BERT or how Google understands better!


The internet is full of articles online about this news, but we wanted to update you on this nonetheless. In this article, we’ll take a quick look at what BERT is and point you to several resources that’ll give you a broader understanding of what BERT does.

Welcome, BERT!

At the end of October Google launched what they call

“the greatest advance in the last five years, and one of the greatest advances in the history of the search”.

Or what is the same, an update of its BERT algorithm, which, simplifying, is Google’s way of better understanding the way humans speak.

What does BERT stand for?

BERT stands for Bidirectional Encoder Representations from Transformers:

  •    Bidirectional: this is quite simple; it means that something can go in two ways
  •    Encoder: another way of saying that it is a software program
  •    Representations:  representing someone or something
  •    Transformers: this is already more technical, and it means that a transformer in the NLP (Natural Language   Processing) architecture solves sequence-to-sequence tasks while handling long-range dependencies.

What is the difference between BERT and RankBrain?

BERT is very similar to RankBrain in the sense that it is a machine learning algorithm that aims to better understand the queries and content of a web page, but in no case replaces this or other language algorithms, but it complements them.

How does it work?

BERT is conceptually simple and empirically powerful. It is technically based on neural networks for the pre-training of natural language processing (NLP), which allows Google to understand queries that are more similar to human conversations, naturally and based on conversations. He has been trained with a general-purpose “language comprehension” model with a large text corpus, such as Wikipedia. What is worth noting here is that Wikipedia is not optimized for search, it is true that they have titles and descriptions with keywords, but the content is not optimized. So, it is an excellent training ground to understand the context.

Does BERT have any impact?

Yes, it will! Google says that the impact of BERT will be massive since they have calculated that 10% of the total queries will be affected. It also says that BERT does not assign values ​​or positions to specific pages or websites.

Can I “optimize” for BERT?

Don’t try to optimize for BERT, try to optimize your content for humans

The traditional “optimization” approach does not fit with BERT. From the point of view of SEO, there is little that can be done in this regard. With the BERT update, the opportunity lies in focusing on the content. It is not necessarily about creating extensive content that talks about different things or topics, it is rather about answering a search engine’s question as quickly as possible and providing value. The content was written for humans, not for search engines, as it has been doing until now.

Therefore, looking at what position we appear in will not help us at all, what we should do is observe the search panel and not forget the analytics data to ensure that our organic traffic continues to rise. We are much more likely to lose traffic if we are a blog or an information page.

Our opinion is that, at the moment, BERT will only affect the results of searches with long-tail keywords, those complex questions that are difficult to answer if something as subtle as the intention is not understood.

“Optimizing” from now on means that we can concentrate on good and clear writing, instead of subjecting the creation of our content for machine understanding.

Keep producing quality content that fits your users’ goals and make your site as good as possible. So, we’re not going to present a silver bullet for optimizing for the BERT algorithm because there is none.

And if we want to do well when it comes to keyword ranking, we should be specific in answering search queries and using more long-tail terms.

We are moving towards a world of voice search that is increasingly easy to use. With sales of devices like Google Home, Amazon Alexa and many others constantly increasing and voice search as a real option on all phones, it’s only a matter of time before people start using it constantly on traditional search methods, which will make search phrases even more diverse and difficult to return an exact result.

Which other Google products can be affected?

They have already confirmed that it is used globally and in all languages ​​in featured snippets. Although they insist that only searches will be affected, the Assistant may also suffer.

At the moment Google confirms that it is not being used in ads, but who knows when it will be so!

Should you need far more in-depth technical info, please visit this site.

Cara Walsh
Content Crafter at

Though sometimes it seems like navel-gazing, Cara has written many blog posts about online marketing for small businesses, B2B marketing and using SEO to drive traffic and now she is writing them for Quarizmi.


Leave a comment:

Your email address will not be published. Required fields are marked *