The Knowledge Graph is Google’s system for organising information about millions of well-known entities: people, places, and organisations, to build a map of how information is interconnected. It’s a knowledge base used by Google to enhance its search results with the use of human language technology (e.g. entity recognition and linking) and the semantic web (graphs of linked data).
The Knowledge Graph is used behind the scenes to help Google improve its search relevancy, but also to enhance search presentation: When a search query contains a notable entity—such as a well-known brand or a famous person—a Knowledge Panel, which is a type of rich answer, is displayed on the results page, populated with instant answers and links to related content.
This is because Google has learned that when users enter a search query, they’re not simply looking for answers, but are open to invitations to explore a whole network of connections. So to enhance the user experience, Google has started to generate insights based on the most popular questions people ask and associated content.
Here’s what that card typically looks like:
The card is populated with information Google has determined as being relevant to the search query you have entered, in this case “Roger Federer”. It includes a basic overview including images and date of birth, but also includes more intuitive results like number of Grand Slam singles titles won. It provides relevant social and web links, and contextualises results by offering information on topics ‘People also search for.’
The Knowledge Graph can can also appear as a drop-down image carousel, like this search for the “best tennis player”:
Bing has a similar tool, called the Satori Knowledge Base, which provides “snapshots” with information about an entity. The cards are available on desktop, tablet and smartphone.
How does it work?
The Knowledge Graph is an intelligent model that understands real-world entities and their relationships to one another. Google calls it “things, not strings”. Think of the Knowledge Graph a little like Google’s own encyclopedia, made up of Freebase, Wikipedia, CIA World Factbook etc. entries. It’s used to supplement/enhance queries and improve the search experience. It’s essentially a database that collects millions of pieces of data about keywords people frequently search for, as well as the intent behind those keywords, based on content that is popular with searchers.
Google doesn’t publish guidelines for generating a Knowledge Graph, so the criteria regarding exactly how it works can be a little fuzzy. Nevertheless, certain areas trigger Google’s interest more than others. In the past, before Google Hummingbird, the number of trustworthy links that your web page earned counted for everything. Link building and associated link metrics were the major ranking factors, even for brand-new blogs. But, since Hummingbird Google is now far better at determining what matters to the user. That means the sites with the best user experience, well structured and good quality content will bear fruit in the long run.
If a search for your brand doesn’t already trigger a Knowledge Graph panel, the reason is that the search engine doesn’t count the brand as being a popular enough entity for searchers to benefit from seeing additional information.
The only way a search engine understands that an entity or brand is important is because it is talked about on the web. Improving your local SEO ranking is one option Google suggests may help generate an accurate knowledge graph, and your SEO agency can help with onsite copy and content creation strategy to boost your business’ relevance.
How does the knowledge graph affect SEO?
While the Knowledge Graph seems like such a user-friendly feature for users, helping them find information easier and faster, it poses a number of problems for SEO. Here’s Google’s search result for “Apple”, for example:
Looking at the snapshot above, we can see that Google seems to be relying on the major data sources it collaborates with, in this case its Wikipedia data, which, like Freebase, relies on public contributions from volunteers. The bottom line is that the Knowledge Graph can be both friend and foe. It can link to your brand-owned content, but it can also link away to other sites, including sources you cannot always control.
And that can cause a problem. First of all, if we take a look at Apple’s’ official website, we can see that there’s more to it than facts about the company’s history, stock price or executives. It’s highly promotional and there are several call-to-action buttons, such as “shop” or “learn”, but these cannot be found on the card presented to the right of the search results. If a searcher can reap the information needed from the knowledge graph, they are less likely to enter the official site, and therefore will not be exposed to the call to action buttons. This way, the chances of converting that user into a buyer will seriously decrease.
Apple probably can’t control the fact that their executives are pulled through into the Knowledge Panel either. Clicking on “View 10+ more” on the panel takes the user to a special search page for “apple executives” featuring a film strip at the top. It’s a result Google has determined as being relevant to your initial query based on association and data stored in its Knowledge Graph, but for Apple’s digital marketing team, it’s most likely just an unwarranted distraction from their primary focus: sales.
While the Knowledge Graph data can be difficult to control, most of these queries tend to be informational and thus don’t hold that much transactional value for a business website anyway. A good agency will try to do what they can to improve Knowledge Graph entries (this may include helping optimise your presence in branded search results) but will spend considerably more time on structured data to improve rich snippets. Rich snippets and rich results are how Google interprets web page information based on how the website’s data has been marked up, usually using schema or some other form of structured data.
Rich results take effect when the query is a bit less clear (usually not just person or place but an actual question). Google’s results are then taken directly from a website (rather than the Knowledge Graph).
How do you optimise your knowledge graph?
There are three key ways your SEO agency can help your site appear in rich snippets:
Structure your data
Structured data refers to site data with a high level of organisation, such as contact information or, say, a food recipe. When information is highly structured and predictable, search engines can more easily organise and display it in blended search results in creative ways. When you’re optimising for Google and other search engines, the best strategy is to spoon-feed the search spiders with your content. Make it as easy as possible for them to find their search terms on your page.
Structured data markup is a strategy that applies searchable tags to your site data. It typically uses the schema vocabulary to can indicate key business or company details, provide business contact mark up and indicate other key content on your site. An SEO agency can perform effective schema markup for your site.
Improve your on-site structure
The Knowledge Graph considers your site relative to context and the user’s intent, which means you not only need to make your site navigable to search spiders, but to real life human users too. What matters is understanding the way your audience finds certain information on your site and what other information related to the search they might be looking for. This also includes external linking to social media profiles.
Knowing which keywords fit certain pages and a balanced structure of your pages and sections has always been important in terms of SEO. Still, the knowledge graph makes this features indispensable. Your SEO agency can help structure your site navigation to optimise for richer search results
Be an authority
If you’re the authority for certain content, Google can treat the structured data on your site as factual and import it into the Knowledge Graph, where it can power prominent answers in search and across Google properties.
A search engine’s main function is to answer queries. The Knowledge Graph understands this need and is now focusing on answering real questions, rather than simply generating a list of websites that match keywords. As Google is doing a great job of returning highly relevant and related content, it’s necessary for websites to step up and do the same.
Using this information properly doesn’t mean just adapting to the knowledge graph’s rules but also to your audience’s needs. If you want to rank high in search engine results pages (SERP) with the Knowledge Graph algorithm, you should consider targeting key search queries such as “how to…”, “how do I…”, “who is…” etc. in your on-site content. It’s the role of your SEO agency to determine the most effective content strategy to predict these questions.