Google Update: BERT significantly changes search results

The search engine can now better understand the true intent of a search query by interpreting the relationships between individual words in a query.

What changes does the Google Update BERT? 

With the Google BERT update, website operators experience the biggest algorithm change in the last 5 years. About every 10th search query is affected by this update. With the BERT Update Google wants to deliver more relevant search results by improving complex long-tail search queries. Google can now understand the semantic context of a search query much better by using Natural Language Processing. The BERT Update is already rolled out in Germany.

Brief assessment of the Google BERT Update

BERT can be seen as a logical development of Rank Brain, Hummingbird and Panda. This update does not focus on indexing and ranking. Google is more concerned with analysing the context of a search query and displaying the best results.

What is behind BERT?

Behind the abbreviation BERT the following is hidden: Bidirectional Encoder Representations from Transformers. Behind this name is an algorithm model based on neural networks. With the help of NLP (Natural Language Processing) computer systems try to understand the difficulty of human language. For a better understanding: Through BERT, Google tries to better assess the context of a search query and to understand the relationships between individual words more precisely. This innovation is based on new calculation principles called Transformers: A word is put into relation to other words in the sentence. This means that not only the meaning is tracked word by word. This new feature pays particular attention to prepositions and the position of certain words in a search query.

Why is this update so important?

Google has found that every 6th search query is submitted for the first time every day. The typed in search queries are becoming more and more similar to real, human language, also encouraged by technical innovations such as Voice Search. It is predicted that in the next few years every 2nd search query will be launched via Voice Search. In addition, the length of search queries is increasing more and more. Already now, two thirds of search queries are assigned to the long-tail segment. Users would like to have an exact answer from Google in seconds to long question sets. BERT and the technology behind it contribute to this. For many years, Google has been working on neural networks with which more extensive search queries lead to even better results.

Which search queries are affected by BERT?

In general, only long queries (Long-Tails) are affected by BERT. Through BERT the typed in long-tail can be better understood. This also applies to language-based search queries.

What impact does BERT have on webmasters and SEOs?

This cannot be answered so easily. There is no simple tactic to make a website rank better. It becomes more and more important to write the web page not for the Google algorithm, but for the buyers and users of the web pages who also want to interact with the page. "Content is king" is the best long-term strategy for good ranking.