In October 2019, Google launched one of the biggest algorithm updates for its search engine, dubbed the Google BERT.
The BERT algorithm is expected to affect around 10% of all search queries, so if you already have a website and published content, it’s very possible that your content is affected in one way or another especially in SERP ranking or organic traffic.
In this guide, we will discuss all you need to know about Google BERT and how it will affect your SEO and content marketing. By the end of this article, you’ll know how you should adjust for the new algorithm update and how to capitalize on the new opportunities.
What Actually is Google Bert?
Google Bert is a major Google algorithm update, with the latest major algorithm update being the Google Fred, launched in March 2017.
However, many will argue that Bert is a bigger update compared to Fred and Possum (2016), the biggest update since Google released the AI-based Rankbrain algorithm in late 2015.
So, there is a huge chance that Bert will impact your site and its SEO performance.
BERT is actually an acronym that stands for Bidirectional Encoder Representation from Transformers. “Transformer” here refers to the name of a popular machine-learning attention model that is now implemented in language modeling.
Too technical? We know. So, in a nutshell, BERT is an algorithm update so Google can now better understand natural language processing—voice recognition, text-to-speech, and so on—.
Even before the implementation of BERT, in the past half-decade or so, Google is already pretty good in making these smart suggestions related to NLP, for example, you can now type in something like “bitcoin price” and you’ll get a rich result featuring a chart. Ask something like “restaurants near me”, and Google can figure out your location and provide a Google Maps result.
So, how will BERT improve Google search and affect SEO? Let’s use Google’s own explanation and examples.
Before BERT, Google has a difficulty recognizing whether the query is about a Brazillian traveling to the US or the other way around, and can’t understand the importance of the word “to” and “need a VISA”, which will be fairly obvious for human cognition. With BERT, as we can see in the right image, Google can now understand the ‘nuance’ behind this rather complex, long-tail query and provide a relevant result.
In this another example, the problem lies with the word ‘stand’. Before BERT, as we can see in the result on the left, Google can’t properly recognize the word stand and its related context and instead used the matching keywords approach to find “stand-alone” in the result page. After BERT, however, Google can now understand that the context behind the word “stand” should properly reflect the physical demand of the esthetician as a job, and thus display a more relevant result.
Google BERT Is all About Context
As we can see, BERT mainly deals with context, and Google can now better understand search queries—especially long-tail, complex keywords— in its whole context. In the past, Google (which we sometimes forgot that it’s still a machine), only processes queries as a string of words.
Meaning, before BERT and as we can see from the above examples, Google normally interpreted each individual word of a long-tail query and is not very good at understanding the relations between these individual words to each other.
Google BERT Changes in Featured Snippets
Again, this is Google’s official example:
As we can see, Google offers very different results before and after BERT implementation, and Google themselves admitted that queries like this have confused Google’s system in the past, as the pre-BERT algorithm placed too much importance on “curb” and neglected “no”. According to Google, the BERT model has improved featured snippets not only for US English results but also in different languages like Korea and Hindi.
How Google BERT Affect SEO
So, how will Google BERT affect your SEO? This will depend on whether you have implemented SEO best practices and strategies in the past, or not.
Google BERT is aiming solely at one thing: making it easier for users to search more naturally, especially with long-tail keywords. This is aligned with Google’s mission to organize information on the internet and make them relevant and useful for human readers.
So, if you already focus on publishing content for the human audience as the core of your SEO, Google BERT shouldn’t change this fundamental. However, if you still rely on grey-hat and even black-hat SEO strategies—especially in your content approach—, the implementation of Google BERT will mean even fewer opportunities for these grey-hat tactics.
Simply put, Google BERT will help websites that are already rich in information to reach their relevant audience, and won’t help websites that are poorly written.
In general, there are three types of search intents people have when they performed a search:
- Informational: the user is looking for (often over-arching) Information, for example, they might search for something like “how to improve productivity”
- Navigational: the user specifically wants to navigate to a certain website, for example, they have stumbled upon a favorite productivity blog, let’s say, xyzz.com, then their navigational query is “xyzz blog” or something similar.
- Transactional: here the searcher is specifically looking to purchase a specific product, for example, the same xyzz blog is selling an ebook, then the query is something like “xyzz ebook”.
One of the key ‘improvements’ of Google Bert is that Google can now better understand the queries and their context, in relation to the specific search intent, and this will affect SEO in so many different ways. Not to mention that better understanding of search queries can help you in terms of Google ads management.
For instance, there are long-tail keywords that might seem obvious but can be difficult to target due to the limitations of the pre-BERT algorithm, for example, “Top 5 photoshop alternatives available in 2020”.
In the past, we have to include “Photoshop” and “Photoshop alternatives” here and there in the content so Google can correctly understand what your content is about. Quite often, the content writer will include a specific—in-depth—section about Photoshop to properly ‘convince’ Google.
With BERT, however, we can now pack the content with the actual discussion about the alternative solutions and software, and in theory, Google can now properly understand it.
BERT will also mean that we—finally—should stop focusing on keyword density,100%. Yes, a lot of SEO practitioners have understood that SEO is no longer about keyword for the past 3 years or so, but you’ll be surprised to hear that many SEO ‘experts’ and digital marketers are still concerned about keyword density and even keyword stuffing.
The Future of SEO
What actually is SEO in its relations with content marketing and Google’s algorithm?
We can think of SEO as the “bridge” between the two.
As mentioned above, Google’s mission is to provide structured, relevant content for the human audience, and Google’s algorithm is simply Google’s way to understand what’s inside the content and how relevant it is to the searcher’s intent. SEO, on the other hand, is the content marketer’s effort to ‘convince’ Google that the content is actually relevant to the user’s intent—with a series of optimizations—.
So, we have to understand that the more advanced and perfected Google’s algorithm is in understanding the user’s intent and natural language, in theory, the fewer SEO optimizations are necessary, and the more focus we should put on content quality and relevance.
Google BERT and Long-Tail Opportunities
Google BERT algorithm update especially open up opportunities to create highly-specific content to cover a topic or to answer a question.
It’s no secret that in the past, longer content tends to rank higher. However—as mentioned above—, as Google algorithm is getting better at recognizing the intent of content and how it fits a certain context, shorter, more specific content can perform better.
This, however, doesn’t mean that long-form content is dead, far from it.
Instead, it simply means that we’ll have more freedom and versatility in developing our content. What matters is quality.
Long-tail keywords generally have fewer competitions, and with BERT, targeting these keywords will be easier—especially as the BERT implementation is getting perfected in the future—.
The implementation of Google BERT in October 2019 shouldn’t affect your site’s traffic significantly as the Penguin update or RankBrain update, and in most cases wouldn’t result in a penalty.
In fact, if your traffic drops a little after the BERT update, it is probably a good thing. After all, BERT is implemented so that Google can better direct traffic to relevant sites, so, if some of your traffic is ‘adjusted’ after the BERT update, it is most likely irrelevant traffic in the first place that won’t result in a conversion.
On the other hand, Google BERT—as discussed above—, has opened and will open more opportunities to target specific, long-tail queries. These long-tail keywords generally have lower competition and can be very valuable opportunities today and in the future.
Mike Khorev is a marketing consultant and SEO expert who helps SaaS, Software, IT, B2B, enterprise and technology companies generate more leads and sales and grow revenue online. He also offers expert advice on marketing your company the right way through performance-based digital marketing, web design, SEO social media, search engine marketing, and many other online practices.