While most of us go through our lives feeling unheard and not having been understood by our partners or loved ones, there is someone out there who is trying to get better at understanding each of our queries and serving exact solutions to our problems and questions we pose. Well.. no points for guessing that right, that someone is the Google Search Algorithms which run every time you punch in a search query and are looking for information or, trying to educate yourself on a topic.
Yet again, Google has rolled out a search algorithm that is a step towards decoding and understanding search queries in a more natural and human way.
Birth of BERT
Bidirectional Encoder Representations from Transformers, more popularly known as BERT in the SEOs and Google circles, is the greatest change since RankBrain was rolled out five-years back. Likely to affect the results of 1 in 10 search queries, it has been introduced to make the search results be a closer match to the queries. In short, BERT makes the responses more relevant to the queries thus creating a better user experience.
How does BERT increase the efficiency of the Google Searches?
In the pre-BERT search results, users would often find that the search responses were based on some of the keywords that they typed in the search query. This would, many a times, lead to multiple searches for the same query, each time worded differently, till they found the perfect result. To avoid such multiple searches and offer more relevant responses to users, BERT was rolled out.
BERT is a more language sensitive algorithm that not only processes the keywords but also the context of the words other than the keywords. This means that the connector words will also be important in a search query now.
Now to understand that better, consider the search query, “Can you get medicine for someone pharmacy”. Earlier, the algorithms would only consider the keywords ‘medicine’ and ‘pharmacy’, throwing up search results which would include generic responses about filling prescriptions. Come BERT and it’s not just the keywords but also other words ‘for someone’ that become important and it is context based results that one gets.
Thus, BERT helps improve understanding of the nuances and context of words in search queries and thus generate more relevant results.
What is BERT?
Without diving deep into the technicality, in short, BERT is the result of Google’s research on transformers. Transformers are models that process words in relation to all other words in a sentence. These models process words in a context rather than individually, in an orderly fashion. Thus, BERT can consider the full context of the word by taking into consideration the words before and after that word. This is what makes it more human-like in processing a query, and gives it the natural language processing skills and enables it to understand the intent behind a search query. Thus, no more will you have to concentrate only on the keywords in a query and can type a query like you would in a human conversation.
So how do my page rankings change after BERT?
While users will have a better experience post-BERT as they get more relevant results, some SEOs might have to rethink their SEO strategy! BERT is likely to affect the results of one in ten search queries. Whether your page still ranks for one of those queries depends on how specific your content is.
To understand this better, let’s dive a bit into how algorithms rank content.
We all know the different kinds of search queries that exist. Out of these, it is the informational related keywords, which are also known as ‘top-of-the-funnel’ keywords, which are easiest to rank for. Given this fact, often times, we have been told that our page will rank higher if the content is lengthy and has a high keyword density. This is further reinforced when we commonly come across long-form content ranked high in searches. But the point we often miss to see is that these content rank high owing to the quality and not just the length.
In other words, a content piece that is just 300 words long may rank higher than a piece which is 10000 words in length and is about twenty different subjects. And, it is such relevant pages and content that are going to get a fillip post-BERT, as this algorithm does not just look at keywords or their density, but the context as a whole. This means that, a page which offers the exact response to that search query will feature higher on SERP.
Gear up for BERT!
So, now that you understand BERT and its nuances better, you know what kind of content can get you the eyeballs and push your competitors lower in the SERP. Go all out and create content in any format, images, audio-visual, text or, any other and focus on serving value to the visitor. After all, in an age of consumerism that we live in, it is only prudent to create a great customer experience and nothing less!