Search

Knowledge graph

Google Knowledge Graph · entity graph

Definition

Knowledge graph is a store of entities and relations — people, companies, places, works — that a search system uses to fill panels, disambiguate names, and ground answers. Google’s public Knowledge Graph is the famous one. Other engines keep their own graphs or borrow Wikidata.

If two companies share a name, the graph is why one of them gets the panel. Earned corroboration (Wikipedia, Wikidata, consistent news names) is how you get into that graph; a blog post about “our knowledge graph strategy” is not.

How it works

Align the legal name, logo, and sameAs set across the site, Wikipedia if notable, Wikidata, and major profiles. Earn coverage that uses the same name string. Do not create a fake wiki.

Worked example

“Acme Robotics” the warehouse-bot firm keeps losing a panel to Acme Robotics the 1980s toy brand. Adding Organization schema, a Wikidata item with the ticker and founding year, and three trade profiles that use the full current name is the path. A tweet is not.

How it differs

Brand entity is the node. Wikipedia and Wikidata are common sources for that node. Schema is a claim on your page. The graph decides whether to believe it.

Common errors

Buying Wikipedia. Stuffing 40 sameAs links to unused socials. Changing the public name every quarter.

Sources