Losing clicks to competitors? Find out exactly why, and what to fix first.

Show me why
Skip to content

On Page SEO

How to Use Entity Salience for SEO Content Optimization

Add saadrazaseo.com as a preferred source on Google

Free and no obligation. A real reply from me, not a bot.

Short answer: Entity salience is how central a person, place, brand or concept is to a piece of text. Google’s Cloud Natural Language API scores it from 0 to 1. To use it for SEO, give each page one clear main entity, name it in the title, heading and opening sentence, make it the subject of your sentences, describe it through its attributes and related entities, and back it up with internal links and structured data.

0 to 1.0The range of the salience score in Google’s Natural Language API; closer to 1.0 means more central to the documentGoogle Cloud docs, 2026
2014Google researchers introduced entity salience as a task: giving each entity in a document a relevance scoreGoogle Research, 2014
500 million+objects in Google’s Knowledge Graph at its 2012 launch, described as “things, not strings”Google blog, 2012

This guide is for writers, editors and SEOs who keep hearing about entities and want something practical to do with the idea. You will learn what entity salience actually measures, what it can and cannot tell you about rankings, how to test a draft in a few minutes, and the edits I make when a page’s main topic comes out weaker than it should.

What is entity salience?

An entity is a distinct, identifiable thing: a person (Saad Raza), a place (Karachi), an organisation (Google), a product (iPhone 17) or a concept (technical SEO). Salience is a linguistics term for prominence. Put together, entity salience is a measure of how central each entity is to the meaning of a document as a whole.

The idea comes from Google’s own research. In 2014, Google researchers Jesse Dunietz and Dan Gillick published a paper introducing the entity salience task: assigning a relevance score to every entity in a document, trained on documents paired with their summaries. The public version you can test today is the analyzeEntities method in Google’s Natural Language API. Its API reference describes salience as a score in the 0 to 1.0 range, where scores closer to 1.0 are highly salient, and attaches a Wikipedia URL and Knowledge Graph MID to many entities when they are available.

An important caveat

The Cloud API is a public product, not Google Search’s ranking system. Google has never said that Search uses this exact score, and a page will not rank because it hits 0.8. I treat the score as a diagnostic: a quick way to see whether a machine reading my page agrees with me about what it is mainly about, much like checking how Google might connect it to the Knowledge Graph. If it does not, readers skimming the page probably feel the same confusion.

Why entity salience matters for SEO

When Google launched the Knowledge Graph in 2012, its announcement described the shift as “things, not strings”: understanding real-world entities and the relationships between them rather than matching words. Since then, search has leaned further on language understanding. That has three practical consequences for content:

  • Relevance is judged on meaning. A page that clearly centres on one entity, and describes it with the attributes people expect, is easier to match to the full range of queries about that entity.
  • Ambiguity costs you. “Jaguar” could be an animal, a car brand or a football club. Related entities around the main one remove that doubt.
  • AI answers quote clear statements. Systems that summarise pages pick up sentences where the subject is explicit. “Entity salience is…” is easier to extract than “It is…”, which is why it matters when you optimise content for AI Overviews.

Salience fits inside the wider approach I describe in my guides to semantic SEO and entity-based SEO. Salience is the page-level part: making sure each individual page is unmistakably about its own entity, so that together they build topical authority around a subject.

Rather have an expert do this for you?

Hand it over and get back to running your business.

Hand it over to me

Entity salience vs keyword density

QuestionKeyword densityEntity salience
What it measuresHow often an exact string appears relative to word countHow central a concept is to the whole text
Counts synonyms and pronouns?NoYes, mentions of the same entity are grouped
Affected by sentence structure?NoYes, subjects and early mentions weigh more
Can be gamed by repetition?Easily, and that is keyword stuffingMuch less; repetition alone moves it little
Best useSpotting overuseChecking a page’s focus and topic clarity

If you still use density checks, my article on keyword density explains why it is only useful as a warning for extremes.

What raises an entity’s salience

From testing many drafts in the Natural Language demo, and from the factors described in the original research, these are the signals that consistently move scores:

  1. Position: entities named in the title, first heading and opening sentence score higher than those introduced halfway down.
  2. Grammatical role: an entity that is the subject of sentences (“Entity salience measures…”) outranks one that only appears as an object or in a subordinate clause.
  3. Mentions across the text: named mentions, common-noun references and pronouns all count, but a spread through the document beats a cluster in one paragraph.
  4. Clear naming: using one consistent name. Switching between “GBP”, “Google Business Profile” and “Google My Business” can split one entity into several weaker ones.
  5. Relationships: describing the entity through its attributes and links to related entities (what it does, who makes it, where it is) reinforces it as the centre of the text.

How to use entity salience for content optimisation, step by step

Step 1: Decide the one leading entity for the page

Write it down before you draft. One URL, one main entity. If you cannot choose between two, you probably need two pages, which also avoids keyword cannibalisation.

Step 2: List the supporting entities and attributes

Search the main query and note which related entities the top results, “People also ask” and Google’s knowledge panel connect to it. For a page about a dental implant service: implant, titanium, crown, bone graft, dentist, the city, price, recovery time. These become the vocabulary of the page.

Step 3: Draft with the main entity as the grammatical subject

Open with a definition sentence that names the entity first. Use it as the subject of topic sentences in each section. Keep one consistent name and introduce any abbreviation once.

Step 4: Test the draft

Paste the main content (not the menu or footer) into the free Natural Language API demo and open the Entities tab. Check three things: is your main entity at the top, is it typed correctly (for example ORGANIZATION or OTHER), and is it linked to the right Wikipedia page where one exists?

Step 5: Fix what the test shows

Typical fixes are moving the main entity into the opening sentence, rewriting sentences so it is the subject, cutting tangents that introduce bigger rival entities, and merging inconsistent names. Re-test after each round.

Step 6: Reinforce it outside the copy

Point internal links at the page with descriptive anchors, and add structured data. Google’s structured data introduction says markup gives “explicit clues about the meaning of a page”. A Service, Product, Organization or Article type with about and mentions properties makes the main entity explicit; my schema markup generator builds the JSON-LD.

Worked example: rewriting an opening paragraph for higher salience

This is an illustrative test, so the exact scores you see will differ with the full text.

Before: “Many businesses in Karachi want more customers from Google. There are lots of ways to do this, and one that helps is making sure your listing is set up correctly, because it shows on Maps.”

Here the most salient entities are “businesses” and “Karachi”. The thing the page is actually about, Google Business Profile, is never named.

After: “A Google Business Profile is the free listing that shows your business on Google Search and Maps. For businesses in Karachi, an accurate profile with the right category, hours and photos is the quickest way to appear when nearby customers search.”

Now the profile is named first, is the subject of the opening sentence, and sits next to its attributes (free listing, Search, Maps, category, hours, photos). In the demo, that change should move it to the top of the entity list.

Mistakes I see when people chase salience

  • Treating the score as a target. Pushing a number up by repeating the entity reads badly and is just keyword stuffing with a new name.
  • Testing the whole HTML page. Navigation, sidebars and footers add entities that distort the result. Test the main content only.
  • Long introductions about something else. Opening with the history of the internet makes “internet” more salient than your topic.
  • One page covering several products or services. Each competes with the others and none comes out clearly on top.
  • Inconsistent brand naming. Your organisation’s name should be spelt the same on every page, in schema and on profiles, so it resolves to one entity.

Check your understanding

What does a salience score close to 1.0 mean in Google’s Natural Language API?

  • The entity is mentioned more than any other word
  • The entity is highly central to the document as a whole
  • The page will rank well for that entity

Salience measures centrality to the document. It is not a word count and it says nothing directly about rankings.

Which edit usually raises the main entity’s salience most?

  • Adding the entity name to the footer
  • Repeating it ten more times in the last paragraph
  • Naming it in the opening sentence as the grammatical subject

Position and grammatical role carry the most weight. Repetition in one place adds little.

Is the Natural Language API salience score a Google Search ranking factor?

  • Google has not said Search uses it; treat it as a diagnostic
  • Yes, pages need at least 0.7 to rank
  • Yes, but only for news sites

The API is a public Cloud product. It is useful for checking focus, but no threshold is required for ranking.

Why can switching between “GBP”, “Google My Business” and “Google Business Profile” hurt salience?

  • Google penalises abbreviations
  • The mentions may be treated as separate entities, splitting the signal
  • It increases keyword density too much

Inconsistent names can stop mentions being grouped as one entity, so each version scores lower.

Frequently asked questions

Is entity salience a Google ranking factor?

Not in the sense of a published score Google ranks by. Google uses language understanding to judge relevance, and salience is a public way to check whether your page’s focus is clear. Clear focus helps; a particular number does not.

What is a good entity salience score?

There is no official benchmark. What matters is order: your main entity should be the most salient one in the main content, ideally clearly ahead of the next. Absolute scores fall naturally on longer pages that mention many entities.

How do I check entity salience for free?

Use the demo of Google’s Cloud Natural Language API, which shows entities, types and salience for pasted text. TextRazor and other NLP tools offer similar analysis, but their scores are not directly comparable to Google’s.

What is the difference between an entity and a keyword?

A keyword is a string people type. An entity is the thing behind it, which can be described by many keywords and in many languages. “Karachi”, “Karachi city” and “کراچی” are different keywords for one entity.

Does schema markup increase entity salience?

Schema does not change the API’s text score, but it gives Google explicit information about the page’s main entity and related entities. It complements clear writing rather than replacing it.

Sources

  • Google Cloud, “Natural Language API reference: Entity” (accessed 2026).
  • Google Research, Dunietz and Gillick, “A New Entity Salience Task with Millions of Training Examples” (2014).
  • Google, “Introducing the Knowledge Graph: things, not strings” (2012).
  • Google Search Central, “Introduction to structured data markup in Google Search” (last updated 2025).

If you want your key pages checked for focus, entity coverage and structure, my on-page optimisation service covers it, or send me a page you are unsure about.

Rather have an expert do this for you?

Tell me about your website and goals, and I will reply with a free first review.

Your details are only used to reply to your request. No spam.

Grow my trafficCall