Google Core Update – Bert: Google’s latest search algorithm to better understand natural language

Google Welcomes BERT to the Search Algorithm

Google has long since been the masters of the search, so much so that ‘to Google something’ has become a synonym for searching on the web even if you don’t use google. This means that whenever they make a significant change to the way their searches work then the world sits up and takes notice. Well, Google is rolling out the most significant change to their algorithm since the introduction of RankBrain almost 5 years ago.

Enter BERT

BERT stands for Bidirectional Encoders Representations from Transformers.

The idea behind this is to get a better understanding of the context and nuances of search queries to bring back more relevant search results. It also updates how snippets are featured on the site. The technology behind this was open-sourced last year and has been in development ever since. It is now in the process of being rolled out across the website on a language by language basis starting with English.

What does BERT Do?

The idea behind BERT is to understand how questions are phrased when entered into Google and make better evaluations on to give answers. The example that Google gives for this is when you search for visa information. You might search for ‘a UK to USA visa’ and google would decide that the word ‘to’ was very common and not relevant to the search and would, therefore, include results about visas from the USA to the UK, this is not what the user was searching for. Therefore BERT will understand that actually the word ‘to’ has a huge impact on the nature of the query and will then provide better results.

How do the Changes to Snippets Work?

The snippets that google provides in response to search queries are also being updated with more natural language processing from BERT. When you were searching for how to do things and needed quick advice then the snippets would give you an overview. However, these weren’t always the most accurate. Again Google said it is about evaluating the importance of certain words. For example, if you searched for ‘computer with no operating system’ previously Google might have determined that the word ‘no’ wasn’t nearly as important as ‘computer’ and ‘operating system’ and therefore would have ignored it. This means that you would get the wrong results. The changes will now help to pick up the importance of words like that.

What Does This Mean in Terms of SEO Strategy?

Google has claimed that you weren’t able to SEO optimise for RankBrain and the changes that are happening with BERT will not change that situation. What is clear is that natural language queries will start working much better. This means that as long as you are crafting content that answers questions then it is much more likely that your content will show up higher in results. It is recommended that you check your current search results and see if your site has been impacted by these changes.