Skip to content

Everything You Need to Know about BERT

By: Annika Peveril

Everything You Need to Know about BERT

Have you always heard about BERT, but never knew exactly what was meant by it? Well, it’s the most important algorithm update Google has made in the past five years. So, let’s get you up to date as well.

In this article you will learn:

  • What is BERT
  • Why is everyone talking about BERT?
  • How Bert works
  • Why you should care about BERT
  • How to optimize your content for BERT

What is BERT?

The Bidirectional Encoder Representations from Transformers (BERT) is the all-in-one combo that solves majority of the Natural Language Processing (NLP) tasks. It was developed by Google and is being leveraged to better understand user search queries, which should mean more relevant results. 

For example:

The word “to” can have different meanings, which is easier for humans to understand but harder for search engines to interpret the meaning. Now with BERT, the search engines can distinguish between them. Let’s take the example that Google has provided: “2019 brazil traveller to usa need a visa”. 

Image source: Google

Before, algorithms would result about U.S. citizens traveling to Brazil. Now with BERT, it understands between such nuances and facilitates more relevant results. Google is looking at the searcher’s intend and getting a better understanding of what that is. 

BERT arrived in Google’s search system the week of October 21, 2019 for English-language queries. Over time, Google will expand the algorithm to all languages.

Why is everyone talking about BERT?

It’s unique because it is the first deeply bidirectional, unsupervised language representation, and is pre-trained using only a plain text corpus. It is open-sourced, allowing anyone to use the software for any purpose. Therefore, anyone with machine learning knowledge can easily build an NLP model without the need for sourcing massive datasets for training the model. This saves a lot of time, energy, knowledge, and resources. 

It is estimated that 50% of all searches will be voice searches by 2020. Thus, this progression is logical as Google needed a way to better understand more conversational search queries. 

How BERT works

Let’s begin by defining what natural learning processing (NLP) is. NLP aims to enable computers to understand the way humans communicate. It relates because BERT is an advancement in NLP through bidirectional training. It’s bidirectional because BERT can understand both previous and next context. BERT is the first to successfully pre-train bidirectional in a deep neural network.

BERT allows the language model to understand the entire word context based on the set of words in a sentence. Before, it would just learn the context based on the immediate word before or after it.

Why should I care about BERT?

BERT will enhance your search on Google! Google has stated that it will impact one in ten searches, as it will work in more complex queries. Most importantly, BERT is aimed at content overall.

We’re changing slowly from conducting “keyword analysis” to “search analysis” as Google is becoming smarter and distinguishing the context of a search. If you create content that is geared toward a specific idea or topic you can rank, it doesn’t matter what keywords people type in. As long as the intent of their search will lead them to your content. 

Check out this video to learn more about why you should care!

Can I optimize my content for BERT?

This is a frequently asked question since BERT came out just a few months ago. Honestly, you can’t optimize for BERT, but you can optimize your website’s content for visitors.  

There a couple of things that you can be aware of that helps BERT make sense of your content, such as:

  • Write grammatically correct English 
  • The flow from one sentence to another is important to get a good ranking
  • Make sure your content answers questions that people are searching for
  • Keep your target audience in mind, as Google is trying to better serve searchers by introducing BERT

Danny Sullivan, an employee at Google, focuses on educating the public about search and is very active on his Twitter account. 

Google insiders, like Sullivan, often provide the same piece of advice to write high quality content garnered at their users. Now with the BERT update, this advice should be taken more seriously as Google is taking steps to better understand humans. 

Interested in other ways to optimize your content for search engines? Check out my article, “SEO 101: A Guide for Beginners”.

What should I take out of this article?

  1. BERT helps Google to better understand human language
  2. BERT looks at contextual nuance
  3. You can’t optimize for BERT, however you can focus on writing content for users

Thanks for reading! Tweet us if you have further questions about BERT. Interested in checking out my other articles? Click here