Entity-to-Token Ratio: Packing Your Content with Citable Facts
How named entities (people, places, numbers) affect whether AI cites your content.
What it measures
Entity-to-Token Ratio counts how many named entities — people, places, brands, organizations, numbers, and dates — appear relative to your total word count. A higher ratio means your content is rich in specific, citable information rather than vague statements.
Why it matters for AI
AI systems like ChatGPT, Claude, and Perplexity extract named entities to build knowledge graphs and generate answers. When your content is packed with specific facts, AI has concrete data points to quote and cite. Content full of vague generalities gets skipped in favour of pages with real information.
| Entity Ratio | Score | What it signals |
|---|---|---|
| < 5% | Low | Too vague — mostly filler with few specific facts |
| 5–10% | Average | Some specific information but could be richer |
| ≥ 10% | Good | Information-dense — AI can extract many data points |
How to improve
- Replace vague phrases with named sources — Instead of "experts say", write "Dr. Jane Smith, Professor of AI at MIT, says"
- Add specific statistics and percentages — "87% of marketers" is far more citable than "most marketers"
- Mention real people, companies, and locations — Named entities give AI concrete reference points
- Include dates and numbers — "In Q3 2024, revenue grew 23%" is more useful than "revenue grew recently"
💡 Quick win
Find your three vaguest paragraphs and add at least one named entity to each — a person, company, statistic, or specific date.