About /

Thought Leadership

Google Knowledge Graph

Google Knowledge Graph Overview

Google launched the Knowledge Graph on May 16, 2012. This new functionality amends the way search results are delivered to the user. Rather than just matching queries to results, Google is now looking deeper to understand the real-world entities that people are searching for and their relationship to one another.

For example, a search for the term “giants” could be intended to get results for the NY Giants, the San Francisco Giants, Andre the Giant, the band “They Might Be Giants,” etc.  The Knowledge Graph buckets the search engine results into categories with which the search query is associated and displays the buckets to the right of the typical Google results.

In the example above, results would appear in separate buckets for the football team, the baseball team, etc. The searcher can then click on one of the buckets to see those specific results.

Google created the Knowledge Graph based on 3.5 billion facts, which include information or relationships about 500 million entities.

Executive Summary

Google's new Knowledge Graph buckets search results into categories to allow the searcher to determine which bucket is relevant to their query. The user clicks on a bucket and then is shown only those results. This functionality is available on desktops, tablets, and mobile devices.

Google's goals are designed to help searchers:

  • Find the right thing
  • Get the best summary
  • Go deeper and broader into new information related to their search
  • Monitor KPIs to identify any negative effects form this functionality.

Why Did Google Do This?

According to Amit Singhal, SVP of Engineering at Google, the Knowledge Graph enhances search in 3 main ways:

  1. Find the right thing – Searches can be vague and unclear at times. By understanding the entities that people are searching for and their relationship to each other, Google makes it easier to find the relevant results you are seeking.
  2.  Get the best summary – By better understanding what you are looking for, Google is able to summarize the data relating to that query more effectively. To do this, Google examined the collective history of what its users have been seeking for a given query. As Amit states on the Google blog, “It is not just a catalog of objects; it also models all these inter-relationships.”
  3. Go Deeper and Broader – this relates to the serendipity experienced when a search for one topic leads you to a tidbit or fact that leads you to subsequently search for more information on it.

What About Mobile Devices and Tablets?

Google is also rolling out the Knowledge Graph functionality to mobile devices and tablets, with the goal of making it as easy as possible to get the answers you are looking for while on the go. It takes into account what people have historically been looking for on a given query specific to those devices.  This is important as user behavior varies from device to device.

On tablets and mobile devices, users can tap or swipe the Knowledge Graph to see the results for the category they are interested in. This means users will spend less time typing when trying to refine search queries.

This functionality is rolling out to most Android 2.2+ and iOS4+ devices. On Android, it is available through Google in the browser and the Quick Search Box. On iOS, it is available in the browser and will be coming shortly to the Google Search App.

What Rosetta Thinks of This and How it Will Evolve

With searchers able to find exactly what they want more quickly within the search results, there is likely less of a need for them to go directly to a site to get that answer. Searchers may also be able to find their answers within the actual results themselves, eliminating the need to go to a site at all.  This creates a better experience for the user, but will negatively affect site traffic in some cases. The site traffic that remains should be more engaging, however, as Google and the searcher weed out the irrelevant search results. This should have a positive effect on site engagement KPIs.

The Knowledge Graph could present the opportunity for a new ad format down the road, which would allow advertisers to serve paid ads at the category level rather than the keyword level. This could take shape in a way similar to a comparative shopping engine that utilizes category level ads, and would be beneficial and highly recommended for brands that are looking to “own” their core categories. 

Google results are already affected by the searcher’s Social Graph; this information can also be input into Knowledge Graph results. If your “friends” are all looking for the same type of result for a particular search or talking about a certain topic online, these things could be used to determine the results that searchers see. This also reinforces the importance of brands having the “Plus 1” button on their sites, which Rosetta has recommended since its release.

The Knowledge Graph results appear on the right side of the normal Google search results, on the side rail, where sponsored ads generally appear. This may cause click-through rates (CTR) for paid ads on the right rail to fall as a result of the ads being pushed further down the page to make space for the Knowledge Graph results.

Next Steps

While it will take some time to determine the exact effects of the Knowledge Graph and how its progression will unfold, Rosetta's Analytics Team is continuously monitoring KPIs to quickly identify any negative effects to our clients and their KPIs (such as CTR for side rail ads) resulting from the constant changes in an ever-evolving media landscape. As data is collected on this functionality, we will update our POV.

Additional Links on the Topic