Everything You Need to Know About the BERT Update

Posted on October 29, 2019 • Written by Ben Smith
What you need to know about the BERT update

On October 25, Google announced that they were rolling out a major new update to their search algorithm. As Google is the dominant force in search, people in the Search Engine Optimization community are always sensitive to changes in their algorithms, and were immediately filled with questions: What exactly is BERT? How will it impact searches? How large of an impact will it have on rankings? And how can site owners optimize to benefit from it? This article will address all of those questions.

What is BERT?

BERT stands for Bidirectional Encoder Representations from Transformers. Google describes it as an open-source neural network-based technique for natural language processing pre-training; but what does that actually mean? In simple terms, it is a technology that trains systems to understand language more like humans do. Instead of just looking at the words in a query one by one in order, using a BERT model will allow Google’s search engine to examine the relationship between all the words in a sentence in order to determine the intent of the user’s search.

How Will BERT Impact Searches?

Google provided several sample queries as examples of how BERT will allow their search engine to better understand user intent. Take, for instance, the query “2019 brazil traveler to usa need a visa”. Previously, the search algorithm would not be able to understand the significance of the word “to” in the query; and so, analyzing the words “traveler”, “brazil”, and “usa” in isolation, it would deliver results relating to travel from the U.S. to Brazil in addition to travel from Brazil to the U.S. With BERT, on the other hand, the search engine will understand that “to” — a word which it previously would have ignored for being too small and common to search for — actually conveys important information about the intent of the user’s search, and thus know that the most relevant results would be ones about travel from Brazil to the U.S. rather than the other way around.

How Large is BERT?

Google has described BERT as the largest change to their search algorithm since their introduction of RankBrain five years ago. They expect it to impact 10% of all queries, in that it will change the results that rank for those queries. This includes Google’s featured snippets, which are also affected by BERT. While BERT is currently rolling out for English language queries, Google plans to expand it to other languages in the future. It is already being used globally in all languages for featured snippets. Notably, despite Google’s statements about the size and significance of the BERT update, people in the SEO community watching for ranking changes thought the visible impact seemed rather small compared to previous core algorithm updates, and automated tracking tools did not register the massive changes that might be expected from an update impacting 10% of all queries. The reason for this is that, being aimed at helping Google understand natural language, BERT’s biggest impact is on longer, more conversational queries. These types of long-tail queries tend not to be tracked by site owners, who typically focus on short, keyword-heavy phrases. BERT thus stands in contrast to previous Google algorithms like Panda and Penguin; which, being focused on cracking down on spam and low-quality content, had a much larger impact on the rankings for the short keyword queries that SEOs track. Therefore, despite BERT’s size, much of its impact may fly under the radar.

How Do You Optimize for BERT?

Whenever a new Google update comes out, the first impulse of site owners is often to ask how they can optimize their sites in order to benefit from the algorithm and rise in the rankings. In this case, however, Google was quick to state that there is nothing to optimize for. BERT has nothing to do with the content on your site: it is only concerned with helping Google understand the intent of queries submitted by users. To those hoping to rank well, Google offers the same advice as always: write your content for users, not algorithms. Google’s goal is to deliver its users the most useful and relevant results to their queries; and whatever the specifics of their algorithm updates, they are all aimed towards rewarding great content.

Conclusion

Google’s BERT update is a large change to their search algorithm designed to assist it in understanding the intent of user queries. Though it is a large update affecting 10% of queries, the scope of its impact might not be easily visible to site owners, who generally focus on tracking shorter queries than the ones BERT is aimed at. And because BERT looks at queries rather than site content, there is no way for site owners to optimize for it, other than by producing high-quality content that is useful to searchers. With this information, you are now fully informed about the nature and potential consequences of BERT, and can make a knowledgeable decision about how to proceed.

Want to improve your website’s SEO?

Get in Touch

References

“Understanding searches better than ever before” by Pandu Nayak. Google Webmaster Blog. October 25, 2019. https://blog.google/products/search/search-language-understanding-bert

“Welcome BERT: Google’s latest search algorithm to better understand natural language” by Barry Schwartz. Search Engine Land. October 25, 2019. https://searchengineland.com/welcome-bert-google-artificial-intelligence-for-understanding-search-queries-323976

“Why you may not have noticed the Google BERT update” by Barry Schwartz. Search Engine Land. October 28, 2019. https://searchengineland.com/why-you-may-not-have-noticed-the-google-bert-update-324103

“Google: There Is Nothing To Optimize For With BERT” by Barry Schwartz. Search Engine Roundtable. October 28, 2019. https://www.seroundtable.com/google-bert-no-optimization-28435.html

“The Google BERT Update Felt Small But Was Big” by Barry Schwartz. Search Engine Roundtable. October 29, 2019. https://www.seroundtable.com/google-bert-update-felt-small-but-was-big-28444.html

Ben Smith

Written by Ben Smith

Ben Smith is a Researcher at Fruition, specializing in Google's Algorithm changes. Ben is a graduate of the University of Denver’s Mathematics program, and he enjoys learning about Google’s search algorithm updates. He's a vital asset of the Fruition team, and he one day hopes to publish a book. In his free time, you can find Ben enjoying the outdoors of Colorado.

Related Articles:

View All Blog Articles

From our team & partners