Entity Density

Entity density is the frequency of named references to specific things (brands, people, tools, companies, locations, data points) within a piece of content. An analysis of 1.2 million ChatGPT citations by Kevin Indig found that cited content averages 20.6% entity density, compared to 5 to 8% in typical English text. High entity density signals specificity and verifiability to AI systems, making content more retrievable and more citable.

Why Entity Density Drives Citations

AI retrieval systems use named entities as anchor points for matching content to queries. A passage containing “Salesforce CRM integrates with QuickBooks, Xero, and FreshBooks for small business accounting” is retrievable for queries about any of those five named entities. A passage containing “the software integrates with leading accounting platforms” is retrievable for none of them because it contains zero named entities.

Every named entity in your content creates a potential retrieval pathway. Higher entity density means more pathways, which means more fan-out sub-queries your content can satisfy. This is why specific, data-rich content outperforms generic marketing language in AI citations: the entities themselves function as retrieval hooks.

Increasing Entity Density

Name specific tools, people, and companies. Replace “industry experts recommend” with “Dan Petrovic at DEJAN recommends.” Replace “leading CRM platforms” with “Salesforce, HubSpot, and Pipedrive.”
Include specific numbers. Replace “significant improvement” with “38% improvement.” Replace “most searches” with “over 60% of Google searches.” Numbers function as entities in retrieval.
Reference specific studies and sources. “Research shows” contains zero entities. “A 2025 Princeton study published in the proceedings of the Web Conference” contains three.
Eliminate zero-entity sentences. Marketing superlatives (“world-class,” “industry-leading,” “seamless”) contain zero entities and zero retrieval value. Every sentence should contribute at least one named reference.

For the complete content optimization framework, see the Generative Engine Optimization guide.