What is Google's Knowledge Graph?
The Knowledge Graph is Google's database of real-world things (people, places, organizations, works) and the relationships between them. Google launched it in 2012 so a search could be matched to the entity it means, not just the words it contains. Knowledge panels and many of the factual answers in Search come from it.
Key takeaways
- The Knowledge Graph is Google's graph of entities: things, not strings of text, joined by their relationships.
- It is fed by public sources, licensed data and what owners contribute themselves, for example when they claim their knowledge panel.
- Knowledge panels are created by Google's algorithm when the web has enough information; they cannot be bought or ordered.
- A store helps Google recognize its brand with Organization data on the homepage; its products live in a different Google graph, the Shopping Graph.
- It is not the @graph of a JSON-LD block: that one is a page's graph; the Knowledge Graph is Google's.
"Knowledge graph" in general means any database that stores things and the relationships between them, and there are plenty, Wikidata among them. But when someone in SEO writes Knowledge Graph, capitalized, they mean Google's: the map of entities Google uses to work out which thing you are talking about when you type a word. It matters to a store for two reasons. It is where the panel that sometimes appears beside the results comes from when someone searches your brand, and it is the idea behind almost everything said today about "entities" and AI search.
How it works
Google introduced it on May 16, 2012 with a slogan: "things, not strings". Each thing in the graph is an entity. Google's example was the search "taj mahal": to a classic search engine that is two words, but a person might be thinking of the monument, a musician or a casino. The Knowledge Graph is the model that understands real-world entities and their relationships, so it can tell those meanings apart, summarize the key facts about the thing searched and suggest related things. At launch it held more than 500 million objects and more than 3.5 billion facts about them and the relationships between them, and it started from public sources such as Freebase, Wikipedia and the CIA World Factbook.[1]Google, official blog (Amit Singhal). Introducing the Knowledge Graph: things, not strings
Today Google describes it as its database of billions of facts about people, places and things, which it uses to answer factual questions like "how tall is the Eiffel Tower?". The data comes from three places: public sources, data Google buys under license (sports results, stock prices, the weather) and what the content owners contribute directly, for example those who claim their knowledge panel and suggest changes.[2]Google Knowledge Panel Help. How Google's Knowledge Graph works
Knowledge panels are the visible face of the graph. Google says its algorithm creates them automatically when there is enough information on the open web, that its automated systems can stop showing them, and that the subject or an official representative can claim the panel and propose changes. Businesses with a physical location also have their own panel, the Business Profile one, which is managed separately.[2]Google Knowledge Panel Help. How Google's Knowledge Graph works
The graph can also be queried, within limits. The Knowledge Graph Search API returns entities with Schema.org types, in JSON-LD format, each with its own identifier. Google warns that it returns individual entities, not the graph of relationships between them, that it is not suitable for a critical production service and that it is moving it to Cloud Enterprise Knowledge Graph.[3]Google for Developers. Google Knowledge Graph Search API
Where your site comes in. Google recommends putting the organization's structured data, the Organization type, on the homepage: it helps Google understand the organization's administrative details and tell it apart from other organizations in the results, and some properties influence the logo it shows and the knowledge panel. If you are a merchant, you can influence more details of the merchant panel, such as the return policy, the address or the contact details.[4]Google Search Central. Organization structured data To link your brand to its profiles on other sites there is sameAs, the URL of a page on another site with more information about your organization,[4]Google Search Central. Organization structured data and Google says it can use sameAs and the rest of Schema.org in general, beyond its rich results.[5]Google Search Central. Introduction to structured data markup in Google Search
And products go in a different graph. For shopping, Google uses the Shopping Graph, which in May 2025 held more than 50 billion product listings, with reviews, prices, colors and availability, and of which more than 2 billion are refreshed every hour. It is the one Google combines with Gemini to shop in AI Mode.[6]Google, official blog (Lilian Rincon). Shop with AI Mode, use AI to buy and try clothes on yourself virtually
In a store
The block a store puts on its homepage to introduce itself as an entity. Fictional example: the brand, addresses and profiles are made up.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "Calella",
"url": "https://www.example.com/",
"logo": "https://www.example.com/img/logo-calella.png",
"description": "A shop for linen shirts made in Barcelona.",
"sameAs": ["https://example.net/profile/calella", "https://example.org/calella"]
}
</script>
Each property does a different job. name, url and description say who you are. logo is one of the properties that can influence the logo Google shows in the results and in the panel.[4]Google Search Central. Organization structured data sameAs points to your profiles on other sites, so Google can recognize that they all talk about the same brand.[4]Google Search Central. Organization structured data It goes on the homepage or on a single page that describes the company, such as "About us", not on every page, and Google recommends using the most specific Organization subtype that applies.[4]Google Search Central. Organization structured data
What the block does not do is create a panel. Google decides whether your brand gets a panel based on the information it finds on the open web, and markup is only part of that information.[2]Google Knowledge Panel Help. How Google's Knowledge Graph works
What gets confused
Believing a knowledge panel can be requested or paid for. Google says its algorithm creates panels automatically when there is enough information on the web, and that when, where and how one is shown is decided by its automated systems, something Google itself says it cannot and would not want to influence.[2]Google Knowledge Panel Help. How Google's Knowledge Graph works What the owner can do is claim the panel that already exists and suggest changes.
How to check it
- Search your brand on Google from a computer and see whether a knowledge panel or a Business Profile panel appears. They are different things and are managed separately.[2]Google Knowledge Panel Help. How Google's Knowledge Graph works
- If the panel exists and you represent the brand, claim it so you can suggest changes from the verified account.[2]Google Knowledge Panel Help. How Google's Knowledge Graph works
- Check the Organization block on your homepage with the Rich Results Test and fix the critical errors, as Google asks in its organization markup guide.[4]Google Search Central. Organization structured data
- Review every
sameAslink: it has to lead to a profile that exists and is yours.[4]Google Search Central. Organization structured data - If you have technical access, query the Knowledge Graph API with your brand name to see whether Google has it as an entity, with which types and which description. It is a read-only query.[3]Google for Developers. Google Knowledge Graph Search API
Related concept: @graph
The @graph of a JSON-LD block is also a graph, which is why the two get confused. But it is a different scale: according to the W3C specification, @graph expresses a graph within a document, and gathers several nodes that share a context into a list.[7]W3C. JSON-LD 1.1 On a product page, those nodes are your organization, your site, your breadcrumbs and your product. The @graph is what your page says about itself; the Knowledge Graph is what Google believes it knows about the world. The first is one of the sources of the second, never a guarantee of getting into it.
To keep going, the entity entry explains the unit the graph is made of, and the structured data reading path puts the rest of the terms in order.
Synonyms
- English
- Google Knowledge Graph, knowledge graph
- Español
- grafo de conocimiento, gráfico de conocimiento
Sources
- [1] Google, official blog (Amit Singhal). Introducing the Knowledge Graph: things, not strings. Accessed . ↑
- [2] Google Knowledge Panel Help. How Google's Knowledge Graph works. Accessed . ↑ ↑ ↑ ↑ ↑ ↑
- [3] Google for Developers. Google Knowledge Graph Search API. Accessed . ↑ ↑
- [4] Google Search Central. Organization structured data. Accessed . ↑ ↑ ↑ ↑ ↑ ↑ ↑
- [5] Google Search Central. Introduction to structured data markup in Google Search. Accessed . ↑
- [6] Google, official blog (Lilian Rincon). Shop with AI Mode, use AI to buy and try clothes on yourself virtually. Accessed . ↑
- [7] W3C. JSON-LD 1.1. Accessed . ↑
Change history
- First English version.
This definition is published under CC BY 4.0. Reuse this definition freely, crediting Clione with a link to this page.
Related terms
Entity
What an entity is in SEO: an identifiable thing, not a keyword. How Google recognizes one and how a store describes its brand and products.
sameAs
What the Schema.org sameAs property is, which links it takes and which it does not, where Google uses it (Organization, authors) and how to review yours.
Organization
What Schema.org's Organization type is, which properties Google recommends, where it goes in an online store, and why it has to match your site.
@graph
What @graph is in JSON-LD, what the W3C specification says, how it links a product page's product, store and site, and how to review your store's.
Structured data
What structured data is, which pieces make it up, what a store gets from it according to Google, and how to check it matches the product page.
Rich results
What Google rich results are, which ones matter in a store, why valid markup doesn't guarantee you'll see them, and how to check.
Make your catalog easy for bots to read
Clione fills every product page with what bots read best: the JSON-LD graph with its structured data, the meta description and the FAQs. You review and press Push; then we check your live store so it all stays published.