Google Bert: Are there really effects on the SERP?

Pandu Nayak had called it "the biggest leap forward” in the history of Google search engines…

so much so that the world of experts had trembled a little at the idea of ​​having a hot potato in their hands. One week has passed from the Google Bert implementation, which should have impacted at least one in ten queries. Yet this "revolution" takes a while to reveal its effects.

Waiting for (the effects of) Google BERT

On September 25, the vice president of Search, Pandu Nayak, announced the introduction of Google BERT, an innovative artificial neural network technique to facilitate the understanding of human language. The acronym stands for "Bidirectional Encoder Representations from Transformers”, BERT. This system makes it possible to automatically process written or spoken information in a natural language and to better understand what is important in queries.

Many words, many interesting expressions and a breakthrough which, according to Google, would have sanctioned the biggest leap forward in the last 5 years in terms of search engines, impacting 10% of queries in English in the United States. Yet, a week after implementation, BERT is demonstrating a much smaller impact than hoped for. We tried to figure out why.

What is BERT for, in a nutshell?

BERT's goal is to affect long, conversational queries, or for searches where prepositions have an influence on the meaning. Before BERT, the algorithm could not fully understand the context of words – but now the code's approach is much more natural and personal.

All very nice. But why is this “Google revolution” not delivering the desired results? There are many possible answers. One of them is that i currently available trackers are not advanced enough to measure the impact of this update. In practice, it is not possible to accurately measure long tail keywords and how BERT has recalibrated the search engine's approach towards them.

What has changed?

Those expecting a Penguin or Panda-style upheaval have been disappointed: at the moment the main results (in the United States, because the change only concerns the English language at the moment) are substantially unchanged compared to two weeks ago, with some negligible swing.

Google BERT: Do you need to make changes to your website?

At the moment, BERT does not require onpage optimization for its effects to become tangible. The function of the system is to help Google better understand user searches, i.e. the search intent, but at the same time it is based on people's natural language. For once, in short, it is the algorithm that adapts to the way people search – and website administrators just have to wait to benefit from its fruits. Naturally, as long as the content created for users respects the required level of quality in order to be taken into consideration.