BERT (Bidirectional Encoder Representations from Transformers) is a Google update that was launched in Q4 2019, which uses a neural network-based technique to help Google better understand search queries by users.
Regarding BERT, Google stated “These improvements are oriented around improving language understanding, particularly for more natural language/conversational queries, as BERT is able to help Search better understand the nuance and context of words in Searches and better match those queries with helpful results. Particularly 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 also stated that the BERT algorithm update was “the biggest leap forward in the past five years, and one of the biggest leaps forward in the history of search.”
How large was this update?
One in 10 search queries was affected by this. To put this into context, Google takes around 3.5 billion queries per day, so this means that 350,000,000 queries were affected by the update.
This is the biggest update since Panda (most recent update was in July 2015), which focused on the quality of link value to a site, and Penguin (most recent update was in September 2016), which penalised thin & spammy content. Both updates are now part of Google’s core algorithm update.
What does it mean?
As well as making it easier for Google to understand conversational nuances in search queries, this update also prepares Google’s search engine to better understand more conversational longer tail search queries, such as those used for voice search.
For example, if someone searched for the term “2019 brazil traveller to USA need a visa”, The word “to” and its relationship to the other words in the query is particularly important to understanding the meaning. This search is about a Brazilian travelling to the U.S. and not the other way around.
Previously, Google’s algorithm couldn’t understand the importance of this connection and returned results about U.S. citizens travelling to Brazil. With BERT, Google can now grasp this nuance and understand that the word “to” is significant to the query, therefore providing much more relevant results for this query.
Here is a screenshot below of the searches for this term before and after:
What we can see above is that with the BERT Update (right screen) Google is doing a better job of understanding the query and directing users towards the information they are searching for.
How can you take advantage of this?
Focus on user intent:
There are three main types of search queries; informational, navigational and transactional.
Informational queries are when consumers seek specific answers to a particular question: for example, if a person wants to go on holiday to Italy, they may ask google, “how long is the flight from Melbourne to Rome?
Navigational queries are queries that have the intent of finding a specific website or page. Using the same example, a user may search for “Qantas flights” in order to get to the Qantas site.
Transactional queries are queries that have a direct purchase intent. Following on from our previous examples, a user may search for “cheap Qantas flights from Melbourne to Rome” In order to find a flight to Rome that suits their availability.
Content on the website should be specifically targeted to answering these types of queries, as Google will better understand the intent behind the search, and reward content that is interesting, useful, and unique.
As BERT gets better at understanding conversational queries, it is important to start focusing more on longer tail keywords, as they are more specific than common, high search volume keywords. Whilst these keywords may have a smaller search volume, they can have a much higher conversion rate, as they are more specific and align to user search intent better.
Additionally, content should be marked up with structured data such as rich snippets, breadcrumbs, local listings, knowledge graphs, etc, as structured data helps search engines better understand your content, which complements the BERT Update.
The BERT Algorithm update is a step forward for Google’s understanding of user intent and what the user is searching for. As it continually evolves, companies need to ensure that their content answers user queries as Google will reward this with higher rankings within the SERPs. This understanding of your content will be enhanced by marking up your content with structured data to help increase search visibility as Google and other search engines will better understand your content.