How to Beat BERT: Combating Google's Latest BERT Update
Nov 07 2019 | 06:00 PM | 5 Mins Read | Level - Basic | Read ModeVandita Grover Contributor, Ziff Davis B2B
Connect with Author
Vandita is a passionate writer and IT enthusiast. She is a Computer Lecturer by profession at the University of Delhi. She has previously worked as a Software Engineer with Aricent Technologies. Vandita writes for MarTech Advisor as a freelance contributor.
Google recently rolled out the BERT Update to help users search better and to improve (Google) Search understand one in 10 searches in the U.S. in English, for now. In this article, we explore how BERT is going to impact your site traffic and how you can crack the BERT code.
To begin with, let’s first understand what BERT is.
What Is the BERT Update For Google Search?
BERT is the Bidirectional Encoder Representations from Transformers technology. It is a neural network-based technique for natural language processing (NLP) pre-training, using which anyone can train their own state-of-the-art question answering system.
THE MODERN CONTENT MARKETER’S BUYER GUIDE
Welcome to the 2019 edition of The Modern Content Marketer’s Buyer Guide. About 10 years ago, marketers realized that content is a critical piece of their pie, and have since been working overtime to generate content to help win the prospect’s attention.
DownloadGoogle introduced and open-sourced BERT in November 2018 after their research on transformers. This year they rolled out the BERT update for Google Search. According to Google Fellow and VP, Goggle Search, Pandu Nayak,
We’re making a significant improvement to how we understand queries, representing the biggest leap forward in the past five years, and one of the biggest leaps forward in the history of Search.
Learn More: Google’s June Update, Search Algorithm & More: Implications for Marketers
How Will the BERT Update Change Google Search?
With BERT, Google intends to understand the complete context of a word in a sentence. Here’s how queries will be interpreted with the BERT update:
-
Search terms will be processed according to their relationship with all other words in a sentence.
-
Words that come before or after a keyword will be considered to understand the search intent.
-
BERT models will be applied to both ranking and featured snippets to show more relevant and useful information in search results.
- Google Search will be able to better understand the context of conversational (or long) queries (by factoring in words like ‘to’, ’for’, etc.) This will make the search process more natural for the user.
Example: If a user types, “2019 Brazil traveler to USA need a visa”. The preposition ‘to’ has a significant relationship with other words and thus plays an important role in the search intent, which is to seek visa information for a Brazilian Traveler to the U.S. and not the other way around.
Here are the query results before and after the BERT update.
Image Source: Google Blog
Screengrab of a Google Search Example
The BERT update is being applied to make search results relevant to people across the world using different languages and also to improve featured snippets. Significant improvements are also visible in languages like Korean, Hindi, and Portuguese.
But what does this mean for your site’s search rankings? Should you worry, and how do you combat the BERT Update? Let’s find out.
Learn More: Google Announces Gmail Dynamic Email: Impact on Email Marketing
How to Be 'Found' With the BERT Update?
According to Google's search expert Danny Sullivan, "There's nothing to optimize for with BERT, nor anything for anyone to be rethinking. The fundamentals of us seeking to reward great content remain unchanged."
The BERT Update is more about creating better search experiences for users and presenting them with the relevant results.
So, to rank as one of the most relevant pages, you will have to focus on creating great content along with context.
Here is what you can do to be 'found' with the BERT Update.
1. Create Compelling and Relevant Content
Danny Sullivan said in a tweet, "BERT doesn't change the fundamentals of what we've long said: write content for users."
Compelling, creative content that is relevant and useful to users will help you reach the top of the Search Engine Results Page (SERP).
Example: If a user searches for 'effective home remedies for the flu'. A content page that shares tips and home remedies for the flu is likely to rank higher as compared to a page that sells a product as a home remedy.
Why?
Because this piece is more relevant to the search query.
But, then how do you promote your product with this page? Isn't content creation all about making users aware of your product?
Yes, it is. And you can still do it.
How about arranging your content piece to focus on home remedies and including information about your product as part of the piece. Say, for example, 'if these home remedies for cough don't work for you, you can try' and explain more about the product with a clear call to action.
2. Context Matters More Than Keyword Density
Search Engine Optimization was no longer about keyword density, even before the BERT update. Keywords merely act as pointers to the trending searches but using keywords as clickbait in titles or sprinkling them throughout your content will not help.
As the Google blog says, the focus is more on the context which is determined by processing a word (keyword) in relationship with other words in the sentence, including the prepositions and preceding and succeeding words. It is therefore better to structure your content around how you would address a user query or resolve a problem. Instead of focusing on the content length or keyword density, try to solve users' problems with in-depth, direct answers that match their intent.
Does that mean you have to restructure your content, to keep up with the BERT update?
Image source: Twitter Page Danny Sullivan
Screengrab of Danny Sullivan’s Tweet
BERT doesn't assign any value to pages. Google is attempting to better understand search queries and the context of the pages to be more relevant to the user.
Understanding language is a complex task and an ongoing challenge, so it will take Google's machine learning algorithms some time to learn more about the language, meaning, and context.
What should be your SEO strategy to combat BERT?
Well, firstly you don't have to combat BERT, because it is eventually going to help your content stand out when the internet is flooded with content.
Moreover, you can't optimize for BERT. Until the BERT algorithms are better trained to learn more about the context in English and other languages, not only in the U.S. but across the world, you can research about the pain points of your target audience, what they search, how they search, and invest in quality content creators.
3. Long Tail Terms or Short Phrases
There is a lot of confusion about whether long-form content will work or whether SEO should focus on long-tail terms or short phrases.
Should we focus on long-tail terms and long-form content?
Image source: Twitter Page Danny Sullivan
Screengrab of Danny Sullivan’s Tweet
Yes and no.
Yes, because BERT will be evaluating content at sentence or phrase level so creating content around specific long-tail terms will be effective. Moreover, creating in-depth content that can provide valuable information to your audience will better fit the search result.
No, because it is more about how people search in natural language and not about how long a query is.
Thus, while creating content, think about how a user types or asks their queries. If your content is conversational and can answer specific user questions, you will not need to worry about long content or long and short phrases.
Learn More: Google’s March 2019 Core Ranking Algorithm Update aka Florida Update 2: Tips for SEO Marketers
BERT Will Get Better
No matter what you’re looking for, or what language you speak, we hope you’re able to let go of some of your keyword-ese and search in a way that feels natural for you. But you’ll still stump Google from time to time. Even with BERT, we don’t always get it right...We’re always getting better and working to find the meaning in-- and most helpful information for-- every query you send our way.
~ Pandu Nayak, vice president, Google Search
For now, BERT affects only 10 percent of the searches in the U.S. English, but as BERT models get better trained, Google will bring BERT to more languages and locales over time. Until then:
- Craft compelling and informational content.
- Get the context right.
- Prepare for conversational searches.
- Remember, the search will no longer be an input to an algorithm, rather a query in a more natural form. Think Voice Search.
- Don't optimize for algorithms and updates, optimize your content quality for your audience.
Are you already seeing the impact of BERT? Write to us on Twitter, LinkedIn, or Facebook.