5 Best Practices for Google’s BERT Update

Google’s BERT update enhances search results with advanced natural language processing (NLP). This new algorithm understands the conversational tone of text and concepts.

Learn how to optimize your content for the BERT Google search algorithm.

What Is the Google BERT Update? 

There’s been a lot of discussion about the new Google Bidirectional Encoder Representations from Transformers (BERT) technology and what this will mean for content on the web. BERT is a new technique that has been open-sourced and created for advanced NLP.

The BERT update moves away from monochromatic keyword matching and towards understanding concepts, conversational tone and answering questions more humanly – pretty smart right?

The newly introduced update is paving the way in Google search updates, streamlining the user to information pathway. Google claims that this new update will affect 10% of all searches.

This is a big statistic in the world of search results. Therefore it’s important to understand what BERT even means, how this might affect your content and what you can do to optimize it.

The BERT algorithm is based on the premise that the definition or purpose of a word can change dramatically depending on the context it is in. This is mainly seen in the words preceding and succeeding the target or main subject. The context is very important to understand as it can alter the fundamental meaning of the search.

First, let’s look at it at a basic processing level. If the BERT algorithm were to process the phrase ‘the cat eats the bird’, it could identify the cat as the main subject the verb is to eat and the bird is the object.

It would be able to do this from previous learnings. It can then apply this to any other related situation – the BERT algorithm will continue to advance the more searches it processes.

What Does BERT Mean?

Before we understand how to optimize your content, lets breakdown the meaning of BERT. BERT is an acronym for Bidirectional Encoder Representations from Transformers.

Confused yet? The acronym isn’t as scary as it sounds. Let’s take a closer look.


The bidirectional component refers to BERT using bidirectional language modelling to scan the text. This is probably the most revolutionary aspect of BERT technology.

The BERT update has been trained to see the whole sentence on either side of the target word. This means it’s able to view all the words at once, no matter which side of the main subject. This revelation allows the model to deeply understand it’s environment creating a more accurate search with a better understanding of the context.

This is different from previous models that use a unidirectional language method, scanning text from left to right or right to left. This previous method meant the context couldn’t be taken into account as well.

For example, the word ‘nail’ would contextually be represented the same in ‘I like to get my nails filed at the salon’ and ‘I like my nails filed at the workshop’. The first sentence refers to nails as fingernails and the second as building nails.

In the unidirectional model, the sentence ‘I like my nails filed at the salon’ would understand the keyword ‘nails’ based on the preceding words. This can obviously prove problematic. The essential context ‘at the salon’ is not recognized – the determining factor between fingernails and building nails.

However, the BERT can process preceding and succeeding information. Therefore the whole phrase ‘I like my nails filed at the salon’ would be processed meaning it could identify the double meaning of nails.

ER – Encoder Representations 

Put simply, Encoder Representations (ER) mean the encoded information being received is decoded for processing ease.

T – Transformers

The use of transformers allows BERT to understand natural language and a more conversational tone in searches. This is done by having a large focus on pronouns which play a huge part in the understanding of the conversation. When pronouns are understood and processed the BERT update can make sense of human conversation.

Google explains “for longer, more conversational queries, or searches where prepositions like “for” and “to” matter a lot to the meaning, Search will be able to understand the context of the words in your query. You can search in a way that feels natural for you. ”

Google’s blog provided a great example of why these prepositions matter so much in the search. They used the example of someone searching “2019 brazil traveler to the USA needs a visa”. The word “to” relationship to the context is crucial. It’s about someone coming from Brazil TO the United States, not the opposite.

Google says previous search algorithms would have no chance of identifying this, but the story is different with BERT. The search can now realize “to” is incredibly important and process the information around this which gives the user the result they are looking for.

The BERT algorithm was rolled out across the English language in October 2019, they are now in the process of running the same implementation across 70+ languages.

The system can take its learnings from the English language and apply them bilingually, which means other languages are only going to receive bigger and better results.

We are still a while away from computers outsmarting the humans scenario aka Terminator’s SkyNet, but search algorithms are well on their way to thinking the way we do.

What Does the BERT Update Mean for Your Content? 

It means the conversational tone can be picked up better. In short, if you have the content readers are looking for they can find it easier. The more your content reflects the natural way we communicate between one and other, the better it’s going to perform.

As search engine technology gets smarter like BERT, the importance of being specific and creating quality content increases. It’s important to stay specific, create good quality content and not excessively shove your blog posts full of keywords just for the sake of it.

5 Ways You Can Optimize for the New Google BERT Update 

1) Quality over Quantity

Old school SEO strategy tells you more, more, more. More content means more keywords. As Google searches become more advanced, this method no longer proves to be as valuable.

In the past few years, experts have advised us to move away from this strategy and focus on the quality and uniqueness of your content. The new BERT update reinforces this even further. With the new search algorithm being able to hone in on conversational tone and specificity, your content needs to include exactly what the user is looking for.

An example being if someone searches for ‘what are the best hikes in America’ your content shouldn’t include other tourist activities in the same post. Hyperlinks to other tourist activities are great, but the content should primarily focus on those great hikes in America.

The new BERT update means there’s a high chance your content will be ruled out because it doesn’t relate enough to the users’ search – the content is diagnosed irrelevant if the reader hasn’t put it in their search query.

Density is no longer the focus.

2) Specify

Know your readers, know your topic. With BERT being able to identify the search query better you must be able to answer their question rather than stuffing your content full of keywords.

While there’s not an awful lot that can be done here, make sure that you include conversational and relevant language in your posts, this will make it easier for readers to find you.

Using analytics and marketing strategies can help identify your target audience. This can identify the right language or tone that attracts your readers and provides them with the correct information.

As we now know that trying to target EVERYBODY may no longer work. Instead, use the tone of voice which your target readers will identify with. Focus on the flow of writing, make your writing readable, and really try to answer the topic you set out to answer.

3) Topic Clusters and Pillar Pages

A good way to put these previous optimization steps into action is by focusing on the topic rather than the single keyword. Using topic clusters and pillar pages draws the focus away from single keywords and more towards useful content.

Topic clusters support the idea of choosing one broad useful topic or pillar content and hyperlinking out to other related topics. This develops the owner’s authority on the subject and is more likely to show up on the new google searches.

The pillar content is usually a long-form blog post which covers all aspects of the general topic but leaves room for other blog posts to elaborate. This establishes your blog as an expert in the industry as readers can stay on your page and click through to the information they need – rather than hopping back onto a Google search.

This structure is already a great way to increase your SEO and is going to become even more beneficial as BERT update further develops its magic.

4) Focus on a Balance of Short-Tail and Long-Tail Keywords

Short-tail keywords are focus words equal to three words or less. These keywords are usually searched for in general searches where the individual may not know what they are looking for. Inevitably these short-tail keywords are highly competitive in search rankings.

Long-tail keywords are three words or more. Long-tail keywords cover content that is less often searched for, but this means searches are less competitive and often led to conversions. You might think that the new BERT update would encourage the use of long-tail keywords, due to its conversational nature, but industry experts assure to focus on both.

By focusing on short-tail keywords and providing the user with relevant content you will naturally be hitting the target for long-tail keyword target – especially with the BERT update in action which can sift through long-tail keywords easier.

A commonly used tool for keyword planning is Google AdWords Keyword Planner. This tool lets you search for keyword suggestions as well as vital information such as PPC (pay per click) advertising, competitive and global/local searches.

By carefully selecting a range of well-researched keywords you save yourself time and increase your chances of showing up in Google searches. And just to engrain in your head, don’t go crazy on the keywords, it comes across robotic and can often do more harm than good to your content.

5) Utilize Your Copywriting Skills 

As Google’s search algorithms become more and more human-like it’s important to make sure your copywriting reflects this. Say goodbye to the days of monochromatic and robotic writing, BERT wants you to write well. This doesn’t mean to say make your content as wordy and long as possible but instead write clearly, concisely and well.

If you craft compelling and useful copy, your reader will be more likely to stay engaged and stay on your page. This should be the ultimate goal as otherwise, you risk the reader returning to Google search and landing on a competitors page. Combining good copy with tactics like pillar pages you will increase your chances the reader can find everything they need on one page.

The Bottom Line

Optimizing your content for the new BERT update is valuable. People are only going to click on the relevant, high-ranking content for their search. Continue to use good existing SEO methods and pay attention to the ones in this post. Quality and specific content, topic clusters and pillar pages, good keywords and good writing will all increase your chances rank high in a Google search.

The BERT update encourages the quality of content making the user’s life easier and adding to a more streamlined search. When producing web content aim to put the reader at the forefront of your mind rather than ranking highly. This will naturally mean you are optimizing your content for the BERT update.

At the end of the day, good content will always be the reigning king.

Don’t forget to share this article!

Share on facebook
Share on twitter
Share on linkedin

Related articles