An entity is a thing with a unique and independent existence. An entity has an identity. For example, an entity can be a person, a product, a place, an organization, an event, or a concept.

Entities can be physical, conceptual, historical, or fictional things.

An entity is described by its properties 1. For example, "John Doe" is a person and has a name property.

Some properties have as value another entity. For example, "John knows Jane." Properties that have another entity as value are called relations.

While entities by definition have an identity, they may also have an explicit "ID" property. This is especially useful for machines to build knowledge graphs from a set of entities.

It is useful to express properties with a common vocabulary 2 so that entities from different sources can be merged into a single knowledge graph.

The Role of Entities in KGO

Search engines, like Google, extract a global knowledge graph. They process the text of a web page with Entity Recognition Algorithms. Furthermore, they extract statements from the text that indicate properties. For example: "John was born on June 5th, 1978." or "The monitor has two HDMI inputs."

With precise copy, links, and linked data markup we can help a search engine recognize the entities on a page with greater certainty.

Linked data markup and links allow search engines to link different bits of information from different web pages together into a single entity in its knowledge graph. If the information from the different sources is identical it corroborates the properties. Google will only include properties in its output when it has sufficient corroboration of the information.


  1. A property is a specific attribute of an entity. A property describes an entity by its value.

  2. A controlled vocabulary is a list of words to tag units of information in a text or document. Controlled vocabularies can be used for information retrieval.