How Google’s Knowledge Graph Has Affected Search
In May, Google began to roll out the latest and, very likely, the most significant change to their search engine that they have made so far. This new feature on Google, called Knowledge Graph, has changed the way we can use search engines and has had an almost immediate impact on some areas of the World Wide Web. In the simplest terms, Knowledge Graph is a new display on Google’s SERPs that shows a summary of the information that Google holds in its Knowledge Graph database for the terms used in the search. The reason that it is such an important change to search engine technology is because it is a major introduction of a semantic feature into Google’s results pages and so it heralds the new age of searching the internet. Google’s search engines have just gotten a whole lot smarter.
Semantic search is an idea that has been around since the 1960s and comes out of the Artificial Intelligence field of computer science. At its base it uses huge, superfast computers to interpret the user’s intention when they enter a search term. In the Google Official Blog article that announced Knowledge Graph, Amit Singhal, SVP of Google’s Engineering1, used the example of Taj Mahal as a search term. This could be either the building in Agra, a casino in Atlantic City or a blues musician. Knowledge Graph will analyze what it knows about Taj Mahal and offer the option of choosing search results from the three as it generates the SERP, taking users straight to the search results that they were after. When the preferred option is chosen, Knowledge Graph will then assemble a summary of what it knows about that particular Taj Mahal and display it on the previously blank white column on the right of Google’s SERPs.
This semantic search is made possible by the way that Knowledge Graph connects the points of data on the web. Google’s database of knowledge is based more on the nature of the connections that users make with words rather than with finding an exact match to a keyword term. To do this, Google has factored their users’ most common associations into their calculations. For many search terms, there are several alternate meanings, but these can easily be defined by the words most commonly associated with them to find the exact meaning that users are looking for. So for a search term like Taj Mahal, Google would associate India with the building and poker with the casino, making many long-phrased search terms like ‘Taj Mahal Palace in India’ point to a distinct set of information. Once this has been done for a large enough database of search terms, Google could preempt the intentions of their users in many cases, making for a smoother search engine user experience and returning better results.
The overall intention of Knowledge Graph is to streamline the search process for users. In many cases, the exact information that users are looking for is found in the Knowledge Graph summary meaning that some wiki sites will see a significant drop in traffic. On the other hand, Google says that it has seen a surge in the number of searches since Knowledge Graph was introduced2, which they have tentatively attributed to users doing deeper searches after seeing the Knowledge Graph summary and how it points to the information that users are really seeking. Google’s main competitor, Bing, has already shown signs of introducing a similar element into their SERPs3 in order to keep up with Google, and Yahoo is certain to follow. As Gary Marcus said of Knowledge Graph in The New Yorker, this may be the moment that we look back on later as the turning point when machines began to think- just a little bit4.