Recently, Google introduced BERT, a new update to its search algorithm. According to Google, this update will affect 10% of all English language queries in the United States. An expansion of this update will take place gradually across different countries, although Google has not decided upon a fixed timeline. Regarded as Google’s most important update to its search algorithm in five years, BERT has the potential to impact SEO in a significant way.
What Is BERT SEO Update?
BERT stands for Bidirectional Encoder Representations from Transformers. It’s introduced as a neural network-based technique for Natural Language Processing. In layman’s terms, Google implemented BERT in order to better understand more conversational, longer-form queries. BERT processes words in a search query in relation to each other, allowing for a more accurate list of results that more closely relate to the question asked.
In a blog post made by Pandu Nayak, the Vice President of Search at Google, the effects of BERT on search results were demonstrated through a few examples:
In this example, the search query refers to the physical demands of an esthetician. Previously, Google would interpret ‘stand’ as a key word, matching it to whichever result had ‘stand’ in its content – in this case ‘stand-alone esthetics’ schools’. However, BERT analyzes the contextual elements of the question, producing a result that matches meaning, not just key words.
Another instance shows the importance of words such as ‘to’, ‘from’, and so on. The question shows a Brazilian traveler asking about a visa to enter the US. Before, Google would direct the question to a result that included key words but had no relation to the question. With BERT, the question is immediately directed to the website for the US Embassy in Brazil.
For the everyday Google user, this results in a more convenient, accurate and streamlined search experience. For SEO, however, this produces significant changes in what shows up on the first page of the results list.
How It Affects SEO
Although the BERT update appears to be a relatively minor adjustment to the search algorithm, it’s indicative of the direction that Google is taking in terms of implementing a more contextual approach to the search engine. As such, it’s important to think about what a step towards Natural Language Processing means for SEO in the long-term.
Due to its focus on word context, BERT mainly affects long-tail search queries, as shown in the examples above. Because tracking tools focus more on short-tail queries, there hasn’t been much of a visible change in Google Search results. However, according to Google, the update will impact 1 out of 10 search queries, affecting a site’s visibility and traffic rankings. Given its relative newness, it’s difficult to see how big this impact will be. However, this update gives an idea of the direction in which Google will be heading in the years to come.
Previously, when Google’s search algorithm relied mainly on keywords, SEOs encouraged long-form content to maximise keyword usage and optimise visibility. However, with BERT, keyword density doesn’t do much to help optimisation. Instead, it requires high-quality content that is specific to its topic, in order to effectively produce results that attend to its user’s needs.
How to Optimize Your Content to Accommodate BERT
Since the release of BERT, web content creators have been wondering how to effectively optimise visibility and traffic to accomodate it. According to Danny Sullivan of Google’s Search division, there seems to be no way for SEOs to optimise their content to adjust to Google’s latest update to the algorithm, other than to focus on engaging, well-focused and high-quality content. Because BERT is an update that mimics and responds to the conversational nature of humans, the best way to optimize for it would be to produce high-quality content that people want to see, rather than keyword-laden articles.
For more tips and advice on how to do this, or to explore any other SEO-related topic, check out SEO Company Melbourne – the experts in all things SEO