Entity recognition is the automated process of finding named entities — people, places, organizations, products, and concepts — inside page content and labeling what type each one is.
| Term | Entity Recognition |
|---|---|
| Category | Content and Topical Relevance |
| Also known as | Entity Extraction, NER |
| Where it appears | TextRazor enrichment and entity reports |
What it means in RankGear
When RankGear enriches a page through TextRazor, entity recognition is the step that reads the text and pulls out the distinct things it is actually about — not just keywords, but the named people, brands, locations, and concepts, each tagged with a type. Those results feed the entity reports, where you can see which entities your page mentions and how that set compares against the pages already ranking for the query. It turns a wall of prose into a structured list of what a page covers.
How to interpret it
Read the entity set as a coverage signal, not a score to maximize. If the top-ranking pages consistently reference entities your draft never names, that is a concrete gap worth closing — provided the missing entity genuinely belongs in your piece. The point is topical completeness that serves the reader’s intent, not matching a checklist. A page can name every competitor entity and still be thin; a shorter page that covers the entities its audience expects can be the stronger one. Treat the comparison as direction, then decide with editorial judgment.
Example
You publish a guide on cold-brew coffee and RankGear’s entity report shows the ranking field repeatedly surfacing entities like Toddy, nitrogen infusion, chicory, and steeping time — none of which your draft mentions. Your page covers the main phrase but misses the supporting concepts the field treats as part of the topic. You add a genuine section on brewing time and equipment, and the entity set moves closer to the field without a word of padding.
Important considerations
- The entities that appear across ranking pages are comparative indicators of what the topic tends to cover — they are not Google ranking scores, and covering them does not by itself make a page rank.
- Recognition depends on the provider. TextRazor classifies entities against its own models and knowledge base, so type labels and confidence reflect that provider’s scale, not an absolute truth.
- Add entities only where they preserve search intent and genuinely inform the reader. Repetition, stuffing, or forcing terms in to hit a target does not build topical authority.
- Named-entity extraction can misclassify or miss ambiguous mentions; use the report to guide coverage, not to grade prose mechanically.
Related terms
Part of the RankGear glossary · how RankGear measures · the 870 factors.