Home Technology What was the Google BERT Update?

What was the Google BERT Update?


What was the Google BERT Update?: Most people use Google to search for answers to their questions without ever thinking about how the Google search engine interprets and successfully serves up relevant search results (SERPs) consistently.

The search engine we use to discover answers every day is made up of very complex technology. To keep the search engine running efficiently and effectively returning the best results, Google has to make updates to their algorithm regularly.

What was the Google BERT Update?

Google BERT

Visit this article for a full list of Google algorithm updates.

Google continues to refine their search technology to provide the best results each and every time you search. One of the key issues with search is giving the search engine the tools to understand your query, even if you don’t know the right words to ask or don’t use the correct spelling.

Google’s BERT Update

In order for Google to understand your question, it requires an understanding of language. That is a lot more complex than using a dictionary. If you’ve ever tried to learn another language with just a dictionary, you know that you won’t find culturalisms, slang, and regional expressions.

These parts of language, especially spoken language, require context and a deeper understanding of the culture in which the language is spoken. And even if you speak the same language, English, for instance. It is not spoken in the same way in Texas as it is in New York.

Those differences grow when you compare American English with Australian or British English. In each culture, the language changes to fit local usage.

This is where Google’s BERT Update comes in. In 2018, Google engineers unveiled their new technique for natural language processing of search queries. This was named the Bidirectional Encoder Representations from Transformers, or BERT

How Does Google’s BERT Update Work?

The development in how technology understands written and spoken language has advanced greatly with the introduction of the Transformer, which is a new kind of neural network architecture that is focused on a mechanism of self-attention, and should be better equipped to understand language. 

The key is that instead of translating the meaning of each word in a sentence one-by-one, the Transformer looks at the relationship of each word in relation to the other words in the sentence.

BERT models can therefore consider the full context of a word by looking at the words that come before and after it—particularly useful for understanding the intent behind search queries.

Google engineers discovered that improving the software wasn’t enough to give the BERT update what it needed to work properly. They also had to upgrade the hardware. To this end, Google started using Cloud TPUs to process the search queries. These were the newest cloud-based supercomputers built for artificial intelligence (AI) learning.

Starting with the English search used on Google.com in the U.S., BERT helped the search engine better understand one in 10 search queries. And Google engineers are using what they learn from the update in English, and applying it to other languages.

Languages change over time, so the challenge of interpreting queries will continue for written and voice search.

BERT and On-Page SEO

Since BERT focuses on language, words, and context as it processes search queries, that means it will do the same for your online content. BERT is using machine-learning and sophisticated modeling to match SERPs with queries. BERT understands words based on the context they are found in. Many words have more than one meaning. Words like “rose” or “bass.” BERT relies on the context to interpret those meanings.

Optimizing for BERT means that your content needs to use words that consistently provide clear meanings. Sentences should be well-constructed to provide BERT with a clear understanding of the meaning and context of your words. 

But, you no longer need to repeat the same words or phrases over and over to ensure that the search engine picks up your main topic. You can produce high-quality content that offers your visitors a deep, rich experience or interaction. If your content is focused and well-organized, you will be able to give BERT the context needed to match query and result. 

For this type of in-depth work on your site, you would be best served to establish a considerable about of time to handle this, hiring an in-house SEO specialist, or working with a professional SEO agency

Organizing Content for BERT

Many of the optimization techniques you can use for BERT will improve your content for human visitors. Techniques include:

  • Clear title and subtitles
  • Long-tail phrases
  • Concise sentence and paragraph structure
  • Content that offers a deeper meaning of your topic
  • Relevant content

Review your content to consider whether it can be improved with clearer, structured language.

  • Does your content flow properly from one topic to another?
  • Is the meaning clear to someone visiting for the first time?
  • Do my title and subtitles help clarify the topic?
  • Does your content offer context to explain words with different meanings?

Featured Snippets

BERT also improved the featured snippets results on Google. A featured snippet is what you’ll find at the top of many SERPs that gives a quick answer to your query without clicking to a content page. These often come from highly-ranked content, but not always.

Featured snippets may offer bullet points, a quick answer, or questions with an attached answer that you can find all without leaving the SERP. The most common types are a paragraph, list or table. They often have an image or video accompanying the text.

The best way to have your page included in a featured snippet is to create content the way Google likes it. And that goes back to creating relevant, high-quality content for your visitors.

Content Marketing Strategy for BERT

  • Top-of-the-funnel keywords – These keywords are usually informational, therefore most impacted by BERT.
  • High-quality content over long wordy content – High-quality content will help in many ways, including content strategy for BERT.
  • Keyword density – BERT reduces the need for a lot of keywords. BERT understands keywords through context.
  • Long-tail keywords – More than ever, long-tail keywords are important for BERT to hone in on the most relevant content.
Previous articleSpaceX launching NASA DART spacecraft to crash into an asteroid
Next articleThe U.S. Asks Court to Unblock Biden Vaccine Mandate for Big Companies


Please enter your comment!
Please enter your name here