Google BERT: Moving to Customer-Driven Dialogue Marketing

#1
Posted by Ann Smarty - Brand and Community Manager at Internet Marketing Ninjas, Co-Founder of Viral Content Buzz and Founder of MyBlogU.

Google's BERT algorithm helps the search engine understand natural speech better to provide more relevant search results. What does this mean for marketers?

Modern technology allows consumers to dictate their questions directly to their mobile device. However, Google's more spontaneous voice queries are not always easy to interpret.

This is why the search engine is constantly working to improve the understanding of spoken language in general and voice queries in particular.

To date, Google has released two major updates aimed at improving the natural language processing algorithm:

  • Hummingbird (and RankBrain)
  • BERT

The Hummingbird update was released back in 2013. This was an earlier attempt by Google to understand words in context rather than match verbatim queries to web pages.

With the release of Hummingbird, the digital marketing community has learned a new concept of things-not-strings. This meant that Google now uses its existing and ever-growing Knowledge Network (i.e., its database of facts) and semantic analysis to understand not only the literal meaning, but also the context and intent of each request.

0_10c41a_b6e29a99_l__71dec3ab[1].jpeg

Two years later, Google also launched RankBrain, an artificial intelligence system that helps it better process and understand search queries. It has been stated that RankBrain is designed to help Hummingbird's algorithm perform better.

No other details of these algorithms were disclosed.

What you need to know about updating BERT

The BERT update was announced in 2019 as an attempt by Google to better understand search queries in general. It was very similar to the concept of "things, not strings", but only appeared five years later. BERT is a big step in the same direction.

The example below clearly shows the difference: after implementing Hummingbird, the algorithm did a good job of understanding the context, and after launching BERT (after five years of machine learning), it interpreted the query itself better:

bert-result-800x501__d547d4a9[1].jpg

Simply put, these algorithms have a lot in common, but BERT is more advanced, especially for "longer and more conversational queries" - those that can be interpreted in different ways, especially if you are a robot with no sense of humor.

Is it possible to optimize for BERT and Hummingbird?

Optimizing for them is not possible because both updates were released to help Google understand search queries better. By default, a person does not need to do anything extra, since the human brain already copes with this task quite well. The only problems were the cars.

Algorithm updates like these should make our lives easier as well, because digital marketers won't have to do anything extra to get the meaning of the content on their page to Google. Of course, it hasn't come to this yet, but these updates bring this day closer.

For the same reason, not many people noticed any changes after the release of these updates. Hummingbird was deployed several months before its announcement, and no one noticed it. The impact of BERT was also quite moderate.

What then is to be done?

In fact, these updates do not require any changes in tactical optimization. However, they are intended to hint any digital marketer about the need to move to a customer-centric content marketing model.

Customer-centric content marketing

A customer-centric strategy puts the target audience in the spotlight. Meanwhile, many digital brands are targeting search engines (organic rankings) and word of mouth (virality) rather than their actual customers.

In addition, a customer-centric strategy involves the use of tools and research methods that are not found in most traditional content strategies.

In particular, this strategy includes building a customer profile, personalization, testing the sales funnel and optimizing for intent.

There are many tools that can help you with this, including:

  • Google Analytics and Finteza (profile building)
Finteza offers detailed demographic reports showing the user's location, device model, installed apps and organizations:

3on0wr-80agrn-2020-12-16-e-14.11.45__3974e4ac[1].png

The tool collects a lot of information to form a user profile and even allows you to estimate the average income and age of the audience.

The demographic and interest category reports in Google Analytics help you better understand your audience today and create content targeted to each user segment / profile.

google-analytics-personas-800x408__8b7abdc7[1].png

You can also add these reports to your WordPress dashboard to make it easier for your authors to target the right audience category.

Alter (personalization)

Alter allows you to personalize content suggestions using artificial intelligence - it allows you to predict what will be interesting to read next for each new or returning visitor.

alter-personalization-768x385__1bc13be4[1].png

  • Text Optimizer (optimization for intent)
Finally, if you understand the intent of your content (intent), then you should also understand whether you have articulated and structured it well. Text Optimizer is the fastest way to know if your content is targeting the right intent.

textoptimizer-intent-768x247__83d614e8[1].png

Start monitoring social media

Google's BERT and Hummingbird algorithms have shown one obvious trend: the shift to conversational marketing.

People interact with their devices in a different way these days. They talk to them by asking questions or giving commands. Therefore, Google's actions to understand and interpret these “conversations” can benefit companies as well.

You need to build your overall digital strategy in the same way. You should study exactly how people speak in your niche. Your brand and your target customer must speak the same language.

Social media monitoring is the best way to collect live conversations, archive and act on them.

Agorapulse offers powerful social monitoring tools to track conversations in a specific niche, process them, and even delegate any questions to other team members so they can use them when planning content:


Conclusion
Hummingbird and BERT are two of Google's many attempts to better meet the needs of a smarter consumer who uses devices in new ways as part of their daily lives.

There is no specific optimization tactic for any of these updates, but following Google's lead is not a bad idea. Take a similar approach: a customer-focused marketing strategy that puts the customer in the spotlight and seeks to better understand and meet their needs.