<img height="1" width="1" style="display:none;" alt="" src="https://ct.pinterest.com/v3/?event=init&amp;tid=2613462408584&amp;pd[em]=<hashed_email_address>&amp;noscript=1">

What Google's BERT Update Means for Marketers

May 7, 2020 4 min read


Google’s advice for how to optimize for its latest major update is: don’t even try to optimize

Instead, Google just wants people to just write for humans and stop trying to game the major “BERT” algorithm update that rolled out last month. 

“Write better content” is great advice in general. But it would be naive to think that a major search engine update won’t create winners and losers. 

As marketers, we’re responsible for making sure our clients don’t get hurt by the new changes to search engine rankings. After reading up what's known about BERT so far, we came away with one main takeaway:  In a post-BERT world we should devote more time developing short pieces that can answer highly-specific user questions. 

BERT will greatly help searchers find answers to these specific conversational queries. In practice, our new approach will likely mean devoting less time to obsessing over rankings for main head keywords, although those short high volume keywords will remain important for the foreseeable future. 

This optimization strategy is a bit more specific than “write better,” although not as specific as the SEO tricks of yesteryear.     

Meet BERT, friend of 🐧 and 🐼

BERT is the newest of Google’s major algorithm updates. 

Google makes changes to its algorithms multiple times per day. Some affect only small numbers of searches. Others completely re-write the rules of the search economy and make us in the digital marketing business want to tear our hair out. 

Why all the constant changes?

"At its core, Search is about understanding language," explained Google Vice President Pandu Nayak in a recent blog post. BERT's major update aligns search results with how users are using specific context in their search queries today.

For 2019, Google designed BERT to understand questions like “do estheticians stand a lot at work?” which are simple questions for humans to understand, but were traditionally challenging for machines to interpret because of the complexities of human language and grammar. 

According to Nayak, previous Google systems would not have understood that the words “stand” and “a lot” should be grouped together in the “do estheticians stand a lot at work?” query. The context of these terms are key in order to return relevant results about the physical demands of the work of an esthetician.

BERT is an acronym for Bidirectional Encoder Representations from Transformers, the machine learning technique that helps Google interpret and sort these conversational queries. The largest updates get a cute codename. Some big ones from the last decade include Caffeine, RankBrain, Hummingbird, Pinguin, Panda, and now BERT.

Google began rolling out BERT on the week of Oct. 21, 2019.

How does the BERT update affect results?  

As a natural language processing tool, BERT is fundamentally about understanding longer conversational search queries  and serving related results based on content that typically lives deeper within a page as opposed to a title or main keyword. Its ability to connect contextual queries with relevant results happens here, beneath the surface, in the world of long-tail keywords

To continue with Google’s esthetician example, BERT is unlikely to affect the search results for a query like “esthetician near me,” or high volume head keywords like “massage,” or “spa.” 

Since its launch, BERT affects only 10 percent of Google searches, according to Nayak’s post. Considering that Google handles over 5.6 billion search queries a day, that’s still a significant change to the world of search. (And yes, fellow linguists, the BERT update is being applied to languages beyond English.)

The widely-accepted explanation for BERT’s arrival with such *relative* little turbulence is that BERT doesn’t do much with the head keywords that current search marketing tools track. 

What is search engine turbulence?

In order to help marketers and industry experts stay up with Google’s constant updates, trackers exist to measure the fluctuation in result rank.

Tools like Moz’s MozCast demonstrate how Google’s tinkering is affecting search engine result pages. MozCast calculates search engine turbulence by comparing how the search results page for 1,000 handpicked keywords changes from one day to the next. These changes are reported to users using a weather forecast as an analogy. Warmer, stormier days indicate lots of upsets and changes in search engine results. 


The MozCast for the week of Oct. 21, was fairly cool and calm, compared to major Google updates in the past. Other search turbulence tools showed a similar pattern.

Below is a screenshot from Algoroo, another search engine turbulence tracker that shows the turbulence from BERT as compared to an earlier update, Google’s September 2019 Core Update.


Quick takeaway for businesses: The lack of head keyword turbulence should mean businesses won’t see dramatic changes to their website traffic from head keyword dominance suddenly disappear as Google rewrites the rules of the game. BERT doesn’t demand urgent attention from website administrators.

That doesn’t mean the new algorithm doesn’t create opportunities for generating new traffic to business websites. 

Our post-BERT plan 

Luckily, our agency’s basic framework for SEO as a foundational marketing tool is largely unchanged. Our strategy around optimizing for BERT means pivoting webpage content to focus on anticipating future search behavior.

Primitive versions of the search engines of 10 to 20 years ago taught people to search using search engine-esque queries with explicit keywords for the bots to understand, like the input [esthetician, “physical requirements,” job, standing].

Improvements in search (including BERT), as well as the popularity of mobile devices and voice-activated digital assistants (Siri, Alexa, Google Home, etc.) mean more people in the future will ask “do estheticians stand a lot at work?” and be able to get more relevant and useful answers. 

This shift doesn’t mean that marketers should stop competing for high-value head keywords, which are likely to remain powerful traffic generators even in a world where most queries are long conversational questions. But as content marketer Neil Patel observes, the new algorithm creates opportunities for attracting traffic with long-tail keywords. 

AKA: As more people go to the internet with long, complex human questions, the businesses that answer them will get more traffic.  

Case in point: the question “do estheticians stand a lot at work?”

The top Google search result is currently claimed by the personal blog of some guy named Nathaniel Tower. He’s clearly using the query as a sandbox to play with Google’s new algorithm. But marketers should notice that his short post designed to rank for “do estheticians stand a lot at work?” addresses the question more directly than the actual esthetician career web pages, which include a forum on Indeed.com. For now, Nathaniel Tower’s post ranks above them all.  

So much for the argument that you can’t optimize for BERT.

Thousands of Marketing and Sales Pros Get Their Inspiration from The Latest

Marketing insights and sales strategy delivered to your inbox bi-monthly.

Become a DMG Insider

This post was originally published November 13, 2019