequential. Sitewide CTR can therefore hide the difference between traffic loss and business loss.
The principle is similar in paid acquisition, where Google Ads lead generation benchmarks become more useful once the business asks what happens after the initial conversion. A front-end rate tells only part of the story when the downstream value differs by query or audience.
Desktop and Mobile Should Not Be Blended Blindly
Device segmentation has become more important because desktop and mobile did not move in parallel during 2026. Advanced Web Ranking recorded a sharp desktop decline during the spring, with position-one desktop CTR across organic results falling from 45.6% in February to 14.6% in May before recovering somewhat.
Mobile moved differently over much of the same period, rising from around 30% late in 2025 to roughly 39% by June. AIO-specific July figures also differed by device: desktop position-one CTR with an AI Overview stood at 9.69%, while mobile was around 12%.
Those observations do not prove that device alone caused the divergence. Interface tests, query mix, AI Mode integration, the amount of answer content visible before scrolling, and other SERP changes can all influence behavior. They do show why a blended average can hide materially different desktop and mobile behavior.
A report saying “organic CTR fell from 4.8% to 3.6%” is therefore incomplete until the device split is checked. The same site can appear stable in aggregate while desktop and mobile are moving in opposite directions.
Other SERP Features Can Compound or Change the Effect
AI Overviews rarely occupy an otherwise empty results page. People Also Ask, local packs, videos, images, product modules, sitelinks, and knowledge panels all compete for attention, but those features do not push CTR in the same direction.
Answer-oriented combinations tend to absorb clicks, while some media-rich layouts can create additional entry points and stronger engagement. Advanced Web Ranking’s feature-combination analysis makes that variation visible rather than treating every AIO SERP as one category.
An AI Overview by itself was associated with a 36% reduction from its position-one baseline. AI Overview plus People Also Ask showed a 59% decline, while adding videos to that combination pushed the reported loss to 81%. The most crowded combination in the dataset — AI Overview, local pack, products, People Also Ask, and videos — left the first traditional organic result with only 0.15% CTR.
Other layouts behaved differently, particularly where images or videos created additional visual opportunities. The practical point is not that every added feature damages CTR; it is that “AI Overview present” can be too coarse a description of the SERP a user actually sees.
Why One 2026 Dataset Showed CTR Recovering
The early-2026 Seer rebound prevents the story from becoming a simple downward line. Organic CTR on queries where an AI Overview appeared rose from roughly 1.3% in December 2025 to 2.4% in February 2026, an increase of about 85% from the trough.
Queries without an AIO still performed better, and later datasets continued to show substantial suppression, so the result should not be described as a universal recovery. Google’s interface can change, as can the queries that trigger AI Overviews, citation distribution, seasonality, user behavior, and the composition of the websites being tracked.
The defensible conclusion is narrower: one large cohort recovered materially after an unusually weak period. That is enough to make both extreme interpretations risky, because one declining month does not prove that organic search is collapsing and one rebound does not prove that users have adapted.
Real User Behavior Confirms the Zero-Click Pressure
Search Console and rank-tracking datasets show impressions and clicks, while Pew’s behavioral research adds a different layer: what users actually did after seeing the results. Its study followed browsing activity from 900 US adults during March 2025 and covered 68,879 unique Google searches.
When an AI summary appeared, users clicked a traditional result in 8% of visits. Without one, that happened in 15%. Users also ended their browsing session more often after an AI summary: 26% of visits compared with 16% on pages without one.
The Pew Research Center study supports the zero-click interpretation without pretending to provide a rank-one CTR benchmark. A behavioral panel and a Search Console dataset can reinforce the same broader story while still measuring different things.
Search Console Can Now Separate Generative AI Visibility
Measurement became more useful in 2026 because Google introduced dedicated generative AI reporting in Search Console. Google announced the feature in June and says it had rolled out globally by August 31, 2026.
The reporting provides a dedicated view of visibility from generative search features such as AI Overviews and AI Mode while keeping that activity within the broader Search performance environment. This should not be mistaken for a perfect standalone AI Overview CTR report, because analysts still need to connect that exposure to rankings, clicks, query intent, citation status, and business outcomes.
Google’s Generative AI performance reporting nevertheless removes some of the guesswork that previously forced site owners to infer AI-related changes entirely from ordinary Search Console trends. Dimensions such as pages, countries, devices, and dates now provide a more useful starting point for separating generative visibility from the rest of search performance.
A Falling CTR Does Not Automatically Mean SEO Is Failing
CTR is produced by both the ranking and the environment around that ranking. Suppose impressions remain stable, important positions hold, generative AI visibility grows, and CTR falls; that pattern can point toward a changing SERP rather than deteriorating SEO execution.
Now change the pattern. Rankings fall, impressions contract, and CTR declines at the same time. AI Overviews may still be present, but attributing the entire loss to them would hide a ranking problem.
CTR therefore needs the metric before it and the business outcome after it. The same logic applies in other channels: a high Instagram click-through rate does not automatically make a campaign successful, while a lower rate is not automatically weak.
| Pattern | Possible Reading |
|---|---|
| Impressions stable, rankings stable, CTR down, AIO visibility up | SERP behavior may be absorbing clicks |
| Rankings down and CTR down | Ranking loss may be contributing materially |
| Cited queries outperform uncited AIO queries | Citation may be preserving part of the click opportunity |
| Desktop declines while mobile holds | Device mix and interface may matter |
| Informational CTR falls but transactional demand holds | Traffic loss may be larger than revenue impact |
| Impressions rise while clicks stay flat | CTR can fall even as visibility expands |
These patterns are diagnostic rather than causal proof. Their purpose is to narrow the investigation instead of assigning every CTR decline to AI Overviews, and they work best when viewed alongside ranking, intent, device, and conversion data.
A Better 2026 CTR Benchmark Is Conditional
The old question “What is a good organic CTR?” is becoming too broad to answer usefully. A useful benchmark now needs conditions around the number rather than a single curve applied to every query.
Ranking position, AI Overview presence, citation status, query intent, device, and surrounding SERP features all change the meaning of CTR. Industry and branded versus unbranded demand can shift it further.
That makes benchmarking less convenient but more informative. AI Overviews have clearly changed the distribution of clicks, particularly on informational and feature-heavy SERPs, but they have not created one universal percentage that every site should subtract from historical organic traffic.
The practical shift is not from “CTR matters” to “CTR no longer matters.” It is from treating CTR as a standalone performance score to interpreting it in the context of the ranking, the query, and the SERP that produced it.