What is the Google Knowledge Graph?
Here's what to know about the Google Knowledge Graph, aka, the massive knowledge base of billions of entities and the relationships between topics.
#SPONSORED
NewzDash wins the Best Global SEO Software Suite at Global Search Awards 2023
🚀 The ONLY SEO tool that provides: 1) LIVE Search Visibility 2) Real-Time Search Volume 3) Unlimited Trends Monitoring & Rankings 4) Historic Performance & KPIs 5) Content Recommendations 6) Global Competitive Analysis 7) Publish-to-Indexing Tracking 8) Instant SEO Recommendations and much more!
Hello, and welcome back. Jessie here, fresh off a (2023) record six softball games in one weekend — including three in what was, effectively, a downpour. The softball hot streak included the company tournament, where yes, we returned the trophy to its rightful home (editorial).
This week: The Google Knowledge Graph, aka, the massive knowledge base of billions of entities and the relationships between them.
Join our Slack community to chat SEO any time.
Let’s get it.
In this issue:
What is Google’s Knowledge Graph?
How the Knowledge Graph works.
Entities and entity relationships.
THE 101
What is Google's Knowledge Graph?
What is it: The Google Knowledge Graph is a database of entities, including information about those topics and the relationships between them. The objective of the Knowledge Graph is to enable Google's systems to "discover and surface publicly known, factual information when it is determined to be useful."
The Knowledge Graph powers some features — including Knowledge Panels and rich snippets — in search results.
It’s akin to a massive encyclopedia: A dataset that makes sense of all the information on the internet. Google refers to it as a “collection of information about people, places and things.” As of 2020, it consists of more than 500 billion facts about five billion entities.
The data in the Google Knowledge Graph comes from “hundreds of sources from across the web.” That includes public sources, licensing agreements the company has (for example, for sports scores, stock prices, etc.), and from content owners who edit their Knowledge Panels (more on this later).
The “Feedback” link at the bottom of a Knowledge Panel (and other search features) sends a report to Google about any inaccuracies (which are reviewed by Google).
Why it matters: Readers seek out the same information in different ways.
The Knowledge Graph helps the search engine understand what exists in the world and the relationships between those things.
When readers turn to Google with queries, it’s able to see the meaning behind the search.
The Knowledge Graph enables the search engine to recognize entities or attributes in natural language queries. All the internet's information organized into entities makes for faster retrieval and more accurate information results.
What are entities?: Any object or concept that can be distinctly identified is an entity. This includes tangibles — people, places, organizations or things — as well as concepts, abstract notions or feelings. Google refers to entities as “topics” in reader-facing documentation.
On queries such as, "Who is Canada's prime minister?," the answer — Justin Trudeau — shows up on the SERP quickly and accurately. That's thanks to the Knowledge Graph.
On the right of the SERP is Justin Trudeau's Knowledge Panel, which confirms the entity and surfaces what is known in Google's database.
Searching the query "Justin Trudeau" also returns a summary of basic information about the Canadian prime minister — all of which is stored in the Google Knowledge Graph. (The SERP also returns a news box in Canada, thanks to the recent cabinet shuffle.)
The Google Knowledge Graph is not just a "catalog of objects," the company said in a blog post when launching the product in 2012. Rather, it creates and understands relationships between entities — how two more topics are connected to each other. It is not just a list of everything that exists on the internet, but rather a map that creates and understands the links between things. The Google Knowledge Graph "models all these inter-relationships."
In the Trudeau example above, the “People Also Search For” panel on the right hand-side includes related entities: His wife, father (and former Prime Minister Trudeau) and brother, as well as the country’s opposition leaders. Google understands these individuals are connected IRL.
What is a Knowledge Panel?
Also in the SERP: A Knowledge Panel — the clearest signal an entity exists in the Knowledge Graph. This feature takes up real estate in SERPs on desktop and mobile (but the position of the features changes depending on the device).
Worth noting: The Google Knowledge Graph is not the same thing as a Google’s Knowledge Panel. A Knowledge Panel surfaces (visually, in search results) a subset of data pulled from the Google Knowledge Graph (the underlying data set).
What is included in a Knowledge Panel: Google says its aim is to display the “most relevant and popular information for a topic” in a Knowledge Panel. The information surfaced will vary — because “no topic is the same” — but generally includes:
A title, summary of a topic and key facts;
A description of the subject;
Images for the entity;
Links to social profiles and official websites.
Depending on the entity, the Knowledge Panels might include songs, episodes of a television show or the lineup for a sports team.
There are many results where a Knowledge Panel is a sufficient response. The question, “How old is Justin Trudeau,” can be answered directly in the SERP because it’s a simple piece of data. These are often zero-click searches (which generally have low traffic potential).
The Knowledge Panel is the most common way data from Google’s Knowledge Graph appears in search results.
If your organization or the people who work there have a Knowledge Panel, you’re likely to see increased visibility in SERPs on related queries.
In addition, thinking about entities makes for a more holistic news SEO approach.
Focus on topic authority: Instead of aiming for precise, specific queries, focus your strategy on ranking for many queries on a given entity.
That means writing content and demonstrating your expertise for every facet of an idea, rather than chasing the top spot for one specific keyword.
A ‘strings not things’ approach
The Knowledge Graph allows Google to go beyond just keyword matching and return more useful results.
“We’ve always believed that the perfect search engine should understand exactly what you mean and give you back exactly what you want,” the company declared in its original blog post, referring to ‘things not strings.’'
For example, these queries are different words, but concern the same entity:
These are informational queries that largely return basic biographical details and related content. That’s because Trudeau is a known entity in Google’s Knowledge Graph.
While the SERPs vary, they all concern the same person.
The same is true of an institution. People search for Bank of Canada information in variety of ways, including:
All three queries refer to Canada’s central bank. Google will surface the relevant, accurate information because it’s an understood entity in the Google Knowledge Graph.
(While returning similar results, look at the placement of Top Stories: It’s highest for ‘BoC’, suggesting a higher news-seeking intent.)
Finally, look no further than querying the film roles of Canada’s most famous ex-Mickey Mouse Mouseketeer: Ryan Gosling.
The films he starred in — from The Nice Guys to La La Land — are understood by the Google Knowledge Graph system as entities connected to him. The search, “Barbie movie ken actor” also returns Gosling’s Knowledge Panel.
The SERP makes clear it understands the relationship between the noted entities.
Side note: “Knowledge graph” refers to any semantic network of data. There are many other knowledge graphs — for example, Wikidata.org. This newsletter considers only Google’s version.
THE HOW TO
Where understanding entities matters most
The importance of entities and the relationships between them is clear on Discover, the queryless platform from Google. Discover does not require user input (a search query), but curates content based on a person’s interests in topics.
That means: If your stories on a specific entity (Bank of Canada, inflation, famous Canadian Ryans — Gosling or Reynolds) already perform well on Discover, Google understands your outlet to have strong authority in those areas. Once you understand which topics (entities and concepts) work for your publication, double down on coverage.
This entity-focused work has the benefit of shaping a holistic overall approach to news SEO. Combine the ‘things not strings’ approach with ‘your coverage is a tree’ planning, and you’re well-positioned to effectively serve your audience’s needs and interests.
THE HOW TO
The Knowledge Panel: How to get it
For readers, SERPs that return Knowledge Panels convey strong authority signals. While it’s not possible to force your way into a Knowledge Panel, you can try to influence it.
For the organization: Use organization schema (use the sameAs property to cite your social presence and Wiki entries).
For reporters: Use the author schema, as covered in our Authorship issue (use the sameAs for all places on the internet your reporters have profiles).
It is not possible to request a Knowledge Panel. And it might depend on your overall authority, meaning smaller sites are less likely to get that visibility. However, strong signals — like correct schema, well-structured web pages, a solid backlink profile, a Google My Business profile, a wikidata.org entry or Wikipedia page — work to convince Google the entity is worthy of the feature.
On queries where Knowledge Panels do exist, users can claim and verify the panel. Start by clicking the “Claim this Knowledge Panel” button that appears under the feature and follow Google’s instructions. Once verified, click the “Suggest an edit” link to make changes for accuracy and relevancy.
The bottom line: The Google Knowledge Graph is a massive data set that enables Google’s systems to return speedy, accurate results. Entities and the relationships between them play a key role in the Knowledge Graph. The Knowledge Panel is its most visible display in search results. Your organization can try to influence how these show up by using proper schema, well-structured pages and a solid backlink profile.
#SPONSORED - The Classifieds
State of Digital Publishing: Build powerful SEO capabilities in-house with SODP's Publisher SEO course. News & Evergreen SEO best practices included. 20% off with code
WTFISSEO.
PixieBrix: Quickly and easily access your own writing assistant from any webpage. Write for your target audience, follow brand guidelines, and quickly shorten text. AI Writing Assistance Anywhere
What did you think of this week's newsletter?
(Click to leave feedback.)
THE JOBS LIST
These are audience jobs in journalism. Want to include a position for promotion? Email us.
Penske Media is hiring a Senior SEO Manager located in either New York or Los Angeles.
RECOMMENDED READING
🤖 Barry Adams: The SEO for Google News newsletter has a new edition — Advanced insights into Googlebot crawling — that follows the expert guidance Barry shared in this newsletter about crawl budget.
🔎Search Engine Land: Here’s how to increase your site’s chances of being featured in snippets.
🗺️SEO for Journalism Slack: Should publishers include sponsored content in sitemaps? (Read our What is a sitemap? newsletter edition for more background.)
📍Search Engine Roundtable: Having the site name in the title tag really shouldn't be an SEO decision, says Google’s John Mueller.
📸Brodie Clark: Here’s how Foot Locker massively increased images being indexed on category pages.
🤔 Aleyda Solis: SEO area importance vs. website size — the answer may surprise you!
Get your company in front of more than 7,100 writers, editors and digital marketers working in news and publishing. Sponsor the WTF is SEO? newsletter!
Have something you’d like us to discuss? Send us a note on Twitter (Jessie or Shelby) or to our email: seoforjournalism@gmail.com.
Written by Jessie Willms and Shelby Blackley