There has been a lot of discussion, of late, about Google’s new search engine algorithm, BERT. Some have heralded it as the most important update to Google’s search engine since … well, ever. Others seem to think that Google search is broken forever.
Before we get into what that is and why people have been getting in a flap about it: keep calm! There’s absolutely nothing to be worried about.
BERT is software that helps computers process requests made in ‘natural’ language. This means that they are making strides in having software that understands how we talk and write, without making us use the limited, specific language of computer programming.
BERT stands for Bidirectional Encoder Representations from Transformers. Of course it does. It’s part and parcel of the research going into ‘NLP’ (that’s Natural Language Processing to you and me) and ‘ML’ (Machine Learning). All of this has become more important since the rise of computers that respond to speech.
Many years ago, Google admitted that it was trying to create a computer like the omnipresent, omniscient ‘Computer’ they have in Star Trek. You know the one: When Capt. Picard walks onto the bridge of the Enterprise he looks at the view screen and says “Computer? Tell me everything I need to know about that planet”. And a woman’s voice from the ceiling does exactly that. As simple and as satisfying a notion as this is, it requires a vast amount of behind the scenes work.
First, the computer needs to have access to all the info about that planet (via search engines). Secondly, it needs to be listening, ready for someone to give it an instruction or ask it a question (like your Alexa or Siri virtual assistant). Then it needs to understand spoken language well enough that it can act upon new questions that it hasn’t been asked before, however they’re phrased. Finally, it needs to know you well enough that it can answer the question in a way that will be most relevant to you.
The majority of people still access search engines through a keyboard, so BERT helps Google with these text searches too.
The new algorithm helps Google to understand the words you use. That word ‘understand’ is important. Previously, search engines have looked at the words you have used in your search and presented results that match those words. It didn’t understand your search; it just offered you results that matched your search. That’s matching, not understanding.
In order to use a search engine properly, you have to format your question accurately, maybe using Boolean techniques. Essentially, you have to communicate with the computer in its preferred language, rather than yours.
As Pandu Nayak (Google VP for Search) said in his blog introducing BERT: “When people like you or I come to Search, we aren’t always quite sure about the best way to formulate a query. We might not know the right words to use, or how to spell something, because often times, we come to Search looking to learn – we don’t necessarily have the knowledge to begin with”.
Hasn’t that been the age-old frustration with dictionaries… How can I look up the way to spell a word if I don’t know how to spell the word?
Clearly, search engines are much more sophisticated than this but, the point is, they didn’t understand your motives for the search, nor did they take into account the different meanings that words and phrases can have, to different people, in different combinations. They certainly don’t take into account the searcher’s lack of prior knowledge. If you ask a human being a question, the answer you will get will have more empathy, intuition and nuance than a computer’s answer. They’ll understand what you mean, even if it’s not what you said.
BERT is part of the ongoing work to make it possible for computers to do that, to understand us much more completely, so they can help us more effectively.
The quick answer is: The same way you always optimise.
Write great content that is information rich and works well for organic users (or ‘people’, as we used to call them). Therefore, Google’s John Mueller suggested, on Twitter: “Our recommendation is essentially to write naturally”.
So, keywords are still important and it’s still important to not overuse them. It’s even more important to use them the way that searchers will use them. So, do your research into the search phrases, as well as search terms, that are likely to bring people to your site.
Search engines basically respond to three types of queries:
This final category is the one where BERT can make a big difference. In order to optimise your content to take the best advantage of this, you need to know what questions searchers will ask.
If you can make the most oft-asked question your H1 Heading, then you have greatly increased the likelihood that searchers will find you.
But, it must be said, if you create content that is SEO-friendly, you probably do all that already.
So, keep calm and carry on.