{"id":2011,"date":"2019-11-11T16:46:00","date_gmt":"2019-11-11T21:46:00","guid":{"rendered":"http:\/\/dry-number.flywheelsites.com\/?p=2011"},"modified":"2024-08-28T15:21:15","modified_gmt":"2024-08-28T20:21:15","slug":"google-bert-update","status":"publish","type":"post","link":"https:\/\/www.digitalstrike.com\/google-bert-update\/","title":{"rendered":"The Google BERT Update: What It Is and What It Means for Your Website"},"content":{"rendered":"

BERT is being hailed as the biggest change in search since Google released RankBrain five years ago.\u00a0 That fanfare is warranted because it\u2019s going to change the results that currently rank for one out of every ten search queries.\u00a0<\/p>\n

The bottom line is that Google is becoming smarter about search. And that means so should you.<\/p>\n

Here\u2019s what you need to know about BERT and how it\u2019s going to change your approach to your website content.<\/p>\n

What is the Google BERT Update?<\/h2>\n

Let\u2019s start with the basics.<\/p>\n

In technical-speak, BERT is short for Bidirectional Encoder Representations from Transformers.\u00a0 It is Google\u2019s neural network-based technique for natural language processing (NLP) pre-training.<\/p>\n

\n

In lay terms, it means that BERT is a search algorithm that is going to help computers understand language more like humans do.<\/strong><\/p>\n<\/blockquote>\n

BERT was first open-sourced in 2018 and started rolling out just a few weeks ago.\u00a0 The initial rollout is for English language queries, but BERT will expand to other languages over time.<\/p>\n

Five years ago, RankBrain<\/a> launched and was Google\u2019s first artificial intelligence method for understanding queries.\u00a0 Google designed RankBrain to look at queries and the content of the webpages in Google\u2019s index to better understand the meanings of words.\u00a0<\/p>\n

BERT does not replace RankBrain.\u00a0 It\u2019s an additional way to understand the nuances and context of queries and content so that they can be matched with better and more relevant results.\u00a0<\/p>\n

RankBrain will still be used for queries, but when Google determines that a query is understood better by using BERT, it will incorporate BERT\u2019s methodologies. Depending on the nature of an individual query, Google may actually use several different kinds of methods, including BERT, to understand a query the best way possible.<\/p>\n

\n

With the BERT update, Google is now more inclusive in attempting to understand what the language of a query means and how it relates to existing content on the web.<\/strong> Currently, if you misspell a word in a query, Google\u2019s spelling systems will find the right word for you.\u00a0 Also, if you use a synonym for an actual word that\u2019s in relevant documents, Google can match those as well.<\/p>\n<\/blockquote>\n

This new algorithm handles tasks such as entity recognition, part of speech tagging, and question-answering as well as other natural language processes. \u00a0Also, because Google has open-sourced this technology, others have created variations of BERT.<\/p>\n

BERT adds to existing query tools and creates another way for Google to understand language.\u00a0 Depending on the nature of your query, Google will take one or several of the tools used to decipher a search and provide the most relevant result.\u00a0<\/p>\n

BERT will also impact featured snippets.\u00a0 Featured snippets are a format that provides users with a concise answer to their questions directly on a search results page.\u00a0 As such, users do not have to click through to a specific result.\u00a0 It\u2019s extracted from a webpage and includes\u00a0the page\u2019s title and URL.\u00a0<\/p>\n

Featured snippets provide several benefits to sites that use them effectively. A featured snippet proves that Google chose your page over others as the most relevant to a user\u2019s query. It allows a site to be the quick answer to a specific question, thus beating the competition in terms of visibility and ranking.\u00a0 Because featured snippets provide a visibility advantage, they can also increase traffic to a particular site by as much as 30%.<\/p>\n

How BERT will understand searches better<\/h2>\n

Google research has shown that longer and more conversational searches containing prepositions like \u201cfor\u201d and \u201cto\u201d matter a lot to the meaning.<\/p>\n

BERT accounts for the context of these words in a query. That means users can now search in a way that feels more natural, knowing the results will be more relevant than ever before.<\/p>\n

Google did extensive testing to ensure that the BERT changes are more helpful to relevant search results.<\/p>\n

As part of their rollout announcement, Google cited examples of their evaluation process that demonstrated BERT\u2019s ability to understand a user\u2019s intent behind their search.<\/p>\n

Let\u2019s take a closer look at some of those examples:<\/p>\n

Search for “2019 Brazil traveler to USA needs a visa”<\/h3>\n

\"BERT<\/p>\n

The word “to” and it’s relationship with other words in the query are especially crucial to understanding the overall meaning of the query.\u00a0 Previously, Google\u2019s algorithms would not understand the significance of the Brazilians traveling \u201cto\u201d the United States and would also deliver search results about U.S. citizens traveling to Brazil.\u00a0<\/p>\n

With BERT, this is no longer the case, because it can understand the nuance of “to” in the broader context of the query.<\/p>\n

Search for “do estheticians stand a lot at work”<\/h3>\n

\"BERT<\/p>\n

Previously, Google\u2019s systems tied search queries to matching keywords.\u00a0 In this case, before BERT, Google would have matched \u201cstand-alone\u201d as part of the result with the word \u201cstand\u201d in the query.\u00a0<\/p>\n

BERT recognizes that this is not the proper use of \u201cstand\u201d in this query.\u00a0 Instead, BERT understands that “stand” is related to the physical demands of an esthetician’s job.\u00a0 As a result, it displays more useful results.<\/p>\n

Search for “math practice books for adults”<\/h3>\n

\u00a0\"BERT<\/p>\n

Previously, Google’s search results for this would have included results in the “young adult” category.\u00a0 But BERT can understand that “adult” is being matched differently and can help pick a more relevant and targeted result.<\/p>\n

Featured snippet search for “parking on a hill with no curb”<\/h3>\n

\"BERT<\/p>\n

BERT’s implementation also impacts more relevant featured snippets. In this example, “parking on a hill with no curb” is a query that would confuse Google’s systems in the past. Too much emphasis was placed on “curb” and not enough on “no” and that skewed results.<\/p>\n

Instead of “no curb,” results would be returned that featured “parking on a hill with a curb.”\u00a0 The importance of the word “no” was not taken into account before BERT coming online.<\/p>\n

\n

Featured snippet results are not only improving in English but in many other languages as well. BERT is being used to improve featured snippets<\/strong> in more than two dozen countries where this feature is available. Google already sees significant improvements in languages like Korean, Hindi, and Portuguese.<\/p>\n<\/blockquote>\n

The importance of pre-training data<\/h2>\n

Natural language processing (NLP) relies on training data to help create the best possible search results. One of the biggest challenges is the shortage of training data to better complete this task. Most task-specific datasets only contain relatively small quantities of human-labeled training examples. Sometimes this may range from a few thousand examples to a few hundred thousand samples.<\/p>\n

Modern learning-based NLP models work best when there are more substantial amounts of data. Results improve when there are millions or even billions of annotated training examples.<\/p>\n

To help close this gap and make NLP more relevant and useful, researchers have developed several techniques that use large amounts of unannotated text on the web. That is known as pre-training. Pre-trained models are then fine-tuned on small data NLP tasks that can result in substantial improvements in search accuracy when compared to datasets that are generated from scratch.<\/p>\n

Pre-trained representations can either be context-free or contextual. Why is this important? It means that words can either be stand-alone or related to each other in the context of a search.<\/p>\n

Contextual representations can either be unidirectional or bidirectional. The best results are produced when bidirectional methods are employed.<\/p>\n

\n

Let\u2019s take a look at how that would work.<\/h3>\n