Google’s BERT Algorithm Update

  • February 05, 2020
  • Bradley Taylor

It has always been a goal of Google to provide the searcher with real and accurate content. The challenge has been to scan everyone of the internet’s 1.74 billion websites and find the most helpful content. They use spiders and algorithms to perform this task. Every year they update their algorithms in hopes that it provides better search results for their users. They usually attach a nickname to these updates.

The update for 2020 is called Bidirectional Encoder Representations from Transformers or BERT. One of the goals of this new roll-out is to find human-written content that is helpful and informative. The reason that “human-written” is important is that there are a plethora of websites that generate automatic text, using bots. These content pages may include keywords and links but they are essentially gibberish to a real, actual human. You may have run across pages like this in the past.

Another stated goal of the implementation of BERT is to help Google’s AI understand the way humans talk. When we search for something, we often include words that are relative to the way we speak to one another. In the past, Google would just focus on the keywords of a search. But now, it is learning how words are used in the structure of a sentence in order to convey what the person is actually searching.

Google’s BERT actually started rolling out at the end of 2019 but for all intents and purposes it is a 2020 update. Search Engine Optimization managers have already noticed some difference but nothing drastic.

Google claims that you cannot optimize for BERT. Instead they say to just continue writing helpful, informative content and post it regularly. This is one of the many tasks that Olympus Web performs for our clients.