Study: LLMs Cite a Narrow Platform Set; Wikidata Yields 2.8× Citation Boost
Three independent datasets converge on the same structural finding: LLMs draw citations from a dramatically narrower platform set than the open web would predict. Profound's 680M-citation analysis identifies Reddit, Wikipedia, Stack Exchange, G2, Capterra, YouTube, and LinkedIn as dominant sources — Reddit receives 17× fewer Google visits than the average top-20 domain but 3.5× more LLM citations. First Page Sage's 36,000-query buying-intent study names G2 as the single most-cited domain for B2B software queries inside ChatGPT. ConvertMate's 2026 GEO benchmark shows Wikidata-indexed companies receive 2.8× more AI-assistant citations than peers without entries. Meanwhile, NP Digital's January 2026 survey of 200 companies attributes 92% of measured organic traffic decline to Google AI Overviews.
Why It Matters
The citation graph and the backlink graph are now structurally divergent: optimizing for Google rankings does not optimize for LLM citation share. For B2B SaaS companies, a Wikidata entry and strong G2 review velocity are now measurable distribution levers — not just brand hygiene.