RankBrain is an AI (artificial intelligence) system. Not to be confused with Google Brain, although certainly part of it. RankBrain is not an algorithm update in itself either, it is now part of the overall Google search algorithm. Rumours are that it is the third most important signal of over 200 ranking factors.
RankBrain marks a significant shift in direction from the Google search team. This is likely over time to return less human maintained and updated algorithms such as the (in)famous Panda and Penguin. Where updates need to be triggered by a human, revised and often reversed as well when things don’t work out as expected.
RankBrain does not require humans to update
Being an AI system, RankBrain has the capacity to sift through large amounts of data, and we all know Google has plenty of it, and teach itself how to use that data and the lessons it extracted. Then, apply them to further decisions when retrieving answers to user queries.
RankBrain does not need a human to push a button somewhere to update itself, by its very nature it will do that by itself. This is a fundamental shift, and some have compared it to a toddler in its infancy, well AI and ML (machine learning) are in their infancy, but RankBrain, like all AI systems, will learn and get better over time, at least in theory, I can’t say I see an awful lot of getting better just yet.
What does RankBrain actually do?
RankBrain is initially and primarily concerned with what we call long-tail queries, these may be ambiguous and previously impossible to be “understood” by a computer system. RankBrain is also concerned with never seen before queries or queries Google has never come across before.
Very close to this are two notions: meaning and the use of synonyms.
Since this is a post for non-techies (mere mortals), the simplest explanation would be that currently Google largely needs a given keyword/keyphrase, to be present both in the searcher’s query he/she types on Google, and then these same keywords/keyphrases in the pages Google retrieves. RankBrain and similar initiatives and updates attempt to extract meaning and intent from the user query and will not need exact keywords to be present.
This would be when for example you type a search for “SEO consultant”, and Google will return a page that has “SEO expert” instead. In this example, Google used the synonym to modify the query. This is a very simplistic explanation for good reason.
Expect to see much more of these over time. Again as least in theory.
Other signals and initiatives related to RankBrain
Well, largely anything that is not about exact match keywords is related to RankBrain and the general direction of Google’s ultimate goal of being a search assistant. Some of this past and present are:
- Google Now
- Google+ Authorship (now scrapped)
- Knowledge graph
- Semantic search (extracting meaning)
- Use of synonyms
- Direct answers
There are more but these are a few of the top ones.
How does RankBrain affect your SEO
For now, it is not clear and despite being a major ranking signal no great specific changes to SEO strategies are needed unless your strategy or that of your SEO consultant or firm relies heavily on link building and exact match keywords. In which case moving to a more user, and arguably RankBrain friendly approach may well be a good idea.
An SEO and content marketing strategy that relies on solving problems and specifically answering and anticipating users’ questions, is likely to be more in line with RankBrain, Hummingbird and any future AI and ML initiatives that Google may bring in. These are almost a certainty.