Cited pages earn 2.3x the clicks of uncited pages on the same query. The full 2026 citation CTR breakdown.
AI Overview citation CTR now splits Google traffic into two tiers. Pages cited inside an AI Overview earn roughly 2.1% click-through rate. Pages that rank on the same results page without a citation earn roughly 0.9%. Seer Interactive measured that gap across 53 brands, 5.47 million queries and 2.43 billion impressions from January 2025 through February 2026. The same dataset shows organic CTR on AI Overview queries climbing 85% in two months, from 1.3% in December 2025 to 2.4% in February 2026. Clicks did not disappear. Google redirected them toward the sources it names.
That reframes the last 18 months of AI search panic. The question stopped being whether AI Overviews cost you traffic. The question is whether Google cites you on the queries your buyers run.
Seer Interactive tracked 53 brands, 5.47 million queries and 2.43 billion organic impressions between January 2025 and February 2026, publishing the third edition of its AI Overview CTR study on April 24, 2026. The agency segmented every query three ways: no AI Overview present, AI Overview present with the brand cited, and AI Overview present with the brand absent from the citation set.
That three-way split matters more than the headline decline numbers most reports quote. Earlier studies compared AI Overview queries against non-AI Overview queries and stopped there. Seer's segmentation isolates the one variable you can influence. You cannot control whether Google shows an AI Overview. You can influence whether Google cites you when it does.
The study also separates paid and organic performance on the same query set, which lets you see how AI Overviews reshape the full search results page rather than one channel in isolation.
Organic CTR on AI Overview queries bottomed at 1.3% in December 2025 and reached 2.4% by February 2026, an 85% jump in two months. Seer's own September 2025 model predicted continued decline. The curve broke the forecast.
Two mechanisms explain the reversal. First, user behavior matured. People who want a one-line answer now stop at the AI Overview and never scroll, which strips low-intent clicks out of the denominator. The searchers who remain want depth, pricing, documentation or a vendor shortlist, and they click. Second, Google widened its citation surface. More inline links inside AI Overviews create more entry points per impression.
Read the rebound carefully. CTR recovering does not mean total organic clicks recovered. Publisher clicks fell 58% after Google made AI answers the default output for every query in July 2026. Rate improved while volume kept shrinking, so the pool of clicks got smaller and better qualified at the same time.
Cited pages earn 2.3 times the click-through rate of uncited pages on the same query. Seer's February 2026 snapshot shows three distinct tiers on the same results page.
| Query condition | Organic CTR (Feb 2026) | Index vs no AI Overview |
|---|---|---|
| No AI Overview present | ~3.3% | 100 |
| AI Overview present, page cited | ~2.1% | 64 |
| AI Overview present, page not cited | ~0.9% | 27 |
Two numbers deserve your attention. A citation recovers about two thirds of the CTR you would have earned on a clean results page. Losing the citation drops you to roughly a quarter. The gap between those two outcomes is the single largest controllable variable in organic search performance right now, and it sits entirely inside your content structure rather than your domain authority.
Track this metric for your brand → nobori.ai
Google confirmed on July 10, 2026 that Gemini 3.5 Flash generates a prose answer for every search query across its global surface, not only inside the opt-in AI Mode tab. Ranked links still appear, positioned below the generated answer with sources rendered as inline citations inside the prose.
That change converts the citation gap from a partial problem into a total one. Seer measured the 2.1% versus 0.9% split when AI Overviews covered a subset of queries. Every query now carries a generated answer, so every query now assigns you to the cited tier or the uncited tier. There is no third bucket left to hide in.
The fallout arrived fast. Publisher clicks fell 58%, antitrust filings piled up, and a German court removed Google's search liability shield. For B2B teams the operational takeaway is narrower: your baseline is no longer "we rank well." Your baseline is your citation inclusion rate, and we broke down how far most B2B brands sit from a healthy number in the B2B AI visibility gap analysis.
Comparison queries trigger AI Overviews about 95% of the time. Question-phrased queries trigger them about 86% of the time. Informational queries sit near 36% and transactional queries near 5%, per Seer's 2026 breakdown.
| Query intent | AI Overview incidence | B2B relevance |
|---|---|---|
| Comparison ("X vs Y", "best X") | ~95% | Vendor shortlist stage |
| Question ("how do I…") | ~86% | Problem education stage |
| Informational | ~36% | Category research |
| Transactional | ~5% | Pricing, demo, signup |
Look at where B2B buying happens. Comparison and question queries carry the shortlist decisions, and both sit above 85% incidence. Your pricing and demo pages still live in the 5% zone where classic blue links dominate. That distribution tells you exactly where to spend: citation work on comparison and question content, conversion work on transactional pages that AI Overviews mostly ignore.
CTR on queries with no AI Overview rose from 2.8% in early 2025 to 3.8% by February 2026, a 36% gain. Those queries became scarcer and more productive at the same time.
The mechanism is straightforward. Google routes quick-answer intent into the AI Overview. Queries that still return a clean results page skew toward tasks a generated paragraph handles badly: navigating to a specific tool, comparing live pricing, downloading a template, reaching documentation. Those searchers arrive with a job to finish.
Most B2B teams have never audited their keyword set for AI Overview incidence, so they cannot tell which half of their portfolio pays 3.8% and which pays 0.9%. Pull your top 200 revenue-adjacent queries, flag which ones return a generated answer, and split your forecast accordingly. Two portfolios with identical rankings can differ 4x in click yield based on nothing but AI Overview coverage.
No. Ahrefs analyzed 863,000 keywords and 4 million AI Overview URLs in March 2026 and found that 38% of cited pages also ranked in the top 10 for the same query, down from 76% eight months earlier. Citations from positions 11 to 100 accounted for 31.2%, and pages ranking beyond position 100 accounted for 31.0%.
Roughly two in three AI Overview citations now come from pages a searcher would never see on page one. Google decomposes each query into sub-queries, retrieves candidates across that wider cluster, then merges results. A page that ranks 40th for your head term can rank first for a sub-query you never tracked, and that sub-query win earns the citation.
This kills the assumption that ranking work automatically produces citation work. We mapped the sub-query mechanics and how to audit coverage in the query fan-out optimization guide.
Semrush examined 3,981 domain appearances across 115 prompts, 14 countries and four engines and found that 61.7% were citations without any brand mention in the answer text. On Perplexity the rate reached 52%. The industry calls these ghost citations: your URL feeds the answer, your name never reaches the reader.
Seer's February 2026 study of 20 brands across 541,213 LLM responses, six platforms and five funnel stages found the reverse relationship too. When a model named a brand in its answer, that brand's citation rate was 53.1%. When the model did not name the brand, its citation rate was 10.6%. The five-fold spread runs against the idea that citations produce mentions. Mentions appear to produce citations.
Seer notes it cannot observe token generation inside the models, so it frames the finding as behavioral evidence rather than proven architecture. The practical read holds either way: optimize for being named, not only for being retrieved.
Ghost citations put a hard ceiling on the 2.1% tier. A cited page only converts its citation into clicks when the reader has a reason to click, and an unnamed source rarely supplies one. Run the arithmetic on 100 citations at a 61.7% ghost rate and you are left with 38 branded appearances doing the persuasion work.
The fix sits in your opening sentences. AI extraction weights the first sentence of each section disproportionately. A generic opener produces a generic paraphrase with no attribution. An opener that carries your brand, your dataset or your named method gives the model an attribution string to reuse.
Three edits move the number. Attribute your own data explicitly in the sentence that states it. Name your product where the product is the answer rather than describing it in the abstract. Put your brand inside table headers and comparison rows, since models lift structured rows close to verbatim. None of this requires new content. It requires rewriting first sentences on pages that already earn citations.
Paid CTR held up better than organic. When Google showed an AI Overview, paid CTR rose from 14.6% to 16.2% across Seer's measurement window. When no AI Overview appeared, paid CTR fell from 26% to 21.8%.
That inversion carries a budget implication most teams miss. AI Overviews push the organic results down the page, which raises the relative prominence of ads on those queries. Meanwhile the clean results pages that reward organic clicks at 3.8% deliver weaker paid performance because the organic listings compete effectively.
So the two channels now prefer opposite query types. Ads perform better where AI Overviews appear. Organic performs better where they do not. If your paid and SEO teams still allocate against a shared keyword list without segmenting by AI Overview incidence, both teams are bidding and writing against the wrong half of the portfolio. Segment the list, then let each channel concentrate where its CTR curve points.
Yes, and by enough to change how you value the remaining traffic. Orbit Media analyzed 97 B2B and lead-generation websites covering 29 million sessions through June 2026 and found visitors arriving from AI sources 3x more likely to convert into leads than other organic traffic. A separate study of 312 B2B firms reported 14.2% conversion for AI referrals against 2.8% for Google organic.
Treat those multiples as directional rather than precise. Attribution for AI referrals stays messy, and only 14% of marketers track AI search as a distinct channel in analytics, so most of these visits land in "direct" or generic "referral" buckets and never get credited.
The combined picture explains why click-rate decline and pipeline growth coexist at the same companies. Fewer clicks arrive. A larger share of them come from someone who already read a synthesized comparison and chose to investigate you specifically. Citation decay complicates the picture further, which we covered in the AI citation half-life analysis.
Google is not the only surface where citation status decides your click yield, and the conversion rates diverge sharply by engine. Semrush's 2026 referral data shows ChatGPT traffic converting at 15.9%, Perplexity at 10.5% and Claude at 5.0%.
Volume moves independently of those rates. US generative AI referral traffic fell from 267.4 million visits in October 2025 to 226.8 million in January 2026, a 15% decline over three months, as more engines answered in place instead of handing off. Citation density differs too. ChatGPT attaches sources to roughly 87% of responses and Google AI Overviews cite in roughly 84.9%, so both surfaces give you an entry point most of the time if you make the source set.
Consistency remains the weak spot for most brands. AirOps found in its State of AI Search 2026 report that only 30% of brands hold visibility across AI answers reliably. The other 70% appear and vanish, which turns citation CTR into a metric that swings week to week rather than a stable line on a dashboard.
Published figures range from 0.61% to 2.4% for the same metric, and the spread comes from methodology rather than contradiction. Compare the inputs before you compare the outputs.
Measurement window drives most of the variance. Seer's earlier study reported organic CTR falling from 1.76% to 0.61% across a window ending in September 2025, which predates the rebound. The 2.4% figure describes February 2026 alone. Both numbers are correct for their windows and useless when averaged together.
Brand set matters next. Seer's panel covers 53 brands weighted toward large advertisers with strong existing rankings, so absolute CTR runs higher than a broad-market sample. Studies reporting a drop from 1.41% to 0.64% use different query mixes with more informational and fewer transactional terms.
Definitions differ last. Some studies count a citation as any appearance of your domain in the source carousel. Others require an inline link inside the generated prose. When you benchmark your own numbers, record which definition you used, then hold it constant. A definition change mid-quarter will look exactly like a performance change.
Report the split, never the blended average. A blended organic CTR number now averages a 3.3% tier, a 2.1% tier and a 0.9% tier into a figure that describes none of your queries and hides every decision worth making.
| Metric | What it answers | Healthy direction |
|---|---|---|
| AI Overview incidence rate | Share of your tracked queries returning a generated answer | Rising; treat as exposure, not failure |
| Citation inclusion rate | Share of AI Overview queries where you are cited | Up |
| Cited vs uncited CTR | Click value of winning the citation | Gap above 2x confirms the lever works |
| Branded mention rate | Share of citations that name you in the answer | Up; inverse of your ghost rate |
| Non-AI query yield | Click value of your clean results pages | Protect; highest CTR tier you own |
Four of those five metrics sit outside Search Console. Google reports impressions and clicks without telling you whether an AI Overview appeared or whether it cited you, so citation data has to come from prompt-level tracking against your own query set.
Start with measurement, then rewrite the pages where a citation is already within reach.
Days 1 to 3: pull your top 200 revenue-adjacent queries and tag each one for AI Overview presence. Split the list into AI Overview queries and clean queries. Forecast each half separately using the 2.1%, 0.9% and 3.8% benchmarks rather than one blended rate.
Days 4 to 7: measure your citation inclusion rate on the AI Overview half. Then measure your ghost rate by checking how many of those citations name you in the answer text. Two brands with identical inclusion rates can differ 3x in brand impact based on that second number alone.
Days 8 to 11: rewrite opening sentences on every page that already earns citations. Attribute your data, name your product where it is the answer, and put your brand into table headers.
Days 12 to 14: audit your comparison and "best" pages against the 95% incidence figure. Those pages face a generated answer on nearly every query they target, and they carry your shortlist decisions. Fix them before anything else in your library.
The rebound suggests a floor rather than a recovery. Two months of CTR growth after 18 months of decline points to users and Google settling into a stable division of labor: the generated answer absorbs quick questions, the links absorb everything that requires a real page.
Three developments will decide whether the floor holds. Google's Search Console controls for generative surfaces, rolling out through 2026 with page-level options scheduled for March 2027, will let publishers opt specific pages out and shift the citation pool. Antitrust pressure and the German liability ruling create legal incentives for Google to link more visibly. And ad density inside generated answers keeps climbing, which competes with organic citations for the same screen space.
Plan for a smaller, better-qualified click pool where citation status determines your share of it. The teams that segment their reporting now will know which half of their portfolio to defend when the next shift lands.
What is AI Overview citation CTR? AI Overview citation CTR is the click-through rate a page earns when Google cites it inside an AI Overview, measured separately from pages on the same results page that Google does not cite. Seer Interactive's 2026 data puts cited pages near 2.1% and uncited pages near 0.9%.
Did AI Overviews stop hurting organic traffic? No. CTR on AI Overview queries rose 85% between December 2025 and February 2026, but total publisher clicks fell 58% after Google made generated answers the default output in July 2026. Rate improved while volume declined.
Does ranking first guarantee an AI Overview citation? No. Ahrefs found in March 2026 that only 38% of AI Overview citations come from pages ranking in the top 10 for that query, down from 76% eight months earlier.
What is a ghost citation? A ghost citation happens when an AI engine uses your page as a source without naming your brand in the answer. Semrush measured a 61.7% ghost rate across 3,981 domain appearances in 2026.
Which queries trigger AI Overviews most in B2B? Comparison queries at roughly 95% and question-phrased queries at roughly 86%, per Seer's 2026 breakdown. Transactional queries trigger them on about 5% of searches.
Can you measure citation inclusion rate in Search Console? No. Search Console reports impressions and clicks without indicating whether an AI Overview appeared or cited you. Citation tracking requires prompt-level monitoring against your own query set.
Ready to see where you stand in AI search?
Nobori tracks your brand's visibility across ChatGPT, Gemini, Perplexity, Google AI Overviews, and Claude, updated daily. See who's getting cited, where you're missing, and what to fix.
Get daily AI visibility alerts for your brand → nobori.ai
