What is BERT? Google’s New Algorithm Update

27 October 2019
Reading Time: One Minute

Google introduced BERT, the biggest update to the search system since RankBrain was introduced back in 2015. It is not an algorithm change but an open code system designed to understand better the nuances of the search queries. According to Google, BERT will affect about 10% of the search queries in all languages, mainly Featured Snippets.


What is BERT?

BERT, an acronym for Bidirectional Encoder Representations from Transformers, is a natural language processing (NLP) system based on AI. Not to get too technical, the system helps computers to understand human language the same way people understand it.

In relation to Google, BERT helps the search engine to understand the nuances and the context of search and to improve the results accordingly. The system will mainly be used to Featured Snippets in search.

For example, before BERT the search query “2019 brazil traveler to usa need a visa” showed results that detailed the visa required for Americans who plan to visit Brazil. After BERT, Google will display the relevant results:

google BERT - example

More examples of BERT impact on Google’s Organic Results:

BERT & google search results

BERT on SERP - before and after

BERT is an addition to Google’s ranking system. In fact, it is another tool used by Google to understand the meaning of the search queries and websites’ content. In some searches, Google will continue to use RankBrain, Google’s first AI system. In other searches, Google may use several methods, depending on the context.


Will BERT affect SEO?

For now, BERT’s impact on search rankings is not significant, mainly since BERT is supposed to impact long-tail, conversational keywords, and not necessarily the short keywords usually tracked by SEOs.

As usual, Google doesn’t recommend to optimize for BERT and to continue producing quality content for human users.