What is RankBrain?
RankBrain is Google’s AI “machine-learning” signal, established in early 2015 and now rolled out globally. According to Google, it’s the 3rd most authoritative signal out of Google’s 200 major ranking signals, and makes up part of the Hummingbird algorithm (alongside Panda, Penguin, Payday, Pigeon, TopHeavy, Mobile Friendly and Pirate).
What does RankBrain actually do?
From our research we’ve identified the two main things that RankBrain is responsible for:
RankBrain recognises user intent
The first thing RankBrain does is interpreting unique, long-tail, and ambiguous searches (of which, 450 million occur a day) and identifying their similarities, to retrieve the most useful pages for the user. It uses “word vectors” to mathematically identify words which share some form of semantic relationship, rather than merely recognising letter sequences.
RankBrain ranks pages in Google based on their relevance to a query
Secondly, and as a result, it directly contributes to whether a page ranks in Google, as it’s able to understand content and surmise what a page is about.
The thing that differentiates RankBrain from Google’s Knowledge Graph, which also analyses the relationships between words and things, is that it is devoid of human input and recognises patterns unaided, by processing historical searches and making predictions from them. This process is offline but, if correct, goes live.
Read more ›