Image taken from https://searchengineland.com/faq-all-about-the-new-google-rankbrain-algorithm-234440
Google uses a machine-learning artificial intelligence (AI) system called “RankBrain” to help sort through its search results and find the ones deemed most relevant for particular searches. Hummingbird is the overall search algorithm, think of a car’s overall engine. The engine itself is made up of various parts, with RankBrain being one of the newest. In practice, AI is used to refer to computer systems that are designed to learn and make connections. Machine learning is a computer science that uses statistical techniques to give computer systems the ability to "learn" with data, without being explicitly programmed. In terms of RankBrain, AI and machine learning seem synonymous.
What Does RankBrain Do?
RankBrain is mainly used as a way to interpret the searches that people submit to find pages that might not have the exact words that were searched for. RankBrain uses machine learning to insert massive amounts of written language into mathematical entities for the computer to understand. All learning that RankBrain does is offline. To account for non-exact word searches (if RankBrain sees an unfamiliar word or phrase), the machine will make a guess as to what words or phrases might have a similar meaning and filter the result accordingly.
How Does RankBrain Work?
Launched in 2015, RankBrain is Google’s machine learning algorithm; its goal to help Google understand certain types of search queries, predominantly the ones that Google has not seen before. Nevertheless, it can affect popular queries, too.
RankBrain uses series of databases based on people, places, and things to create the algorithm and machine learning processes. These queries are broken down into word vectors using a mathematical formula to give these words an address and return multiple related results. Over time, Google refines the search results based on user interaction and machine learning to improve the match, or linguistic similarity. Words go in and are assigned a mathematical address. Words are then retrieved based on the users’ query and the words the algorithm locates in the “best fit” vector. These word “interpretations” are used to return results to the user. To make those results more relevant the next time, data is fed continually behind the scenes into the machine learning process. It is important to note, RankBrain is not meant to replace ranking. Instead, as stated by Search Engine Journal, “it is to help better understand queries to deliver the best search results, particularly for negative oriented queries, such as queries using words like without or not.”
RankBrain learns, but not on the fly. As Google has previously stated, RankBrain “can see patterns between seemingly unconnected complex searches to understand how they are actually similar to each other.” Remarkably, the way RankBrain affects the actual search results is not as important as many believe. Gary Illyes revealed “that while RankBrain is “touching” all the search results, it does not necessarily change the order or how those results appear in searches.”
Similarly, advertising spend does not influence RankBrain. Just like with regular organic search results, it does not make a difference whether a company spends $1 million on AdWords advertising or $0. The role of AdWords has no impact on RankBrain or organic search results.
Google has also confirmed that no matter how many times a website is shared or favorited, social sharing does not have any influence over RankBrain results. Search Engine Journal explains that in addition to having no control over future access to social data, social signals are too noisy to be used in search rankings. For this reason, social sharing is not used for RankBrain.
Can You Optimize for RankBrain?
You can optimize for RankBrain by just writing naturally. Writing in natural language is writing content in a way that sounds human. If you try to write like a machine then RankBrain will just get confused and probably just push you back. You can assess content quality by reading the content aloud. Doing so will make things that make your content sound unnatural, such as excessive use of keywords, become very apparent.
RankBrain results are designed to change and bring back better results. Therefore, optimizing for it would be like trying to hit a moving target all the time. So, what makes RankBrain different? Before RankBrain, Google’s algorithm was hand-coded, now RankBrain tweaks the algorithm automatically. Depending on the word search, RankBrain will increase or decrease the importance of backlinks, content, domain authority, length, and so on. Then, it looks at how Google searchers interact with the new search results.