Google BERT: the next big change in SEO

In Oleoshop we always keep our eyes open to every change in Google algorithm . We love keeping you updated with each update  or new functionality, like the brand new changes in snippets robots.txt variation,  favicons, that has a lot to do with SERPs optimization on mobile devices… 

We are so obsessed with helping you understand the search machine par excellence that, believe us, BERT is going to change the way Google understand user´s search intent completely.

Of course, before understanding what changes, we must be clear of what is behind this acronym. Because, although it looks a proper name, in fact, we are talking about an acronym. 

The acronym of Bidirectional Encoder Representations from Transformers form the famous word we are talking about. The truth is that if you read it without stopping, we can understand why they have opted to look for a more kind solution refer to it. 

Behind BERT it is Google workhouse: provide its search machine with the capacity of processing natural language (NPL) in the same way users are able to interact with other people. 

This, by itself, it is not new, we know they have “trained” their algorithm for years to understand users. Algorithm updates like Google Hummingbird had already aimed to make of semantic web a reality further those simple keywords. 

We already knew that machine learning is used through its known Rankbrain to, through these artificial neuronal networks, simulate human brain behavior and learn by itself to refine searches and offer more relevant results based on this knowledge.  

Then... what does BERT change? 

Well, as it usually happens with Google, everything is related. The appearance of that new technology means the capacity to concrete the semantic web and neuronal networks, in a formula to understand the context of a word and, in this way, understand it more efficiently. 

The key is given by the word itself, we talk about Bidirectional Encoder Representations from Transformers. When we say it is bidirectional, we mean it is capable of establishing the relation among all the words in a sentence, according to the meaning they modify one another. 

We do not mean reading the whole sentence in a semantically logical way, from left to right, and re-read it backwards. Bidirectional is in the process of  contextualization. 

It is better to see an example isn´t it?

On Google blog itself, the vicepresident of the company, Pandu Nayak, announced BERT to society, talked about concrete cases.

It is not difficult to see that it will be determining. In the example we have just talked about, and others mentioned by Nayak in the post, we see how the keyword stops having an isolated value by itself. 

So far, we have understood that user looks for effort economy synthetically: better “free hotel Madrid” than writing “not too much expensive hotel in Madrid city”. Virtual assistants have changed the searching method:  voice search is a reality.

In any case, people keep using this kind of searches in a high percentage, discarding articles or prepositions. 

Partly because they are used to it, but in Google aims to another interesting aspect: 15% searches do not match with anything previously searched, are unknown. This means that not even the user himself can carry out a complex search, he needs to rely on the context to get to the resolution.