Google has always been a forerunner in search engine technology and just recently, the search giant has announced the addition of another layer to their increasingly complicated ranking algorithm. This time, the update is more focused on machine learning—something that has been paving its way into search engine algorithms for quite some time now, with more IT giants embarking on their own forays into the connection between machine learning and search. Google’s brand new update is called RankBrain and digital agencies, business owners, and web administrators should always be on the look-out for the changes that the company enacts, particularly when it comes to their search and ranking algorithm. Failure to keep up with these changes and updates can mean your website’s downfall in terms of search rankings. What is machine learning and what does it have to do with search engine algorithms, the following are the key things you should know:
- RankBrain is the name of the artificial intelligence that Google has been using to handle and decipher different search queries. A very large fraction of the search queries that users type per second (this accounts for millions of queries per second) are being interpreted by this new AI system. At least 15% of the new searches that RankBrain receives had never been seen before. The system guesses what words or phrases may have a similar meaning to the searched terms and then filters the results in order to make the engine more effectively in handling never-before-seen searches. This way, listings can be more reliable and relevant to the search. When tested against human/manual interpretation/guesses (eyeballing which websites Google search users will rank on top), humans (Google engineers) guessed right 70 percent of the time, while RankBrain (the new AI) had an 80 percent success rate, which shows that Google searches are becoming much more reliable and thanks to these brand new innovations.
- RankBrain is an effort to remove as much human element from search algorithms because humans can only process what we know and recognize, and collecting/gathering data to learn from is still a very manual process. All rules governing how current search algorithm works, ranging from content crawling to link tracking, and all other activities were all written by a human being and human beings wrote them because they recognized patterns and adjusted algorithms accordingly. It is true that humans have exceptional capabilities to recognize patterns but are still very slow at doing so, not to mention prone to errors. This is why enlisting machines to help govern rules that run these algorithms is important to speed up the process.
- RankBrain is still a computer program—an incredibly powerful computer program, running on previously written protocols and taking predetermined actions, but with some level of learning, which theoretically allows it to evolve processes over time, albeit gradually. Just the same, a human is still responsible for spot-checking its evolution and watching over its progress.
- Google has expressed that signals coming from RankBrain are considered the third most important of all the signals that they base their rankings on. RankBrain looks at the intent of the search. This means that searching for “homes in so-and-so” will follow a synonym code to bring up results matching “houses in so-and-so place.” These codes are still largely written by humans. The hope is that as searches come, the AI will be able to track which results users are clicking on and begin making these correlations without needing the manually written code. This will help the engine to keep up with the never-before-seen slang terms, new topics, and trending queries.
- This update will not affect SEO a whole lot just because it is an algorithm change that allows users to have a better experience. It shouldn’t change how marketers tackle their SEO but website owners should focus on well-designed and fast-loading pages that give searchers great user experience and offer excellent content.