What is an entity?

An entity is a uniquely identifiable thing or concept that exists in its own right and can be told apart from everything else, for example a person, a place, a company, a product or an abstract idea. In SEO, an entity is such a node in Google’s Knowledge Graph, tied to a unique ID that carries its properties and its relationships to other entities. Entities are the foundation of semantic, entity-based SEO: Google understands content not through single keywords, but through the things a text describes and the way they connect.

The term comes from philosophy and computer science, where it simply describes a “being”, an object with an identity of its own. Google adopted the principle in 2012 with the Knowledge Graph, under the motto “things, not strings”. Since then the search engine no longer sorts content by words alone, but by the objects and meanings behind them.

The term at a glance

What is an entity in search?

In search, an entity is a real or abstract object that Google stores as an independent node with a stable identifier in the Knowledge Graph. Every entity carries a machine-readable ID (a Machine ID in the format /m/… or /g/…) that tells it apart from every other thing, no matter what a user calls it.

An entity consists of three building blocks: a type (person, place, organization, product), a set of properties (attributes such as founding year, population or color) and relationships to other entities. The city of Munich, for example, is a place entity with the property “around 1.5 million inhabitants” and the relationship “located in → Bavaria”. It is exactly this structure that makes an entity understandable to a machine.

The difference to classic web search is fundamental. Google used to count words and how often they occurred. Today it recognizes the object behind the word, links it to everything it knows about that object and delivers knowledge boxes, knowledge panels and answers in the SERP.

How does Google recognize entities?

Google recognizes entities through natural language processing, above all through named entity recognition (NER). An algorithm breaks a text apart, picks out proper names and technical terms and matches them to a known entity in the Knowledge Graph. Four signals carry the most weight:

1. Named entity recognition: Google finds the names of people, places, brands and concepts in running text and resolves their meaning from the context.
2. Co-occurrence: When two entities appear together in many trustworthy sources, Google infers a relationship. If a brand is mentioned next to a service again and again, Google connects the two.
3. Structured data: Markup written in JSON-LD tells Google directly which entity a page describes. The term structured data explains more about that.
4. Knowledge sources: Google matches entities against open databases such as Wikidata, Wikipedia and the Google Business Profile and takes on the facts documented there.

The sameAs property inside structured data plays a key role here. It links your own entity to its records on Wikidata, LinkedIn or in business directories and confirms to Google that all of them describe the same thing. The more consistent the evidence, the more confident Google is in the match.

Entity vs. keyword: where is the difference?

A keyword is a string of characters that a user types in; an entity is the unambiguous concept behind it. The keyword “Golf” has several meanings: it can stand for the car, for the sport or, in German, for a gulf. An entity, by contrast, is always unambiguous. The VW Golf, the sport of golf and the Persian Gulf are three separate entities, each with its own ID.

That clarity solves the biggest problem of pure keyword optimization: ambiguity. Google no longer has to guess which meaning is intended, it reads it from the context of the surrounding entities. When “Golf” sits next to “horsepower”, “gearbox” and “trunk”, the car entity is the one that is meant.

In practice this means that keywords and entities do not rule each other out, they work together. The keyword stays the trigger of the search, the entity is the meaning that Google serves. Anyone who only counts keywords optimizes the surface; anyone who names and classifies entities cleanly meets the intent.

Why do entities matter for SEO and GEO?

Entities matter for SEO and GEO because Google and AI systems alike assemble their answers from entities and their relationships, not from single words. Whoever is anchored in these knowledge networks as a clearly defined entity gets pulled in as a source more often, in classic search results as well as in AI answers.

Three developments make entities the central SEO lever:

• AI answers and AI Overviews: Large language models and the AI overviews in Google Search prefer facts that are documented as unambiguous entities. This is the core of Generative Engine Optimization (GEO).
• Voice search and semantic understanding: Spoken queries are longer and richer in context. Google answers them through entities and their attributes, not through an exact word match.
• Trust and authority: A brand that exists as an entity of its own, with consistent facts, sends strong E-E-A-T signals. Google attaches experience, expertise and trustworthiness to the entity, not to a single URL.

In short: rankings increasingly happen where Google has understood a brand, a topic and their connections as an entity. Visibility in AI search follows the same logic.

How do you optimize for entities?

You optimize for entities by making your own core entities unambiguous for machines and by embedding them in a clear topical network. Entity-based SEO follows five steps:

1. Name entities consistently: write your brand, your services and your core topics the same way everywhere, without rotating synonyms. Every fact belongs in a sentence of its own with a clear entity anchor.
2. Mark up structured data: declare the entity of a page explicitly in JSON-LD (for example Organization, Person or DefinedTerm) and point to Wikidata and profiles with sameAs.
3. Build a network of entities: link related topics as a cluster. A pillar page bundles the main topic, while detail pages such as this glossary entry connect the individual entities with each other.
4. Establish your brand as an entity: records in trustworthy sources (Wikidata, business directories, the Google Business Profile) create the evidence Google needs for a node of its own.
5. Deliver context and completeness: answer every logical follow-up question about an entity so that Google can capture type, properties and relationships in full.

In practice these steps interlock: clean content, precise schema markup and a well planned internal linking network form the foundation together. That is exactly what ongoing SEO support works on when it plans for semantic SEO and AI visibility from day one.

Frequently asked questions about entities

What is the difference between an entity and the Knowledge Graph?
An entity is a single node, a concrete thing such as a person or a place. The Knowledge Graph is the whole network of millions of such entities and the relationships between them.

Does my brand need a Wikipedia entry to be an entity?
No. An entry on Wikipedia or Wikidata strengthens an entity, but it is not a condition. Google also forms entities from structured data, consistent mentions and a well maintained business profile.

How do entities and structured data relate to each other?
Structured data is the most direct way to tell Google about an entity. In JSON-LD you declare the type, the properties and, through sameAs, the identity of an entity, so Google does not have to guess.