It’s been nearly 3½ years since Google first announced their usage of RankBrain (October 26th 2015, but it had started being rolled out early 2015, in multiple languages).
In that time, there’s been little in the way of details coming from G about what it is or how it works.
The result is that numerous SEOs have stepped up to fill that void with their own speculations and opinions, and in doing that, have caused all sorts of confusion.
This is my attempt to correct and clean up some of that mess.
(There is a TL:DR at the bottom if you want to skip the verbiage :D)
What does RankBrain do?
Though there isn’t much publicly available, what we do have is fairly specific:
“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.”
– Greg Corrado, from Bloomberg’s Google Turning Its Lucrative Web Search Over to AI Machines[1]}
Or, if you want it more succinct than that;
“… Lemme try one last time: Rankbrain lets us understand queries better. …”
Google receives a fair percentage of queries per day that it hasn’t seen before: 15% at last check.[3]
These may include misspellings and typos, elisions/omissions, unusual phrasing/syntactic structures, the wrong word(s) being used, negations (“not x”), things that have only just happened etc. etc. etc.
RB receives these weird, wonderful, and new searches, and attempts to identify existing