It has recently been revealed that Google is now using a machine-learning artificial intelligence system known as 'RankBrain' to interpret searches and submit the adequate information for the query. But how does RankBrain differ from existing algorithms and the way searches are processed?
What is RankBrain?
Firstly, RankBrain isn't a replacement of Google's current search algorithm, which is the core system that searches through billions of web pages and ranks them. Rather, RankBrain is a part of that algorithm, helping to interpret what a user is searching for.
For example, in the past the search query 'premier league' would have returned only results that contained that exact phrase on a page. In recent years, Google has become increasingly more sophisticated in their interpretation of these queries and may return results such as 'football teams', 'Arsenal' or 'Manchester United'.
RankBrain is considered a new way to process these queries, building upon the ecosystem previously used, armed with the ability to embed large amounts of written language into mathematical entities called vectors that a computer can understand. This means that if RankBrain encounters any unfamiliar words or phrases it's able to make a guess as to what the user is actually searching for and process this to show results with similar meanings. Thanks to RankBrain, Google can now deliver more accurate results for brand new and complex search queries.
How important is RankBrain?
Greg Corrado, one of Google's senior research scientists, states that RankBrain has interpreted a "very large fraction" of search queries and is now the "third-most important signal contributing to the result of a search query".
Any ranking signal that is impacting a significant number of queries is very important, particularly if Google is telling us it is their third most important factor - this is something to sit up and take notice of.
How will RankBrain impact SEO?
RankBrain is an artificial intelligence system that attempts to understand the meanings of and relationships between words and phrases. If content focuses on a very limited set of keywords, then it's unlikely to benefit from the changes RankBrain brings.
Content should be detailed, keyword rich and make use of a wide range of related words, phrases and synonyms to take advantage of RankBrain's artificial intelligence.
How can I find out more about RankBrain?
Google has advised that if people would like to learn more about word vectors, then they should read this article, or this longer research paper.