Google rolled out its latest algorithmic update, known as BERT, during the week of October 21, 2019.
This new update is designed to improve Google’s understanding of search queries that are in natural language.
Instead of relying on strings of keywords to formulate our search queries, this update will allow users to employ conversational language. Meaning we’ll be able to use the same language in a search query as we would naturally ask in a question to another person.
Previously, Google interpreted search queries by looking at words individually and in order. The BERT algorithm now places more importance on how words relate to each other.
This refined understanding of language allows Google to interpret the different meanings words can have depending on their context within a sentence. It’s about recognising the user’s intent so it can deliver more accurate and relevant results.
BERT will particularly improve results for searches where prepositions like ‘for’ and ‘to’ can influence a sentence’s meaning.
Google’s goal is to allow users to be able to search in a way that feels natural.
In these examples, Google illustrates how the new understanding provided by BERT facilitated improved comprehension and consequently returned more relevant results.
This before and after sample shows that BERT was better able to understand how the preposition ‘for’ in ‘for someone’ mattered to the intent of the query. Previously, it missed how a simple preposition could affect a sentence’s meaning.
In the past, Google overlooked the relevance of the word ‘no’ to the intent of this query. It placed more importance on the word ‘curb’ but missed how ‘no’ altered the meaning of this question.
Until now, Google’s system of matching keywords failed to understand the word ‘stand’ in this query. The BERT model now has a better grasp on the subtle nuances at play in the English language.
The BERT model will also be employed to improve featured snippets. Featured snippets are the selected search results that appear above Google’s organic results whose purpose is to provide an immediate answer to a search query.
Google has stated that this update will impact on 1 in 10 queries in the U.S. in English. (They plan to expand it to other languages though no timeframe has been provided.)
It has also said there is no specific way to optimise for this update.
The aim of this update is to improve the search results Google delivers.
Companies often create content as part of an SEO strategy for the purpose of improving their own website’s ranking. The result is content created for machines not humans.
Google wants to reward the producers of meaningful content who are answering the questions people are searching for – thereby providing a better user experience.
SEO expert Neil Patel highlights that the content creation opportunity presented by BERT is for creating content around very specific long-tail phrases.
Ultimately, with the BERT algorithm, Google’s intention is to encourage content creators to optimise for humans by creating authentic and worthwhile content.