Knowledge Graph Navigator

What Are Entities in SEO? (With Examples)

In SEO, an entity is a specific thing a search engine can identify and store as a distinct record: a brand, a person, a product, a place, an event, or a topic. Entities are what queries are about; keywords are the strings people type to ask about them. The practical difference is that one entity is usually the target of dozens of different keyword strings, and a page written for the entity can serve all of them, while a page written for a string serves one phrasing.

Entities versus keywords, with examples

Four cases show the mismatch between the two layers.

One entity, many keywords. The University of Florida is one entity. “university of florida,” “uf gainesville,” “gators school,” and “florida state university in gainesville” are four strings pointing at it (the last one incorrectly, which the search engine has to sort out). A page about the university is relevant to all of them because it is about the thing.

One keyword, many entities. “mercury” is one string that could mean the planet (Wikidata Q308), the chemical element (Q925), the Roman god, or the record label. A search engine cannot rank pages for “mercury” until it decides which one the searcher means, and it uses the rest of the query plus history and location to decide.

An entity with no keyword. A regional accounting firm with forty employees is a real entity with a name, an address, and a founder. It may have almost no search volume. Keyword tools will report it as nothing; the search engine still needs to know whether it exists, where it is, and whether it is the same firm mentioned in a local news article.

A keyword that is not an entity at all. “how to file an extension” is demand for a procedure, not a reference to a thing. Plenty of valuable queries are like this. Entities are one useful unit of planning, not the only one.

Why the distinction changes how content gets written

Once a query is resolved to an entity, near-duplicate phrasings stop being separate opportunities. “neo4j vs postgres” and “postgresql or neo4j which is better” are the same comparison between the same two entities. Publishing one page per phrasing creates several thin pages competing with each other rather than earning multiple rankings.

The productive move is to plan a page around the thing and the questions a reader has about it. For a product entity, that means the attributes a buyer needs (what it is, what it costs, what it works with, who it is for, what it replaces) and the adjacent entities they will look up next. That same relationship structure is what a knowledge graph would hold about the product, expressed in prose. The method is covered at /seo/semantic-seo/.

Writing also has to make the referent obvious. A search engine’s extraction pipeline finds names in text and then decides which known thing each name refers to, using nearby context. Naming an entity in full on first mention and putting a type or place next to it (“Neo4j, a graph database company founded in Sweden”) gives that pipeline what it needs. See /concepts/entity-linking-ner-and-extraction/.

Why it changes how markup gets written

Structured data is the machine-readable restatement of facts a page already contains. Its value for entities is identity: an @id says “this node,” and sameAs says “this node is the thing already known at these other addresses.”

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "@id": "https://neo4j.com/#organization",
  "name": "Neo4j",
  "url": "https://neo4j.com/",
  "sameAs": ["https://en.wikipedia.org/wiki/Neo4j"]
}

Without the sameAs line, a system reading this page has an organization named “Neo4j” and has to work out whether it is the Neo4j it already knows. With it, the page has answered the question. Markup does not raise rankings and no search engine documents it as a ranking factor; it removes ambiguity about what is being described and makes a page eligible for the specific result types Google documents. Format details are at /languages/json-ld/ and property lists at /seo/organization-schema/.

What counts as an entity on an ordinary website

Most sites have four groups worth listing explicitly.

  • The organization itself, with one canonical URL, one primary name, and sameAs links to whatever public identifiers exist.
  • The people, particularly named authors, executives, and practitioners whose credentials the content depends on.
  • The things sold or offered, whether products, services, courses, or degree programs.
  • The topics covered, which are entities in the knowledge graph sense whenever a shared vocabulary has an identifier for them (photosynthesis is Wikidata Q11982 just as Marie Curie is Q7186).

For each one, the question is whether a search engine could tell which thing it is from what is published. That audit, and the work that follows from it, is what /seo/entity-seo/ covers in full.

What entities are not

Entities are not high-volume keywords, and “entity” is not a synonym for “topic.” An entity is a referent with an identity; a topic is a subject area, which may or may not have one.

Adding markup also doesn’t create an entity. Markup describes a thing and helps a search engine recognize it. Whether the engine adds that thing to its own graph depends on evidence from sources other than the site making the claim, which is why entity work involves third-party profiles and public catalogs as much as it involves code.

Related pages

FAQ

What are entities in SEO?

Entities in SEO are the specific things search engines identify and store as distinct records: brands, people, products, places, events, and topics. Queries are about entities; keywords are the strings used to express them. One entity typically matches many keyword variants, so planning content around the thing covers more demand than writing a page for each phrasing.

What is the difference between an entity and a keyword in SEO?

A keyword is a string of characters typed into a search box. An entity is what that string refers to. “mercury” is one keyword pointing to several entities, including a planet and a chemical element, while the University of Florida is one entity reached through dozens of keyword variants. Search engines resolve the string to the entity before ranking pages.

Do I need structured data for search engines to see my entity?

No, but it removes guesswork. Search engines extract entities from page text without markup. Structured data restates the same facts in a parseable form and, through @id and sameAs, states which known thing the page describes. It is not documented as a ranking factor, and it does not substitute for the third-party sources that establish an entity exists.

Sources and further reading

Powered by CookieAdmin