Co-citation

Co-citation happens when two sources are cited together by a third. The idea comes from academic bibliometrics: if many papers cite A and B together, A and B are probably about related things, even if they never cite each other. In AI search, MLforSEO uses the term for brands or pages that an LLM cites together in the same answer.

How it works

Henry Small proposed co-citation analysis in 1973 as a way to map how research fields relate. In traditional SEO, the term came to describe two sites being mentioned or linked together on other pages, implying a topical connection. In LLM-driven search, when a system synthesises an answer and cites several sources, those sources are co-cited: the system has judged them relevant to the same query and complementary to each other.

MLforSEO’s view is that repeated co-citation places your brand in an “entity neighbourhood”. Appearing consistently alongside recognised authorities makes your entity easier to cite confidently, while appearing alone, or only beside weaker entities, gives the system few cues about where you belong.

How it’s used in SEO and marketing

• Auditing which brands and domains AI assistants cite alongside you for your priority prompts.
• Spotting the authorities you are not yet associated with, then creating content that genuinely compares, references or works with them.
• Earning mentions on third-party sources (comparisons, research roundups, industry lists) where you appear next to established entities.
• Formalising relationships with schema properties and consistent entity relationships.
• Tracking change over time as part of AI share of voice.

Example

Someone asks an AI assistant about email deliverability best practices. The answer cites a Mailgun engineering post, an EmailToolTester comparison, a Litmus research report and your guide to SPF setup. If that pattern repeats across related prompts, you are co-cited with recognised authorities in email infrastructure. Specific topics produce sharper co-citation sets than generic ones, which favours brands with focused, niche authority.

Limitations and things to watch

• Co-citation is an observed pattern. No AI provider documents it as a ranking signal. Nobody outside the AI providers knows how much past co-citation influences future answers.
• Answers vary between runs and platforms, so measure across many prompts and repeat over time.
• Being co-cited with competitors is not a bad sign; it often means you are in the right neighbourhood.
• Avoid manufactured associations. Thin pages that name-drop big brands do not create genuine relationships.

Learn it on MLforSEO

• How LLMs Co-Cite: Building Authority by Association
• Why Entity Relationship Mapping Is Critical for Brand Visibility
• LLM Citation Tracking: 4 Step Implementation Guide (resource).
• AI Search & LLMs: Entity SEO and Knowledge Graph Strategies for Brands course, lesson on co-citation in LLM reputation management.