The citation pattern our audits keep finding
Across audits, the domains cited for buying questions cluster into a handful of types, in roughly this order of frequency. Review and comparison platforms: the G2s, Capterras, TripAdvisors and Healthgrades of each vertical. Community discussion: Reddit threads and niche forums where practitioners and customers speak candidly. Encyclopedic and reference sources, Wikipedia above all, for entity facts. Major and trade media for claims and news. Category listicles and best-of roundups on credible blogs. And official sources, brand sites and documentation, mostly for facts about the brand itself rather than recommendations.
The mix shifts by engine and category: Perplexity favors fresh media and comparison content, Gemini leans on Google-adjacent signals like Business Profiles, and health or finance questions pull heavier on institutional sources. But the shape, reviews plus communities plus references plus listicles, holds remarkably steady.
Why these domains win the citations
They share three properties assistants can verify. Independence: they are not the brand talking about itself, so they corroborate rather than assert. Structure: reviews, ratings, comparison tables and threaded discussions are machine-parseable by design. And density: one listicle answers a whole shortlist question, which is exactly the shape assistants need to compose an answer. Your own site rarely tops the citation list for recommendation questions, and it does not need to: it wins the factual questions about you, while third parties carry the recommendation.
Sources cited in AI Overviews specifically
Google's AI Overviews show their links, which makes the pattern easy to check for your own queries: run your buying questions and log the linked domains. Expect the same categories with a Google accent, sites already ranking for the query, Business Profile data for local intent, and YouTube appearing more than in any chat assistant. Overview citations rotate as rankings shift, which is why monthly logging beats a one-time look.
How to use this, and how to check your category
The generic pattern is a starting map; your category's actual list is knowable in an afternoon. Ask the four assistants your buyers' questions, collect every cited domain, and count frequency. The result is your outreach plan, ranked: the review platform where your profile is thin, the two listicles that omit you, the community where your competitors' customers post and yours do not. The free Aethon audit produces exactly this citation map for your category, with the answer screenshots behind it.