Trying to understand what determines the positioning of a site in search engines is the job of every SEO Master , but it is a real undertaking: in fact, some factors have been well known for some time – presence of search keywords, number and quality of links pointing to the site, etc … – but there are hundreds of signals that contribute to the identification of PageRank.
Every so often, however, there are news: the news of the moment – leaked thanks to Bloomberg and Search Engine Land – is that RankBrain also affects the results returned by Hummingbird (the Google algorithm in force). What is this new implementation and how does it affect the SERPs?
The argument is still nebulous, but what we are told is that it is a technology equipped with ” artificial intelligence ” which, by analyzing offline blocks of historical research done by users over the years, tries to predict (without the human intervention, for this we speak of “machine learning” ) which searches are similar in terms of meaning even if in terms of contents / search keywords typed are different, thus providing similar SERPs as a result.
It is not just a matter of understanding derivatives and synonyms of the terms entered in the search, which Google was already able to do: for example, searching for “gym shoes” we could also find “running shoes” or “sports shoes” among the results returned, because they are strongly correlated.
The way RankBrain refines the search is more complex and is perhaps best understood by the example that was reported by Search Engine Land, namely:
if a user writes “what is the consumer at the highest level of the food chain called”
RankBrain understands that the search is similar (and therefore returns similar SERPs) to the search that follows, ie “level at the top of the food chain”
A big step forward for Google, which continues to refine the results it provides to its users, even for those long-tail queries never typed before.