BERT stands for Bidirectional Encoder Representations from Transformers. It works by pre-training this NLP technique to better understand the context of a search query through a bidirectional neural network architecture.
Sound complicated? Well, it is. That’s why it’s taken this long for Google’s top engineers to make it work. As marketers, we’re more concerned with how it affects us and what impact it will have on our website.
1 in 10 search queries are said to be affected by BERT. These are predominantly longer, more conversational search queries, as highlighted by Google. Not the sort of high traffic keywords that are usually worth rank tracking.
Should I optimise for long tail, conversational search queries?
In short, no. BERT will make search results more relevant to users, so if you’re providing quality content then you’re already doing all you can to optimise for BERT. You may see better engagement metrics as search intent will be closely matched to the content you’re providing.
What can I do to optimise for BERT?
The official answer from Google is that there’s nothing you can do to optimise for BERT, reiterating that your website needs to serve users to succeed. As Google’s algorithm gets better at understanding users, the only way to “optimise” your website is to ensure you’re providing quality, unique and engaging content.
What does this mean for the future of SEO?
This isn’t the first time that Google has given little guidance after releasing an algorithm update. Since March 2018 Google has been releasing updates without actionable recommendations, advising only that users create quality content. It’s unlikely that Google will give much more away with any future updates, but it’s becoming increasingly important to think more like a user than a search engine when it comes to SEO.
If you need support understanding Google’s BERT update or want advice on adapting to the changing search marketing landscape, then contact Torpedo.