In case you missed it, Google has just rolled out an algorithm update called BERT. And this change with have far-reaching consequences for SEO.
Introducing: BERT. Or, Bidirectional Encoder Representation from Transformers.
Ok, we’ll stick with BERT.
This algorithm change foreshadows the leap Google is making in natural language processing abilities. Google says the update will affect 1 in 10 search queries. In fact, SEO experts consider BERT the biggest update since RankBrain in 2015.
But first things first:
What Is BERT?
Google has always had trouble understanding certain types of search queries, particularly informational queries. Even Pandu Nayak, Google Fellow, and Vice President, Search admits it.
This is one of the reasons people use ‘keyword-ese’, typing a string of words they think Google will understand—instead of presenting a question in a natural way.
For example, you wouldn’t ask Google something as straightforward as: “is a marketing director’s day-to-day work harder in comparison to other roles in the same company?”
But rather: “marketing director work harder than hi-tech jobs”
Right?
The great minds at Google are aware of the fact that with today’s AI and voice helper disruptions, users would become impatient with search results that are not precise to what they actually need.
Thanks to the BERT update, Google Search is now better-equipped to understand longer, more conversational queries, or queries where prepositions like “for” and “to” matter to the meaning. It improves the accuracy of the SERPs (Search Engine Results Pages) and featured snippets.
But I know what you’re thinking:
Seriously Now, What Is BERT Really?
No better way to explain it than to show you BERT in action:
Let’s say you search for “parking on a hill with no curb.” The word “no” is a key factor in this query, right?
Look how BERT changed the way Google answers this query:
See how precise the “after” result is?
Problem solved for the person who needed a quick answer before he makes a mistake that can cost him a new bumper.
Let’s look at another example:
This time you need to know: “Can you get medicine for someone pharmacy.”
The word “someone” holds the key to your search intent. You don’t need the nearest pharmacy or to buy medicine to cure your own headache.
You want to know if you can get medicine on behalf of “someone.”
Let’s see the power of BERT:
Spot on!
So now let’s talk about how we, as marketers, can harness the power of BERT.
Google’s BERT Update Mostly Impacts Informational Queries
We’ve talked a lot on this blog about the 3 types of user queries: navigational, informational, and transactional.
An informational query, such as “what is home insurance” or “how to use a dating app,” indicates that the user wishes to learn more about a specific subject.
Content marketing owes its existence largely to informational queries. After all, one of the main purposes of content marketing is to provide information about a product or service with the aim of building up brand recognition. This differentiates it from pure brand marketing campaigns, which are better-suited to transactional queries.
Informational queries tend to be longer and more complex than the 2 other types of queries, and therefore need BERT’s helping hand.
Navigational queries, like “Allstate property insurance” or “HelloFresh meal kits,” are fairly straightforward for Google to understand.
And transactional queries, like “best home insurance for mobile homes” or “cheapest home insurance in Pittsburgh,” are also a no-brainer for Google.
Thanks to BERT, Google will provide more accurate responses to longer, more complex informational queries. It will enable users to be more natural in their interaction with Google—in other words, to have a conversation with the search engine.
And we must be prepared with the proper content.
The Good News Is: High-Quality, Focused Content Wins
Throughout its evolution, Google has been transparent about its desire to reward websites that produce high-quality content. The 2015 RankBrain update represented a major milestone in Google’s quest to elevate webpages that provides strong user experience and punish pages that employ outdated tactics like keyword stuffing.
BERT is the next phase in this journey: although powered by artificial intelligence, it has the effect of rewarding content written for humans.
As Danny Sullivan, Google’s public search liaison puts it: “Have good content.”
Honestly, I don’t know that it’s so much our statements on ranking factors as much as SEO who dig for layers not there. Have good content. That shouldn’t make anyone cry. “Does good content mean 300 characters or 1000 characters?” That makes me want to cry….
— Danny Sullivan (@dannysullivan) 11 October 2019
BERT doesn’t really have anything to say about what type of content to produce. It merely reinforces the fact that you should be producing good content—and offers potentially greater rewards when you do.
Word count doesn’t really matter anymore (between you and me, it never really did). It is more about answering search queries as specifically as possible.
If you want to rank higher for informational queries, be super specific and answer the question better than your competitors. Whether you produce videos, images, recordings, or plain and simple articles—do your utmost to answer thoroughly, clearly, and in an engaging way.
And the last word of advice to survive the BERT upgrade:
Write for Humans, Not for BERT
For anyone who has been following Google closely over the years, BERT is another step in the search giant’s mission to master natural language processing. That is, to be able to process and analyze human language with precision. Ironically, the more sophisticated Google’s algorithms changes become, the more Google rewards websites that produce content for people—rather than for search engines.
And writing for human beings is not something I have to teach you.
Right?
If you want to continue the conversation, please feel free to contact Natural Intelligence’s marketing specialists. Simply fill out this short form and we will get back to you.
Partner with Natural Intelligence
