BERT stands for bidirectional encoder representations. BERT is a new technology from Google which will help understand search queries.

How will BERT help?

BERT technology takes into account words in search queries in relation to other words. This helps to better understand the meaning of the search query and generates more appropriate results.

Will this revolutionise the online landscape?

If BERT works according to plan search results pages could be transformed. In the eyes of a user the ease of online searching could be increased tenfold by BERT. Information which previously would have needed to be dug around for may very well be presented in the first few search results. This will drastically improve user experience.

BERT may also be innovative in terms of search questions. Everyone has been in a position where they have typed a question into Google hoping for an immediate and succinct answer, plenty will have not have achieved this. It often appears to be the case that certain keywords are picked out from search queries and these are what takes precedent in results. With BERT the whole search query will be taken into account and will (hopefully) mean our questions will be answered.

BERT could have a knock on affect to PPC. BERT could help filter ‘duff’ traffic from sites as the users who are looking for information rather than purchasing products will be shown pages with this information rather than ads selling a product or service. This means ad spend could be more effective, by reducing the amount of clicks from less relevant traffic.

BERT Computer Search

What are potential problems with BERT?

BERT could have an effect on bounce rates. If the pages in search results are highly relevant and contain all the information the user is looking for then they may not interact with the site in any other way before leaving. This could increase the bounce rates on sites and potentially be seen as problems with the site or the sites content. This would need to be taken into account if BERT comes to the masses.