One of the latest examples of this innovation and development is the new BERT algorithm update announced in late October. In Google's announcement, the BERT algorithm update was described as the most important algorithm update since the artificial intelligence-supported Rankbrain algorithm update announced in 2015.
What is the BERT Algorithm?
The BERT Algorithm, like many other algorithm updates by Google, was developed to better understand queries and provide more accurate results to its users. The BERT algorithm, which consists of the initials of the expressions “Bidirectional Encoder Representations from Transformers”, can be explained as a natural language processing technique that uses artificial intelligence and machine learning technologies together. With this technique, Google aims to better analyze the user's search query and provide more accurate results in line with the user's search intent.
With this algorithm update, Google will start to evaluate the words employment database in the queries made, instead of evaluating them individually, together with the words before and after them or similar and synonymous words. In this way, it aims to understand long-tail or complex user queries much better. In addition, Google states that with the BERT algorithm update, it will understand much better what the conjunctions and prepositions used in the queries add meaning to the query. In addition, according to the statements of Google officials, the BERT algorithm will be used for the development of Featured Snippets as well as general searches.
In Google's statement, it is stated that approximately 1 in every 10 search queries will be affected by the BERT algorithm. This statement also indicates how important the BERT algorithm update is. However, although it is an important algorithm update, no direct optimization work can be done against the BERT algorithm. We will discuss the details about this subject in the following headings. Because Google's aim with this algorithm update is not to completely change the search engine results pages, but to simply understand the queries better and provide more accurate results to its users.
BERT Algorithm with Examples
For example, before the BERT algorithm update, when a Brazilian traveler made this query on Google about a US visa, they would encounter the Washington Post's news. Because Google could not correctly understand the user's intention here and presented such a result to the user. However, after the BERT algorithm, it began to directly present the embassy and consulate page to its users, especially since it understood the query more correctly with the "to" conjunction in the query.
In another example, the user wonders whether it is possible to get medicine from a pharmacy on behalf of someone else and performs the following query on Google. However, since Google could not fully understand the search intent, it presented the user with general content about filling a prescription. However, after the BERT algorithm, we see that it fully understands the query and presents the user with content that is the direct answer to his query.
Finally, in the example below, we see that even though there is no expression related to the Eiffel Tower in the user's query, Google analyzes the meaning of this query very well and presents the information its users are looking for in the featured snippet.
As a result, it is possible to say that Google has better understood user queries and the search intent in these queries with its new algorithm update, BERT. If we look at the examples above, it seems to have been successful in this regard. However, after this algorithm update, everyone has only one question in mind: “How can I optimize my website against the BERT algorithm?” Let’s answer this question together.
Things to Consider After BERT Algorithm
With the BERT algorithm update, we can explain Google's aim as providing more accurate results for complex user queries rather than offering a ranking algorithm that will completely change the results pages. Therefore, as Google official Danny Sullivan said, there is actually no optimization work to be done directly against the BERT algorithm. However, let's examine together below both the measures that can be taken indirectly against the BERT algorithm and to meet Google's expectations from websites in general.
