All terms

Algorithm

Entity salience

Also known as: salience score, entity prominence

Entity salience is a measure of how central an entity is to a piece of content, as opposed to merely being mentioned in it. Natural language processing systems assign a salience score based on signals such as position, frequency, grammatical role and surrounding context. It helps a system decide what a document is actually about.

What it is

Salience is usually expressed as a relative score across the entities found in a document, with the main subject scoring highest and passing references scoring low. Google's Cloud Natural Language API is the best known implementation practitioners test with, returning entities with salience values from its analysis. The score is about prominence within that document, not the entity's general importance in the world.

Why it matters

If a search or answer system reads a competitor, a partner or an unrelated topic as the most salient entity on your page, the page is less likely to be retrieved for the topic you care about. Clear salience helps machines connect a page to the right entity in a knowledge graph, which supports both classic ranking and inclusion in AI answers. It also exposes drift, where a page is stuffed with so many topics that none stands out.

How it works

Practitioners run key pages through an entity extraction tool, check whether the intended entity tops the salience list, then edit to strengthen it: name it in the title, opening paragraph and headings, use consistent naming rather than varied synonyms, and cut tangents that pull attention elsewhere. Internal links, structured data and clear on-page definitions reinforce the association. The test is repeated after editing to confirm the intended entity now dominates.

When it applies

It applies when auditing pages that underperform for their target topic, when building topic clusters and entity-led site structures, and when a brand keeps being confused with a similarly named one. It is most useful on pages with a single clear subject, such as product, service, location or glossary pages.

Examples

  • A service page targeting 'roof insulation' returns a boiler brand as the most salient entity, because a case study quote dominates the copy.
  • A glossary entry is edited to name the term in the title, first sentence and two headings, after which extraction tools rank it first.
  • A local clinic page mentioning three nearby towns is split into separate location pages so each town becomes salient on its own page.

How it is measured

  • Salience score of the intended entity from an entity extraction API, before and after editing
  • Rank of the intended entity within the page's extracted entity list
  • Count of off-topic entities scoring above a set salience threshold on key pages
  • Change in impressions, clicks and assistant citations for the target topic after salience clean-up

Related terms in Algorithm

Primary research · August 2026

How ChatGPT Shortlists Software Brands

An audit across 10 categories and 60 buying questions. I recorded what ChatGPT reads, throws away and links to when a buyer asks it which software to buy, and what that decides.

60
Questions asked
10
Software markets
2,680
Results read
367
Links shown
Free35 pages · PDF · 536 KBDiscovery Digest every Friday

Free download

Get the full report

35 pages · PDF · 536 KB. Enter your details and it downloads straight away.

How ChatGPT Shortlists Software Brands downloads straight away. No spam, unsubscribe anytime.