Fresh November news from Mountain View: Google has just introduced the presentation of the new algorithm in its talks, the one that will be defined DeepRank. In this article we will explain everything you need to know about its application dynamics which, for once, are a little less nebulous than usual thanks to the presence of an excellent explanatory video.

The video illustrates a kind of intergalactic guide for Googlers curious to know more about how searches work not only in terms of numbers and machines, but also in terms of people working on a specific goal.

What is DeepRank?

DeepRank is a Google algorithm which bases its operation on machine learning and on the capabilities of the system, already present, of process natural language in a mash more digestible by the mathematical and logical algorithms of the search engine.

In practice, Google takes one more step towards users, who will be able to use a language that is ever closer to the real one, and not to a composition of elements specifically designed to be understood by him. Ask Google for something, therefore, is increasingly similar to asking a question to a learned professor, without having to modify the grammar too much to adapt to his language.

DeepRank has a signal that understands the relationship between human terms: in other words, it helps Google to better understand the language of humans, even if they use very colloquial or spoken jargon. In short, it will be even easier for the algorithm to understand the connections between the various words and idioms that do not have a literal translation pertinent to the research object. In the article on BERT we gave the example of "catching a crab", but today we will propose a slightly different one, always to keep the line of example: "remain stone". Or maybe putty. Thanks to DeepRank, the search engine doesn't just look for the word rock, it finds it connection between the elements of the sentence and identifies the most relevant result, or “to be amazed”. In this way the user has to think less, because Google compensates for the logical holes that the natural language does not foresee.

Naturally, such a project requires very complex and elaborate experimentation work, so much so that the tests carried out on the algorithm have been (and still are) very many, so that Google is prepared to receive as many colloquial expressions (or just not very clear) possible. In other words, with DeepRank Google becomes more and more within the user's reach. Closer and closer to the medium-low user, who does not have in-depth knowledge of the linguistic or computer rudiments necessary to communicate in the old-fashioned way with the algorithm.

A new step towards the future of Google, increasingly simple and intuitive. How will this algorithm change the lives of webmasters? In reality, if the webmaster is already doing an exemplary job with the internal SEO of his site, it is very likely that Deeprank will not affect the positioning of his site, except in a favorable way. Especially for the benefit of long tail keywords!