Leadership1 publisher2 min readPublished
A 4-million-URL study puts AI Overview citation overlap with the top ten at 38%
The number comes from a Forbes Tech Council post by a marketing agency director, and it is the only measurement in the record behind the claim that rank tracking and AI citation tracking now count different populations.
The Board Room · Leadership desk
What happened
- A study of 4 million AI Overview URLs found that 38% of the pages Google cited also appeared in the top 10 organic results for the same query.
- Roughly a third of those citations went to pages that never reached the top 100, and some of the cited pages had no meaningful organic presence at all.
- Jangid wrote that most marketers assume AI recommendation engines mirror the organic results page, so a top-10 ranking will earn a citation.
- Analyses on his firm's RankRabbit platform put the AI assistant as the first and sometimes only touchpoint before a consumer calls a business, making the recommendation the conversion event.
Compiled by The Board RoomSomething wrong?How this is made
Why it matters
- constraint Rank tracking can only report on the 38% of cited pages that sit in the top ten; the other 62% fall outside what the tool measures, so a green dashboard describes a shrinking part of how buyers get shortlisted.
- exposure For multilocation brands, inconsistent business listings remove them from consideration before ranking is weighed at all, and the owner of that data is usually operations, not the content team.
- decision Jangid does not argue that ranking stopped mattering, so citation work has to be funded alongside the existing search programme instead of out of it, and somebody has to say which line item shrinks.
- contradiction Google's own description of query fan-out supports the direction of the finding, while its size rests on a study quoted second-hand and on the author's own platform data.
Thirty-eight percent of cited pages sat in the top ten for the same query [1]. Roughly a third of citations came from pages that never cracked position 100 [2]. That leaves about 29% of citations coming from pages ranked somewhere between 11 and 100 [13].
Jangid attributes the gap to a retrieval technique Google has described in public. In its announcement of AI Mode, Google described the technique as "breaking down your question into subtopics and issuing a multitude of queries simultaneously on your behalf." That, by Google's account, lets Google Search dive deeper into the web than a single traditional query could [3]. One question becomes a cluster of subquestions, each retrieved separately, and the answer is assembled from whichever pages perform best across the expanded set [4]. A page can hold position one for its head term and be passed over because it never addresses the adjacent questions the system decided were also relevant. A page on the fourth results page can win a citation for answering one of those subqueries precisely [5].
Provenance is worth stating before anyone moves money. The piece is a Forbes Tech Council contribution by Hastimal Jangid, Director at Coozmoo, an AI-powered digital marketing agency serving small and medium-sized businesses [7]. It does not name who conducted the 4-million-URL study. The claim that the recommendation itself has become the primary conversion event for many local and service businesses comes from his own firm's analyses of its RankRabbit platform [8]. So the number points at the remedy its author sells, and the record does not say what share of any given company's buyers are asking an assistant at all. Direction is the steadier part: Google's own account of fan-out explains why citation and ranking would come apart even if 38% is the wrong figure.
The two responses on offer differ in size. Auditing AI citations separately from rank tracking is analyst time, and Jangid recommends it alongside optimizing topic clusters instead of single keywords [10]. Rebuilding content production so that individual passages stand alone as complete, extractable answers is a multi-quarter programme [16], and it competes for budget with the ranking work that traditional SEO signals still serve: backlinks, domain age, keyword density [11].
The audit is the decision available this quarter. Run against your own query set, it shows where your own overlap sits against 38% [1] [10]. Next quarter the content roadmap gets written either from that figure or from a study of somebody else's queries.
What to watch
- Publication of the 4-million-URL study with a named author and method, which would let the 38% overlap figure be checked against its query set.
- A second measurement of AI Overview citations, from a party that does not sell the remedy, reproducing or contradicting a top-ten overlap near 38%.
- Any first-party citation reporting from an assistant operator that lets a brand see its own share without buying a tracking tool.