Entities in Sentences counts how many times RankGear’s recognized entities appear inside complete prose sentences on a page. It returns an occurrence count, not a quality score — it says nothing about whether those mentions are accurate or well written.
| Factor ID | RG-ENT-012 |
|---|---|
| Family | Entities |
| Measurement | Occurrence count |
| Measured zone | Parsed sentences from cleaned page content |
What it measures
Entities in Sentences counts entity occurrences within the sentences RankGear recognizes as prose. Rather than scanning raw text across the whole document, it isolates the entities — the people, places, organizations, products, and concepts in the current run’s entity set — that appear inside readable sentences, and tallies every occurrence. The result is a single count that reflects how much the page’s running prose actually names and discusses those entities.
How RankGear measures it
RankGear segments the cleaned page content with its sentence tokenizer, lowercases the resulting corpus, and then sums word-aware matches for each entity in the run. Word-aware matching means an entity is counted where it appears as a whole word within a sentence, and every matching occurrence adds to the total.
| If the count is… | What it suggests |
|---|---|
| Higher | The page discusses its entities in sentence form, explaining them in context rather than only listing or labelling them. |
| Low or zero | Relevant entities may be missing, or they appear only in lists, headings, tables, or markup that the tokenizer does not read as sentences. |
How to optimize it
Explain the relevant people, places, organizations, products, and concepts in complete, useful sentences that establish their relationship to the topic. Because the factor reads prose, entities that live only in bullet lists, tables, headings, or markup may not be captured — writing them into full sentences is what registers. Treat the count as an observation of how thoroughly your page explains its entities, not a target to inflate: repeating the same name adds occurrences without adding meaning.
Important considerations
- Formatting affects sentence segmentation. Punctuation, line breaks, and list structure change what the tokenizer treats as a sentence, which changes the count.
- Repeated mentions increase the count. Every occurrence of an entity is summed, so naming the same entity several times raises the total.
- The factor does not evaluate factual correctness. It counts occurrences only; it does not check whether what a sentence says about an entity is true.
- Entity availability depends on the run. The set of entities RankGear matches against comes from the current run, so the same page can produce different counts across runs.
- Correlation is not causation. This is a prioritization signal that relates to position within a measured result set — a higher count does not make a page rank or guarantee a better position.
Related factors
Part of the Factors reference · how RankGear measures · glossary.