In a recent announcement, Google introduced the latest update to their search engine, one they have typified as the single largest step forward in search technology for the last five years.
Okay, so it sounds less impressive that it should do. Nonetheless, BERT is something that all website owners and designers need to be aware of, especially with its connotations for the future.
What is BERT?
BERT stands for Bidirectional Encoder Representations from Transformers. It is a deep learning computer algorithm that uses natural language processing as part of its calculations.
What this means in layman’s terms is that the BERT algorithm allows a computer to read how a human reads. Not only will it be able to recognise the words in a sentence and know what they mean through definition, it will also be able to understand the sentence by reading each word in context of those surrounding it.
How does BERT work?
There are reams of technical data available online discussing the complex ins-and-outs of BERT and we don’t intend to rehash them here. From a practical point of view, BERT will be most effective on longer, more complex search strings.
By way of an example: consider the search query “can you get medicine for someone pharmacy”.
Under the old keyword method, the search results would focus on getting a prescription filled at a pharmacy. Using the BERT algorithm, Google understands that the words “for someone” changes the desired result. It allows Google to infer – using context clues – that the user wants to know if they can pick up medicine from a pharmacy on behalf of someone else and produce more relevant results.
How will BERT affect my site content?
Short answer: it won’t. At least, not at first. Estimates from Google suggest that context-based searches account for less than 10% of all searches. Most people still pop in a handful of keywords and let the current system do its thing. As context-sensitive searching becomes more commonplace and the BERT algorithm becomes more sophisticated, that figure could increase.
The best advice is to create quality content, as this will attract BERT searches on its own. Context-driven searches mean you will have to demonstrate your expertise to attract interest, rather than relying on keywords.