How Does The BERT Update Affect Your Website?

You may have heard about the latest machine-learning related update to Google’s algorithm. It’s called BERT, and it has a lot of website owners nervous.

But there’s no reason for any law firm SEO expert to fear this new BERT update as it only affects about 10% of searches worldwide.

This guide will answer any questions you might have about Google’s BERT update.


What is the Google BERT algorithm?

Google’s BERT algorithm was rolled out on October 25, 2019. The engineers at Google have hailed it as the greatest advance in their search algorithm in the last five years.

Others inside Google claim it is their biggest update since RankBrian, and ’one of the biggest leaps forward in the history of search.’

BERT is a ‘deep-learning’ algorithm that relies on something called natural language processing, or (NLP).

The algorithm analyzes natural language processes like questions and answers, and entity recognition.

In plain English, Google has adjusted the way its algorithm uses pattern recognition to better understand how human beings communicate so that it can deliver the most relevant search results for its users.


What does BERT acronym stand for?

BERT stands for ‘Bi-directional Encoder Representations from Transformers.’


What does the BERT algorithm do?

The first thing to understand about BERT is that website owners, and SEO professionals should all take a deep breath. Unlike Panda and Penguin, the idea behind BERT is not to penalize websites.

The whole concept of incorporating BERT into the algorithm is simply to help Google better understand the intent behind more complex searches.

BERT is not focused on the content of your website.

Rather it is focused entirely on understanding search queries.

Specifically, BERT is designed to help the algorithm understand the meaning behind a search query by paying attention to the order of the words in the query, and better understanding the context of the query.

BERT pays attention to ‘stop’ words in search queries. Prepositions like ‘to’ and ‘from’ are ‘stop’ words.

Rather than having the algorithm discount the importance of stop words in a search query, BERT pays close attention to those prepositions, and exactly where they appear in a search query.

Here’s an example…

Let’s assume you are from Brazil.

Then let’s say that you wanted to travel to the United States.

So you type in a search on Google something like this:

‘2019 Brazil traveler to usa need visa’

Before BERT was introduced, the first results was how U.S. citizens could travel to Brazil without needing a visa.

Close, but yet so far away from the actual intent of the search query.

The searcher wasn’t asking about how people in the U.S. could travel to Brazil. They wanted to know about traveling from Brazil to the U.S.

Now, after the BERT update, the first result in the SERP is much more relevant…

screenshot of a search query before and after BERT about travel from brazil to the U.S.


Let’s look at another example of how BERT has improved the relevance of the top results to certain types of queries.

Let’s assume your significant had a desperately needed prescription called in by a doctor, but was too sick to drive to the pharmacy to pickup the medication.

You, being the amazing spouse you are, volunteer to go pick it up.

But before you go, you check google just in case to ask…

‘can you get medicine for someone pharmacy’

Look at the top results Before and After the BERT update.


screenshoy of search results before and after BERT about pharmacy question


The algorithm is now better able to understand the subtle nuance in that search query – which implies that the searcher wants to know if they can pick up a prescription for someone else at the pharmacy.