Graph-ology 101 – The Science behind Google Knowledge Graph and Facebook Graph Search

August 13, 2013 | By

Google’s Knowledge Graph

The Knowlegde Graph
Google’s search engine has ‘virtually’ become synonymous with web search, so much so that to ‘Google’ something has come to mean to search for something on the web today.

Aiming to enhance this search experience, Google launched Knowledge Graph more than a year ago for U.S. users and was rolled out globally last August. Over time Google has also made additions to Knowledge Graph for instance, the launch of the Knowledge Graph carousel for local searches.

Knowledge Graph aims to reduce the work users currently do on search and empower them with knowledge about real entities and their relationships. Knowledge Graph lets users look for things, people or places Google has knowledge about: geographical landmarks, movies, celebrities, sports teams, restaurants etc. and gives information relevant to the query.

Knowledge Graph delves into various public databases and information across the web. It contains more than 500 million objects and 3.5 billion facts about and studies relationships between these objects. Knowledge Graph is tuned based on what users search for and what Google finds on the web.

So how does Knowledge Graph use data?

Knowledge Graph makes use of the collective intelligence of the web and sees the world a little more the way people do. This understanding of the way people think helps Knowledge Graph perceive users’ queries to deduce what they exactly want.

For example, if a user types ‘Taj Mahal’, Knowledge Graph displays results for the monument and also the singer understanding that people who search for these terms can mean either.

Google-Knowledge-Graph-Taj-Mahal

In this example, Knowledge Graph uses its data on two different objects that answer to the same keywords. So it dips into its knowledge reserves and shows results for both the real-world entities.

Knowledge Graph also understands that when a user searches for something, they may not just want to see results for that object, but also for other related objects. Understanding the relationship between inter-connected objects is crucial to being able to display information about the right secondary objects. Here’s an illustration:

A user wants to find information on Renaissance painter Leonardo Da Vinci and types ‘Da Vinci’ in the search box:

Google Knowledge Graph

Introducing the Knowledge Graph:

Google Knowledge Graph Da Vinci

The above screenshot is a great depiction of the kind of information Knowledge Graph gathers about entities and how it connects them to related real life objects. It collects information about his birth and death and also understands the connection between him and Italy (his place of birth) and his famous paintings like the Mona Lisa, Michelangelo etc. Knowledge Graph helps users go deeper into queries and make new discoveries by displaying inter-connected results like this.

If users are looking for succinct summarized query results, Knowledge Graph helps Google understand user queries better so that it can summarize relevant content for queries along with important facts the user may need. For example, if a user is looking for information on Nobel laureate Marie Curie, this is how Knowledge Graph would display results:

Google Knowledge Graph Marie Curie

Knowledge Graph here analyzes all the data it has on Marie Curie and gives the above summary with details about her birth, death, family and also about her discoveries and education. It understands that the user may want to know facts about her personal life and so but may also want to know about her education and the discoveries she made.

Knowledge Graph uses data in different ways: for queries that might have different meanings for the same keywords it pulls up data on all the alternatives, for queries that involve lot of details it categorizes the relevant data based on the criteria of importance and it also helps users make surprise discoveries by branching out into other topics related to their query.

Here are two instances of how people are using Knowledge Graph:

Places:

Google Knowledge Graph Canary Islands

Restaurants:

Google Knowledge Graph Restaurants

What Knowledge Graph has in store for marketers…

Continuing with the ‘Taj Mahal’ example, if you are a travel site that talks about the monument Taj Mahal, make sure you mention the same clearly on the site. This way Knowledge Graph can pick up your content for its carousel when a user searches with this keyword. Include other places where your business offers tours in the keywords since Knowledge Graph picks up related content also. Your SEO needs to factor in word relationships into keyword research and copywriting to fit into the Knowledge Graph scheme.

Marketers can also make sure that their promotional content has relevant keywords that will help in being picked up by the carousel. A report that Jack Nicholson attended the Lakers game can affect the Knowledge Graph for Jack Nicholson, Lakers and the NBA. This means brand ambassadors can play a role in getting your brand on to their carousel. Provide as much relevant factual content about your brand to help Google pick it up as data for their carousel.

Google has also indicated that it will be picking up a lot of data from Google+ so a good Google+ page for your brand may help your brand stand out better.

Next comes…

Facebook’s Graph Search

Facebook Graph Search Icon
Compared to Knowledge Graph, Facebook’s Graph Search is still in its infancy having been released in January this year. It is still being rolled out and users who want to get Graph Search need to join a waiting list.

People are increasingly looking for relevant content that jives with their interests when searching for information on the web. Facebook is trying to cater to this need with Graph Search.

Graph Search uses people’s friends as the sieve that filters information. People can search for restaurants near them, best hotels in places they are travelling to, photos shared by Pages they like or games their friends like to play. For instance, people can look for ‘Mexican restaurants my friends like’, ‘Music my friends like’ or ‘Photos I like’.

But how does Graph Search work?

The search bar initially shows the top search suggestions which include people, Pages, apps, groups, places and suggested searches.

These search suggestions take users to unique results pages. These results are based on aspects that include information shared by businesses and the connections of the person searching.

Facebook has also tied up with Microsoft’s Bing for accommodating web searches.

Regardless of whether users have Graph Search or not, they can still see Page and app sponsored results.

Facebook Graph Search

How does Graph Search use data?

The way Graph Search uses data is influenced by the fact that it can only access information that has been made public. When a user searches for something, Graph Search looks at information from that user’s connections. Since each user has a different set of connections, every person sees unique results.

Graph Search studies the likes and preferences of a user’s connections, pulls up information relevant to the user’s search query and displays customized results. When a user is looking for information on a business Page, Graph Search pulls up information from those Pages also and shows search results.

What does Graph Search mean for marketers?

Since Facebook Graph Search is very new the understanding of its implications is still evolving. However, common sense suggests that if people are going to search based on their friends’ brand preferences then it would be advantageous for brands to build large and authentic fan bases. The broader the fan base, the wider the brand’s reach. The more representative the fan base is of the brand’s target audience, the more authentic it will appear and hence compelling. Positive associations with the brand amongst friends will make for powerful endorsements of the brand. In other words, brands can hit the word-of-mouth marketing jackpot.

Graph Search also offers an interesting way for brand to understand the demographics, likes and interests of their fan base helping marketers understand their audience better.

How do Knowledge Graph and Graph Search compare?

The main difference between the two search features is that while Knowledge Graph is more web oriented, Graph Search is more social. Knowledge Graph crawls the entire web to find answers to users’ queries. Graph Search, however, searches within Facebook amongst a user’s connections and only looks at information that is shared publicly.

Knowledge Graph makes use of information on the web along with its understanding of the relationships between objects. Graph Search uses information shared publicly on Facebook along with the common threads connecting the user with their friends and the object of the search. While Facebook’s Graph Search is more interest-oriented, Knowledge Graph is more generic.

What are your thoughts on Google’s Knowledge Graph and Facebook’s Graph Search? Let us know!

Naksha M. Kumar

About the Author: Naksha M. Kumar

With more than 3.5+ years of experience in content generation and management, Naksha builds marketing collateral in different forms at Position2. Aided by an interest kindled by her work, she is continually on the hunt for and writing about trending digital marketing news which will impact B2B marketers. When away from work, she loves studying languages, reading books and interpreting them, watching action movies, learning about different cultures, meeting new people and making friends.

  • http://www.tone.co.uk Anthony

    Thanks for writing such an easy to digest explanation on knowledge graph and graph search. I’ve been struggling to fully understand how they both work but now I’m alot clearer. Excellent use of imagery too! Shared on Twitter :)https://twitter.com/Anthony_Mac85/status/405734389549891588

  • Pingback: Pinterest, Google and Facebook Updates for the Week | Position²()

Get in Touch.
We’d love to hear from you.

 

Subscribe to our monthly and
weekly email newsletters

 

Toll-free (US):

800-725-5507