How often do AI engines cite pages that don’t rank?
- 61.7% of AI chatbot citations point to URLs that do not appear anywhere in the organic top 100 — so-called ghost citations. The figure comes from Profound’s analysis of 340 million prompts. This is the single strongest data point for the claim that AI citation behavior and organic ranking are different systems.
- Top-10 organic pages’ share of Google AI Overviews citations fell from 76% to 38% between July 2025 and March 2026, per an Ahrefs study of 863,000 keywords. Ahrefs itself flags that part of the measured drop reflects improved parsing methodology in the newer study, not purely a change in Google’s source selection — the direction is confirmed, the exact magnitude may be overstated.
- After Gemini 3 took over Google AI Overviews globally on 2026-01-27, SE Ranking’s 100,000-keyword US study found 42.4% of previously cited domains dropped out and 51.7% of newly cited domains were new entrants, while average sources per AI Overview rose from 11.55 to 15.22. Vendor study; directional.
- Google AI Mode replaces 56% of cited domains per week, per SISTRIX’s citation-drift study. For identical local queries, SE Ranking measured 65% domain volatility between consecutive AI Mode runs (5,000 keywords, 5 cities, 15 parse cycles).
- AI Mode generates 8–12 sub-queries per standard query, and only 27% of those sub-queries are stable across repeated runs, per Ekamoira’s query fan-out research. A single visibility probe is therefore a weak measurement instrument on this surface.
What has AI search done to clicks and traffic?
- Position-1 organic CTR on AI-Overview keywords fell from 7.3% (December 2023) to 1.6% (December 2025) in Ahrefs’ 300,000-keyword study (150,000 AIO keywords against a 150,000-keyword control). Ahrefs puts the AI-Overview-attributable reduction at 58%, comparing the actual 1.6% against a 3.7% counterfactual without AI Overviews.
- The flip side: pages actually featured in an AI Overview saw a 91% CTR uplift versus the same organic position without an AIO, per Seer Interactive’s study of 53 brands and 5.47 million queries.
- 68.01% of US Google searches ended in zero clicks in SparkToro’s 2024 clickstream analysis — a pre-AI-Mode baseline worth keeping in view.
- Google search referrals to publishers fell 33% globally from November 2024 to November 2025 across 2,500+ sites, per the Reuters Institute/Chartbeat data reported by Press Gazette. A separate Chartbeat data pull, reported in March 2026, showed small publishers down 60% versus 22% for large publishers — these are two independent releases, not one study, and the small-publisher timeframe is disputed in secondary coverage.
- Baseline before AI enters the picture: 96.55% of pages get zero organic traffic from Google, per Ahrefs’ 14-billion-page study — December 2023 data, never updated, and sampled toward the quality side of the web. And only 1.74% of new pages reach the Google top 10 within a year across all URL types (6.11% for the non-empty English-content subset).
Where do the AI engines actually get their sources?
- 87% of ChatGPT citations match Bing top-20 organic results, per Backlinko’s citation study. That coupling is loosening: Profound’s 240-million-citation dataset shows ChatGPT citation alignment with Google SERPs rising from 12% to 33% between April and July 2025 while Bing alignment fell from 26% to 8%.
- Google properties account for 17.42% of all AI Mode citations as of February 2026, up from 5.7% in June 2025, per SE Ranking’s study of 1.32 million citations across 68,313 keywords.
- Community platforms (Reddit, Quora) captured 52.5% of AI citations versus 47.5% for brand-owned domains in Otterly.ai’s report on 1 million+ citations (January–February 2026). The split varies sharply by engine: brands took 44.7% of ChatGPT citations, 28.9% of Perplexity citations, and 59.8% of Google AI Overviews citations.
- By source ownership, Yext’s 50,000-query, 8-industry study found brand-owned pages account for 52.15% of Gemini AI Overview citations, while third-party pages account for 48.73% of ChatGPT citations.
- Concentration is severe: the top 15 domains capture 68% of consolidated AI citation share across 680 million+ citations synthesized from six studies in the 5W AI Platform Citation Source Index 2026.
- Perplexity averages 21.87 citations per response versus 7.92 for ChatGPT, per Profound’s 680-million-citation dataset. Perplexity’s own internal data, reported by TechCrunch, puts its citation rate 37% higher for content published within 48 hours — a vendor-internal figure, treat as soft.
- Since 2026-05-06, Google AI Overviews directly quote Reddit threads and forums with named attribution, per TechCrunch’s reporting on the update — a structural change in how forum content surfaces, not just a citation-share shift.
What content characteristics correlate with getting cited?
- The foundational academic result: GEO techniques produced up to a 40% visibility boost across AI engines in the Princeton GEO paper (KDD 2024). Within that study, adding statistics to content lifted citation likelihood by 25.9% — a figure often misquoted as 41%; the paper’s table shows 25.9%.
- The GEO-16 framework study (1,702 citations, three engines) found pages scoring at least 0.70 with 12 of 16 quality pillars hit reached a 78% cross-engine citation rate; the odds ratio for citation from a higher score was 4.2 (95% CI 3.1–5.7). The strongest pillars: metadata and freshness, semantic HTML, structured data.
- 38% of AI citations are pulled from a page’s first 100 words, nearly double the 20% a year earlier, per Surfer’s two-wave study (100,000 citation placements, June 2026 wave). Front-loading the answer is getting more important, not less.
- SE Ranking’s 129,000-domain, 216,524-page analysis of ChatGPT citation factors reports concrete thresholds: sections of 120–180 words averaged 4.6 citations versus 2.7 for sub-50-word sections; articles over 2,900 words averaged 5.1 versus 3.2 for sub-800-word pieces; pages with expert quotes averaged 4.1 versus 2.4 without; pages with 19+ discrete statistics averaged 5.4; content updated within 3 months averaged 6.0. The same study found plain declarative headings outperformed question-format headings.
- On layout: SALT.agency’s 2,318-URL pixel-depth study found no correlation between on-page position and Google AI Mode citation likelihood — citations averaged 2,400–4,600 pixels deep depending on vertical. AI Mode selects passages by relevance, not screen position.
- Google states on the record that no special structured data is required for AI Overviews or AI Mode citations, in its AI features optimization guide. Correlationally, Yext measured a 54.53% citation rate for pages with structured data versus 38.12% without — vendor data, correlation not causation.
- Review presence compounds: in Seer Interactive’s 800,000-response study, brands with no Trustpilot profile saw a 1% AI citation rate, a minimal profile (1–13 reviews) 53.5%, and a strong profile (80+ reviews with owner responses) 75.3%.
Does AI referral traffic actually convert?
- ChatGPT referral traffic converts at 15.9% versus 1.76% for organic search, per Orbit Media’s 2026 analysis. The overall AI channel in the same dataset converted at 10.21% versus 2.46% for non-AI direct traffic, across 446,000 AI visits.
- Measurement undercounts it: an industry analysis of 446,000+ visits found 70.6% of AI-driven traffic arrives with no referrer header and lands in analytics as “Direct.” Google AI Mode applies noreferrer to outbound links outright.
What is the scale of these surfaces?
- Google AI Mode reached 1 billion monthly active users within 12 months of its May 2025 US launch, per Google’s own announcement. ChatGPT stood at 900 million weekly active users as of May 2026.
- Meanwhile, 73% of sites analyzed had a technical barrier blocking AI crawler access entirely (robots.txt, CDN/WAF rules), per Otterly.ai’s citation report. And the major AI crawlers — GPTBot, ClaudeBot, PerplexityBot — execute no JavaScript, so client-side-rendered content is invisible to them.
How should these numbers be read?
Three cautions apply across this entire page. First, most large citation datasets come from vendors selling visibility tooling; their figures are directional evidence, not neutral measurement, and are flagged above where relevant. Second, churn rates of 42–65% per measurement window mean any single snapshot is already aging — every figure here carries its measurement date for that reason. Third, several widely circulated numbers in this field failed our verification and are deliberately absent from this page; where a statistic is missing, that absence is usually the finding.
Sources
- Profound — Ghost Citations report: https://tryprofound.com/blog/ghost-citations-report
- Profound — AI Search Shift: https://www.tryprofound.com/blog/ai-search-shift
- Ahrefs — AI Overviews reduce clicks (update): https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/
- Ahrefs — Search traffic study: https://ahrefs.com/blog/search-traffic-study/
- Ahrefs — How long does it take to rank: https://ahrefs.com/blog/how-long-does-it-take-to-rank-in-google-and-how-old-are-top-ranking-pages/
- Search Engine Journal — AIO citations from top-ranking pages drop: https://www.searchenginejournal.com/google-ai-overview-citations-from-top-ranking-pages-drop-sharply/568637/
- Search Engine Journal — ChatGPT citation factors (SE Ranking study): https://www.searchenginejournal.com/new-data-top-factors-influencing-chatgpt-citations/561954/
- Search Engine Journal — Small-publisher referral decline: https://www.searchenginejournal.com/search-referral-traffic-down-60-for-small-publishers-data-shows/569959/
- Search Engine Land — Google AI Mode citing Google more: https://searchengineland.com/google-ai-mode-citing-google-more-study-471042
- Search Engine Land — AI Mode above-the-fold study: https://searchengineland.com/google-ai-mode-above-the-fold-content-study-468578
- SISTRIX — AI citation drift: https://www.sistrix.com/blog/ai-citation-drift-how-stable-are-sources-in-ai-search-results/
- SE Ranking — AI Mode volatility test: https://seranking.com/blog/ai-mode-volatility-test/
- Ekamoira — Query fan-out research: https://www.ekamoira.com/blog/query-fan-out-original-research-on-how-ai-search-multiplies-every-query-and-why-most-brands-are-invisible
- SparkToro — Zero-click search: https://sparktoro.com/blog/less-than-half-of-google-searches-now-result-in-a-click/
- Press Gazette — Google traffic down (Reuters Institute/Chartbeat): https://pressgazette.co.uk/media-audience-and-business-data/google-traffic-down-2025-trends-report-2026/
- Seer Interactive — AIO impact on CTR: https://www.seerinteractive.com/insights/aio-impact-on-google-ctr-2026-update
- Seer Interactive — Reviews in AI search (800K responses): https://www.seerinteractive.com/insights/study-of-800k-ai-responses-how-reviews-shape-brand-presence-in-ai-search
- Backlinko — ChatGPT search citations study: https://backlinko.com/chatgpt-search-citations-study
- Otterly.ai — AI Citations Report 2026: https://otterly.ai/blog/the-ai-citations-report-2026/
- Yext — AI Overview source distribution report: https://www.yext.com/blog/2026/the-ai-overview-source-distribution-report
- 5W — AI Platform Citation Source Index 2026: https://www.prnewswire.com/news-releases/5w-releases-ai-platform-citation-source-index-2026-the-50-websites-that-now-decide-what-brands-are-visible-inside-chatgpt-claude-perplexity-gemini-and-google-ai-overviews-302759804.html
- Discovered Labs — Platform citation differences (Profound data): https://discoveredlabs.com/blog/chatgpt-claude-perplexity-and-google-ai-overviews-how-each-platform-cites-sources-differently
- TechCrunch — Perplexity freshness data: https://techcrunch.com/2026/02/freshness-perplexity-citations/
- TechCrunch — Google adds forum quotes to AI search: https://techcrunch.com/2026/05/06/google-updates-ai-search-to-include-expert-advice-from-reddit-and-other-web-forums/
- Princeton GEO paper (KDD 2024): https://arxiv.org/abs/2311.09735
- GEO-16 framework study: https://arxiv.org/abs/2509.10762
- Surfer — AI citations and the first 100 words: https://surferseo.com/blog/ai-citations-first-100-words/
- Google — AI features optimization guide: https://developers.google.com/search/docs/fundamentals/ai-optimization-guide
- Orbit Media — AI traffic and dark analytics: https://www.orbitmedia.com/blog/ai-traffic-dark-analytics/
- FoundryCRO — Tracking AI search referrals: https://foundrycro.com/blog/tracking-ai-search-referrals/
- Google — AI Mode scale announcement: https://blog.google/products-and-platforms/products/search/search-io-2026/
- Google — I/O 2026 keynote: https://blog.google/technology/ai/google-io-2026-keynote/
- Passionfruit — JavaScript rendering and AI crawlers: https://www.getpassionfruit.com/blog/javascript-rendering-and-ai-crawlers-can-llms-read-your-spa
- Allmo — llms.txt citation analysis: https://allmo.ai/articles/llms-txt