RankBrain is Google’s machine-learning artificial intelligence system, which plays a significant part of the company’s overall search result ranking algorithm. This system is surging in importance due to its role in sorting out results for specific queries based on several signals. However, not only does it sort out the most relevant web pages, but it also affects the response to these queries, particularly complex or obscure ones.
First functioning in early 2015 before Google officially unveiled it in October of the same year, RankBrain is an artificial intelligence that is part of Google’s search algorithm, Hummingbird. RankBrain has two primary duties. The first goal is understanding search queries, or keywords, and processing them. The second goal is understanding the context of how users interact with the search results and rank them in terms of relevance.
RankBrain, as well as other parts of the Hummingbird search algorithm, learn to understand the queries through machine learning. Machine learning occurs when a computer system takes a user’s frequent behaviors and patterns to predict future behaviors and patterns. These predictions influence Google’s future decisions, such as the ranking of search results or the placing of specific results in related keywords.
RankBrain specifically targets “long-tail” queries or keywords that are too long or too complex to lead to a variety of results. The system looks at the initial queries and how an unsatisfied user may change them to something more concise, but related, and how the user interacts with the results.
If the user was satisfied with the later, related queries, RankBrain will make a note to look for similar keywords and associate the unsuccessful queries with the successful ones. RankBrain also looks at the human context for web page visits. If the user is satisfied with the later queries’ results and clicks on the content, the system will associate those successful web pages with the unsuccessful queries in order to improve the results.
When Google revealed the system’s existence in October 2015, a representative of the company claimed that it processes a large fraction of 15% of the queries that Google would have never seen before. These queries are often either new, obscure, or too complex or long to lead to a successful search. While this fraction of queries is small, Google benefits from improving the query results in order to encourage more activity and determine more successful content in general. RankBrain’s success with Google has increased its range. News in June 2016 revealed that Google uses RankBrain on every single query the company processes, instead of just lesser-known queries.
For more information on RankBrain and Google algorithms, check out part 1, part 2, and part 3 of the series.
Ready? Let’s get to work. Call (419) 540.4437 or email toledoseoforgrowth@gmail.com to get started today.