In late 2015, Google in a characteristically cryptic fashion, announced that it had begun utilising an AI named RankBrain, to help determine its search results. By June 2016, Google had confirmed that RankBrain was now being used to help process all of its 1 trillion+ yearly searches. This is of particular importance, as Google has now also confirmed that the RankBrain AI is its 3rd most important organic ranking factor, after quality backlinks and content (these two being in no particular order).
With RankBrain now factoring in processing every Google search query, what does this mean for SEO, and for Google search users?
What is RankBrain?
Google refers to RankBrain as a ‘machine learning technology’. This means a program where the computer is able to teach itself how to perform a new task, rather than having to be manually provided with further programming. The stated aim of RankBrain is to provide more accurate results for user’s search queries by learning to better understand and interpret searches for content that may not contain the exact terms being searched for.
How does it work?
To provide more relevant search results, RankBrain utilises advanced mathematical and language semantic processing capabilities, which represent a generational leap in effectiveness compared to Google’s earlier methods of synonym matching. According to an interview with Google’s senior research scientist Greg Corrado that appeared on Bloomberg:
“RankBrain uses artificial intelligence to embed vast amounts of written language into mathematical entities — called vectors — that the computer can understand. If RankBrain sees a word or phrase it isn’t familiar with, the machine can make a guess as to what words or phrases might have a similar meaning and filter the result accordingly, making it more effective at handling never-before-seen search queries.”
As RankBrain is able to learn ‘on the fly’, it is able to continuously gather data on how people search and what they were actually searching for, and apply this learning to determining future results.
An example of this would be a search for “New Zealand presents”. RankBrain might analyse this query and determine that it is similar to a more popular search like “nz gift ideas”. It may then, unknown to the searcher, decide to translate the first search to the second, Google would do this because it possesses more data for popular searches and can likely deliver better quality, more relevant results.
Rather than a new search algorithm like Panda or Penguin, RankBrain appears to be integrated as part of Google’s exiting Hummingbird algorithm. Nor does RankBrain replace Google’s earlier Knowledge Graph database, which stores information about things and the relationships between them. Instead, RankBrain is another of the many ranking factors utilised by Google to determine its search results.
What’s the point?
Besides search query refinement, the purpose behind RankBrain appears to be helping Google deal with the increasing volumes of long-tail search queries, with around 15% of their three billion daily searches, having never been seen before. By continuously gathering and interpreting this search data, RankBrain is able to see patterns between seemingly unconnected searches, and form an understanding between them. Previous search refinement methods employed by Google were forced to rely heavily on human involvement, such as creating synonym databases.
What RankBrain Means for Users and SEO
For people using Google’s search engine, RankBrain goes almost entirely unnoticed. In the earlier example of a search for “New Zealand presents”, the user would be unaware that RankBrain may have altered their search to “nz gift ideas”.
For those concerned about the SEO ramifications of RankBrain, remember that it’s primarily designed to help users find the best quality, and most relevant content. By producing unique content that answers user’s specific questions (and the questions people haven’t asked yet!), and quality evergreen resources, a site will still see significant improvements to its search rankings.