In the initial press release announcing the advent of Google BERT, Google noted that as many as 15% of the queries they got on a daily basis were on topics that were new to the system. With this in mind, it’s easy to understand while Google keeps iterating on improving its algorithms to try and provide the most effective search results for all users. However, for those on the digital marketing side of the spectrum, the reaction to these changes is a bit different. 

As we can see from previous examples such as the EAT&YMYL algorithm changes, what seems like a simple change behind the scenes can have a radical impact on how your pages rank and alter your entire digital marketing strategy. Here are all the main questions you need answering about BERT and how it will impact future searches.

1. What is the BERT Algorithm? 

BERT was released in the fall of 2019 in the U.S and promises to be the most wide-scale algorithm update since the advent of Rankbrain several years prior. However, a global release date is not yet confirmed. BERT stands for Bidirectional Encoder Representations from Transformers. And if you want to get technical, it was a natural language processing NLP framework before it became a part of Google’s algorithm.

However, to keep things simple, we’ll focus on the relevant context to digital marketing. BERT is an algorithm addition designed to make it easier for Google to process and respond to human speech patterns. In essence, BERT implements a neural network to help pre-train Google to help the computers understand human speech the way a person might.

2. How is Google BERT different?

BERT marks the latest in a series of major updates in the last 10 years (or so) that Google has made to improve overall operations. These include:

• The Panda update, incorporated into Google’s algorithms in 2011. This was largely designed to weed out low-quality search results and put a greater emphasis on effective, optimised content. This created the basis of modern SEO as we know it today.

• The Hummingbird update, released in 2013. This allowed Google to start processing full phrases rather than interpreting individual words, opening the path for long-tail keywords.

• The Rankbrain update, released in 2015. This gave Google the ability to properly process queries that had multiple meanings, like dialogues and slang terms.

BERT’s main purpose, in this context, is to help Google better understand the difference between phrases and sentences in different settings. This is key, because depending on whether a word is used in a phrase or whole sentence, it may have a very different meaning.

3. What is the main purpose of the algorithm? 

The main purpose of the algorithm is to help Google better understand the context in sentences and longer phrases. Those that have been following the world of SEO know there’s a greater emphasis on long-tail keywords and voice search. BERT, in theory, will help Google better handle both. 

As an example, let’s look at the query: “Do barbers stand a lot at work?” Prior to the advent of BERT, Google would try to match together different keywords. As a result, your search may have gotten answers regarding “stand” on its own, which isn’t really relevant. However, now Google would understand the context of “stand” as a physical demand within the sentence.

4. How does BERT affect search queries?

There are a lot of key points to consider here in terms of BERT and how it impacts future queries. Here are a few key examples:

• Pre-training: BERT marks the first example of an NLP-framework trained on pure plain text through unsupervised learning.

• Modeling masked language: In the event that words are masked out in a sentence, BERT now attempts to predict what that word is and factor that into the calculation of search intent.

• Bi-directional context: We know through human reading comprehension that you can’t completely understand the context of a sentence unless you see all the words in a sentence at the same time. In the past, Google struggled with this, so you would see greater weight placed on one word in a sentence. BERT makes sure the whole sentence is taken into account.

• Transformer architecture: Along with looking at every word at the same time, this feature lets Google analyze the meanings of words in relation to each other, cutting down on ambiguity.

5. How does BERT change content marketing?  

In the world of content marketing, BERT doesn’t necessarily change what you should be focusing on, but it makes staying on course more important. Many experts have been emphasizing the importance of writing to humans rather than trying to appeal to search engines, and BERT makes this more important than ever by making Google more human in its understanding.

6. How does this work in practice? 

For one thing, you want to make sure that you’re using the terms, language style, and tone that will resonate with your audience. In addition, you need to make sure that your content does a better job of answering key customer questions. The advent of voice search means that quality is even more important.

7. How does SEO change due to Google BERT?

All the way back to Panda, major Google algorithm changes have been designed to separate the wheat from the chaff. If you want one general piece of advice, make sure to incorporate long-tail keywords in your content. BERT’s ability to better understand the context in these longer terms means that you’ll get more ROI. This doesn’t mean you should go exclusively long-tail, just that you’ll see worthwhile results. Along with this, longer, more detailed content will rank higher, as it allows BERT to better figure out what your context is about and what searches it will apply to.

8. What changes for advertisers because of the BERT update? 

While a lot of the applications in the ad world are more theoretical, many PPC advertisers recommend looking at Dynamic Search ads. These are created automatically based on a search query, so with BERT powering Google’s AI, these may be more effective. One other notable application is that Google will now be able to apply successful PPC campaign knowledge to campaigns in other languages.

9. How does Google BERT affect your website? 

10% of all queries are going to be affected by BERT, so there’s a chance you may see a shift, but it looks like featured snippets and voice search will see the largest changes. This is largely because Google is changing how it determines content is contextually useful. Note that BERT is all about user intent, so you won’t be able to regain ground with technical design changes if your ranking drops.

Major algorithm releases like Google BERT always create a powerful ripple effect in the world of digital marketing. Not only do keyword strategies need to adapt and evolve, but you may need to go back and adjust some of your older content to avoid being penalised for your previous efforts. 

Adapting to a new Google landscape is difficult for any business, but especially for smaller businesses that may not have the time or bandwidth to fully revise their entire website to comply. To avoid losing your rankings, with a full website SEO audit from a digital marketing agency in Melbourne.