Entities & Knowledge Graphs
In shortA distinct, identifiable thing such as a person, place, organisation, product or concept, which search engines and LLMs use to understand meaning.
An entity is a single, well-defined thing that can be distinguished from everything else: a person, organisation, place, product, event or concept. Search engines moved from matching keyword strings to understanding entities, which Google summed up in 2012 as “things, not strings”. For SEO, entities are the units of meaning that search engines and large language models use to connect queries, pages and brands.
A search system needs to do three things with entities:
Tools such as the Google Cloud Natural Language API return each entity’s type, a salience score for how central it is to the text, and, where available, a Knowledge Graph ID and Wikipedia URL. MLforSEO’s semantic keyword research course builds on this with the entity-attribute-variable (EAV) model: an entity, its attributes and the values those attributes take.
A page titled “Lightweight waterproof hiking jackets for women” contains the entities hiking jacket (product type), waterproofing (attribute), women (audience) and perhaps brands such as Patagonia. In MLforSEO’s benchmark of entity extraction tools, the same 221-word bio was run through several APIs and chatbots; the purpose-built APIs also returned salience and knowledge-base links that the chatbots did not.
