Google sure knows how to throw itself a birthday party. During their 15th birthday party this past Thursday, they took the opportunity to let the world know about one of the largest updates they have done in years. The internal code name for the release is called Hummingbird, and one big part of this relates to adding the ability to do comparisons and add filters to a technology that Google calls the Knowledge Graph.
In short, the Knowledge Graph is the beginning of Google’s efforts to move search past simple use of keywords and links to determine what to show in its search results. Instead Google now has mapped the relationships between many things, and can use that to answer more complex queries. For example, consider the Google response to the query “apples vs kiwi”:
Or, consider the query “How far is santa cruz from here”
To provide these results, Google must leverage an understanding of the relationships between different things, that apples and kiwis are both fruits and therefore certain information would be interesting to compare, and that I live in New York (which is “here” in the sample query above), Santa Cruz is another place and it is in California, and that I might want to get driving directions to get there.
The queries you can put together can even become more complex, and Google can now string a succession of queries together. Consider the following sequence of 3 queries:
Query #1: “show me pictures of the Empire State Building”
Query #2: “how tall is it”
Query #3: “show me pictures of its construction”
What this query sequence shows is that Google is remembering the context of the conversation, much like humans do. In queries 2 and 3, Google knows that “it” refers to the Empire State Building, which is exactly what would happen in a natural language conversation. Knowledge Graph searches have had the capability of handling “conversational search” for some time, but now with Hummingbird Google has expanded this to the rest of its search engine.
It is important to understand that the Knowledge Graph remains in its very early stages. At the moment it is focused on handling more complex search queries and relationships. It is our opinion that over time that search will become more natural language in general, and that will evolve to handle increasingly complex concepts.
That is what makes Hummingbird so interesting, because it is a tangible step in this direction. It now has the ability to recognize when the word “it” is a reference to a prior object or name. It has the ability to recognize some simple relationships between different types of objects and people. Just like a you might teach a toddler to recognize certain things, Google is steadily enhancing its Knowledge Graph to understand more and more complex concepts.
The way that search evolved caused a heavy focus on keywords, and in fact, many people learned that the best way to formulate search phrases to get what we want involves finding the minimum number of words in the search query. We tolerated that for a number of reasons, including:
- The method we have had for providing input was a keyboard. As technologies such as voice search become available the demand for more and more natural ways of interacting will increase.
- There was a lack of sophistication in the possible responses. As the complexity of what is possible goes up, the need for more natural interactions will go up as well.
In short, this is evolving into the computer we used to see on Star Trek, where very human interactions with the computer are totally natural. Over time, this shift will reduce the focus on mechanical SEO. Yes, keywords will still matter, but what will matter more and more is the connections we make, both to relevant topics, and relevant people.
Networking with peers and building our own audiences will be more important than ever, as is creating content that people interact with, or cite in their own writings (articles or in social media). It also places a major emphasis on AuthorRank as a way of helping Google identify the true experts on different topics. These are all things that Google wants, and its good news for those who know how to do a great job producing quality content.