Entity Salience: The Ranking Factor Most SEOs Overlook

Most SEO practitioners focus on whether a keyword appears on a page. Google is more interested in whether the page is about the concept behind the keyword. That distinction is entity salience: a numerical score between 0 and 1 that reflects how central any named entity is to the document as a whole.

A page about entity salience in SEO should have entity salience as its highest-scoring entity. If it does not - if the scores are distributed across schema markup, Google, SEO audit, and keyword research without a dominant signal - Google's algorithm has less confidence in the page's relevance for queries about entity salience specifically.

What Is Entity Salience in Google's Algorithm?

Entity salience is a score Google's Natural Language Processing pipeline assigns to every entity it detects within a document. Google has made this score directly observable through its Cloud Natural Language API, which returns entity names, types (Person, Organisation, Location, Event, Work of Art, Consumer Good, Other), and a salience score for any text you submit.

A salience score of 1.0 would mean the entire document is about a single entity with nothing else mentioned. In practice, scores above 0.5 for a target entity indicate strong topical focus. Scores below 0.1 indicate that the entity is present but peripheral.

Google's patents and research papers describe a system that uses entity salience to understand what a document is about at a deeper level than keyword matching. A page that mentions "XML sitemaps" twice in passing has low salience for that entity. A page that defines XML sitemaps, explains how they work, describes how to submit them, and answers common questions about them has high salience - even if the exact phrase "XML sitemaps" appears less often than on a keyword-stuffed competitor page.

How Does Google Calculate Entity Salience?

Google has not published its exact salience algorithm. The Natural Language API demo provides a close approximation of what Google can compute internally. Based on the API's behaviour and the available academic research on entity salience, the key inputs appear to be:

Position and frequency of entity mentions

Entities mentioned early in the document - in the title, the first paragraph, and the first subheading - receive a higher positional weight than entities mentioned only in closing sections. Frequency matters too, but frequency without positional prominence produces a lower salience score than an entity that appears consistently throughout the document from the introduction onward.

Co-occurrence with related entities

Google's entity understanding is relational. When the Natural Language API scores a document, it considers which other entities appear alongside the primary entity. A page about entity salience that also mentions the Knowledge Graph, Google's NLP API, named entity recognition (NER), and the entity grid model signals - through entity co-occurrence - that it belongs in a specific part of the knowledge graph. This co-occurrence pattern increases the model's confidence in both the topic and the expertise of the source.

Sentence-level entity centrality

At the sentence level, entities that appear as the grammatical subject of informational sentences receive higher salience weight than entities that appear as objects or prepositional phrases. "Entity salience tells Google what your page is about" centres the entity differently than "you should consider entity salience." The first construction has higher entity centrality and contributes more to the overall salience score.

Document-level entity density

Salience is a proportional measure. If your page contains 50 named entities and spreads attention across all of them equally, each receives a low salience score. If your page centres on one primary entity and uses 15 to 20 supporting entities that define, contextualise, or relate to it, the primary entity's score rises. Focused documents score higher for their primary entity than unfocused documents, even when total word count is similar.

Does Entity Salience Affect AI Overview Citations?

AI Overviews in Google Search and citation selection in ChatGPT and Perplexity all require the same underlying condition: the retrieval model must be confident that a given page is genuinely about the query topic. High entity salience for the target entity is a proxy for that confidence.

In systematic testing across informational queries, pages with dominant entity salience for their primary topic appear in AI Overviews at a higher rate than pages where the same keyword appears frequently but salience scores are distributed. The mechanism is not surprising: an AI model extracting a passage to cite benefits from a document where one concept is clearly central, because the relevant passage is easier to locate and the citation is less likely to be challenged as off-topic.

This creates a concrete connection between entity salience optimisation and both traditional rankings and AI citation performance. The same content improvements that raise your salience score - topical focus, entity co-occurrence patterns, answer-first writing - also improve your visibility in AI-generated summaries.

How Do You Increase Entity Salience for Your Target Topics?

Increasing entity salience is not about repeating a keyword more often. It is about structuring your document so that your primary entity is unmistakably central to every major section. These are the techniques that produce measurable salience improvements when tested with the Natural Language API.

Name the entity explicitly in the first sentence

The opening sentence of your page should name the primary entity clearly and position it as the grammatical subject. "Entity salience is a score Google assigns to entities in a document" is a stronger opener for salience purposes than "In semantic SEO, there are many factors that affect rankings, including a concept known as entity salience." Both mention the entity, but the first establishes it as the central subject immediately.

Use the entity as the subject of informational statements throughout

Every major section should contain at least one sentence where the primary entity or a direct synonym is the grammatical subject of a declarative statement. This is the sentence-level centrality signal described in the calculation section above. Review each H2 section and verify that the first paragraph centres the entity, not a supporting concept.

Build a supporting entity cluster

Identify the 10 to 15 entities that belong in the same knowledge graph neighbourhood as your primary entity. For entity salience, those supporting entities include: Natural Language API, named entity recognition, Knowledge Graph, entity grid, topical authority, semantic SEO, entity prominence, co-occurrence, information extraction, and document understanding. Weaving these into the document - where they genuinely aid explanation - creates the co-occurrence pattern that reinforces your primary entity's salience score.

Remove or demote off-topic entities

Every unrelated entity you mention is a vote against the salience of your primary entity. A guide to entity salience that also contains a long section on backlink building, social media strategy, or paid advertising will score lower for entity salience than a guide that stays within the semantic neighbourhood. If off-topic sections are necessary, contain them in a brief paragraph rather than a full H2 section, and do not feature off-topic entities in headings.

Structure headings around the primary entity

Headings carry additional weight in entity scoring. H1 should name the primary entity. H2s should frame questions or statements where the primary entity is the logical subject. "How Does Google Calculate Entity Salience?" scores higher for entity salience than "The Google Calculation Method," because the first puts the entity in the heading explicitly rather than relying on inference.

Entity Salience vs Entity Prominence: What's the Difference?

Entity salience and entity prominence are related but operate at different scales, and conflating them is a common source of confusion in semantic SEO discussions.

Entity salience is a page-level signal. It measures how central an entity is to a single document. You control it entirely through content decisions: how you open sections, which entities you include, and how you structure sentences. A brand-new page about a niche concept can achieve high entity salience on its first day of publication.

Entity prominence is a Knowledge Graph-level signal. It measures how significant and well-attested an entity is across the entire web - the volume of references to it, the authority of those references, the consistency of the entity's name and attributes across different sources, and whether Google has a Knowledge Panel for the entity. Prominence builds slowly through editorial mentions, structured citations, Wikipedia presence, Wikidata entries, and consistent brand signals over months and years.

Attribute	Entity Salience	Entity Prominence
Scope	Single page	Entire web / Knowledge Graph
You control it by	Writing focused content	Building consistent citations over time
Time to impact	Hours (on next crawl)	Months to years
Measured with	Natural Language API	Knowledge Graph presence, brand SERPs
Primary use in ranking	Document relevance for specific queries	Entity authority across all queries

The practical implication: a publisher with high entity prominence for their brand - a well-known SEO tool, a major news outlet, a recognised individual expert - can rank a page on a new topic more easily because their entity prominence provides a credibility baseline. A smaller publisher can compensate with exceptionally high entity salience for their target topics, producing content focused enough that Google ranks it despite lower overall prominence.

How to Measure Entity Salience on Your Pages?

There are two practical methods for measuring entity salience on existing content. Both require submitting the page's text to Google's Natural Language API.

Google's Natural Language API demo

Google provides a free demo at cloud.google.com/natural-language where you can paste the full text of any page and see entity detection results including salience scores. This is the most direct measurement available. Copy your page's visible body text, paste it into the demo, and examine which entity has the highest salience score. If your target entity does not score highest - or if it scores below 0.3 - the document has a salience problem that content restructuring can fix.

The Entity Analyzer tool

The Entity Analyzer on this site submits a live URL's content to the Natural Language API and returns a ranked list of entities with their salience scores, types, and Wikipedia links. It also compares your primary entity's salience against a sample of the top-ranking pages for your target query, giving you a competitive benchmark rather than an absolute score in isolation.

Run this audit quarterly on your highest-priority pages, and immediately after any significant content update, to verify that additions have not diluted your primary entity's salience.

For a complete approach to semantic SEO that combines entity salience with knowledge graph optimisation, internal linking strategy, and topical authority building, see the Semantic SEO hub and the Knowledge Graph Optimisation guide.

Frequently Asked Questions

Related Guides

• Semantic SEO: Hub Overview
• Knowledge Graph Optimisation
• ChatGPT SEO: Get Cited in ChatGPT Answers
• XML Sitemaps: Build, Submit and Optimise

Tools on This Site

• Entity Analyzer - score entity salience on any live URL
• LLM Mentions Tracker - monitor AI citations by entity