BERT- Google’s New Algorithm

Lisa Danca - Marketing Associate

Written by: Lisa Danca

Category: Search Engine Optimization

01.14.2020

In our last blog post, we covered the biggest Google algorithms. Today, we’re here to discuss Google’s latest one, BERT, which may be the search engine giant’s biggest algorithm yet according to SEO experts! 

BERT stands for Bidirectional Encoder Representations from Transformers and it rolled out on October 21, 2019. It will impact 1 in 10 of all search queries. It is the biggest new algorithm since Google released, RankBrain, five years ago. 

BERT started with English queries and Google is slowly starting to apply it to other languages. On December 9, 2019, Search Engine Land reported BERT is rolling out to Google search in over 70 languages. BERT is addressing a gap in previous architectures. It helps computers understand language the way humans do. Google has said their new search algorithm helps their search engine to “better understand the nuances and context of words in searches and better match those queries with more relevant results.” 

Besides search queries, BERT will also impact featured snippets. Featured snippets, also known as “answer boxes,” are the text on top of Google’s organic results below the Google ads. Featured snippets aim at answering the user’s question right away.

A screenshot of the Google search for 'what is a featured snippet.'

Now, if you read our last blog post, you’re probably thinking BERT doesn’t sound that novice. In 2013, Google rolled out Hummingbird, the first algorithm to understand the context of search queries.

What makes this update different is how it goes about processing and understanding languages. BERT has pattern recognitions, which helps Google to better understand how humans communicate. Until now, Google had a hard time answering complex or conversational based queries.

Another reason why BERT was invented is queries are getting more conversational. Users are treating their devices like companions and voice search is becoming more popular.  This new update won’t really affect short search queries. It will focus on long-tail queries.

As stated in this article about BERT, in human beings, language is innate and nurtured. In machines, nothing is innate.

Google’s John Mueller has said, “what we’re trying to do with these changes is better understand text. Which on the one hand means better understanding the questions or the queries that people send us. And on the other hand better understanding the text on a page…The text on the page is something that you can influence. Our recommendation there is essentially to write naturally…So if there’s anything that you can do to kind of optimize for BERT, it’s essentially to make sure that your pages have natural text on them.”

Below are some examples Google has made to help marketers better understand their new algorithm. 

Example 1

A screenshot of Google explaining its new search algorithm, BERT. The search is '2019 brazil travelers to usa need a village.'

BERT allows Google to better understand words like ‘to’ and ‘need.’ As we can see, they are very important to understanding this query. 

Example 2

Another a screenshot of Google explaining the search algorithm, BERT. The search is 'do estheticians stand a lot at work.'

In this example, Google couldn’t tell if ‘stand’ was a verb or noun.

Example 3

A screenshot of Google explaining BERT. In this screenshot the search query is 'parking on a hill with no curb.'

In the past, Google would put a lot of emphasis on the word ‘curb’ and not on the word ‘no.’

You can read more about Google’s explanation of BERT here.

Marketers don’t have to do anything to “respond” or “prepare” for BERT. BERT was created to neither penalize sites or benefit them.

If your website does take a  hit from BERT, it’s most likely because you were getting traffic from queries that shouldn’t have been giving you traffic because they weren’t the best fit for the query.

It should be noted RankBrain is not dead. RankBrain was Google’s first artificial intelligence method for understanding queries. It came about in 2015. It looks at both queries and the content on the web pages in Google’s index to better understand what the meanings of the words are. BERT has not replaced RankBrain. It is an additional method for understanding content and queries. But when Google thinks a query can be better understood with the help of BERT, Google will use that. A single query can use multiple methods, including BERT, for understanding that query.

At Sperling Interactive, we pride ourselves on being experts at search engine marketing. You can learn more about to optimize your website over on our search engine marketing page.

Contact Us

What can Sperling Interactive help you with?