Guess RankBrain isn’t the only best invention since sliced bread, introducing BERT designed to impact 1 in 10 of all Google search queries.

Who doesn’t look out for authentic and relevant searches? It is quite obvious that no matter how simple the search may be, we always want the most accurate results to show up to address our concerns. Speaking of which, we all know about RankBrain when Google introduced the search feature over five years ago to enhance the quality of the result. Fast forward to 2020; Google just announced that they’re bringing a huge amendment to their search system, known as BERT. It will bring relatively precise options to choose from.

What could be the reason behind this change?

Search engines are striving consistently to provide not only the most relevant but also the most reliable search results that they possibly can. Reliability and relevance are a matter of concern for Google, particularly and can be seen in the ‘your money your life sector’ (YMYL). With top stories like the global pandemic prevailing search results and the US elections on the brink of happening, it is necessary for a trusted search engine like Google to keep up with updates and provide correct information.

What is BERT, and how is it used?

Bert is an acronym for Bidirectional Encoder Representation from Transformers. It relates to neural networking and artificial intelligence techniques for pre-training of NLP, natural language processing. With the help of BERT, Google claims the new feature allows a precise understanding of words and tones used in context, to address queries with more reliable sources and relevant results. BERT is also used for snippet features.

What difference did it make?

Pandu Nayak, vice president of Google search and a Google fellow, in a recent interview, stated that they’ve been improving their systems over the past few years. It is to make sure they provide authoritative information. He explained further that as news trends develop, it’s not likely that the most recent news on the internet is the most trustworthy. On the other hand, users’ need for reliable information can crop quicker than facts can materialize.

Apart from this, he stated how the announcement mentioned the changes that Google accomplished to make in recent years. Rather than emphasizing on the launch of a new feature solely, some of the tweaks Google made since last year include:

Detection of breaking news in merely a few minutes relative to the wait of 40+ minutes.

Close working with Wikipedia to look out for and eliminate vandalism, something Google might make use of in knowledge panels.
In 2020 alone over 4 billion times of fact check labels were shown.
In an attempt to better evaluate fact check against stories, the new feature is being used in full coverage news stories too.

Unfortunately, you cannot use optimization for BERT. Nonetheless, with such advancements and recent features, Google will do a better job at understanding the natural use of language and tone used, to match for more correct and accurate results.