2. Understand

Knowledge Graph

The Knowledge Graph is Google's database of entities and the relationships between them, which powers knowledge panels and helps Search and AI features understand what a page is about.

Published

The Knowledge Graph is Google’s database of entities and the relationships between them. It powers knowledge panels in search results and helps Search and Google’s AI features understand what a page is about and which real-world thing a name refers to.

What feeds it

Google builds the graph from many sources: structured data on websites, well-known reference sites, licensed data and the web at large. A site cannot write to the graph directly. What it can do is make its own entities easy to recognise and corroborate: consistent facts on the site, Organization and Person markup that matches the page, profiles linked with sameAs, and independent mentions that use the same name.

Knowledge panels

A knowledge panel is the visible result of an entity being in the graph with enough confidence. New, small companies often have none, and there is no reliable way to force one. Google’s Organization documentation describes the markup it may use for a company’s panel; the rest is corroboration over time.

Generated answers lean on the same entity understanding. A company that is clearly identified is more likely to be described correctly, which is what the Understand stage of our method measures.

Sources

  1. Google Search Central: Organization structured data
  2. Google Search Central: Profile page structured data
  3. schema.org: Organization