encompass blog

Making Sense of Google BERT

Making Sense of Google BERT

Written by enCOMPASS Agency

If you know much about Google search algorithms, then you probably know they’re never static. On the contrary, Google makes adjustments and revisions to its algorithms on a fairly routine basis. Their goal is always to provide the end user with a better search experience; naturally, these changes have an impact on marketers and SEO professionals, as well.

A lot of Google’s algorithmic updates are quite minor, hardly worth discussing. Every so often, though, there’s an algorithmic update significant enough that Google names it and offers advice to search marketers whose rankings may have been impacted as a result. Such is the case with the so-called Google BERT update, which is one of the more consequential search shake-ups of 2019.

But what does the Google BERT update actually mean for search users and search marketers? Let’s take a closer look together.

What is Google BERT?

To begin, it’s important to understand Google’s overall aim. Like any company, Google has an interest in providing the best possible product for its customer base, specifically the most accurate and relevant search results for its search engine users. To that end, most updates to Google’s algorithms are concerned with better understanding queries, and more accurately connecting those queries to good, solid information.

That brings us to BERT in particular. Basically, the BERT update was implemented to help the algorithms understand natural language better, specifically in conversational searches. Again, it’s all about Google trying to get a better handle on the kinds of information its search users are seeking. And it’s by no means a minor change. Allegedly, BERT impacts up to 10 percent of all Google search queries!

Actually, it’s important to note that BERT goes way beyond being a standard algorithmic update. It stands for Bidirectional Encoder Representations from Transformers, and while it is indeed an update to Google’s natural language processing ability, it also represents a full-blown research project that Google’s engineers have been working on for a long time.

As a Search Engine Journal article notes, “BERT is a natural language processing NLP framework that Google produced and then open-sourced so that the whole natural language processing research field could actually get better at natural language understanding overall.” To that end, you may notice that most online references to BERT aren’t about Google’s algorithms at all, but rather they represent the more robust research activity surrounding machine learning and language processing.

As for what BERT does, exactly, the basic goal is to help cut through some of the ambiguity that exists surrounding certain search queries; for example, to help machine learning algorithms to better distinguish between different uses of a word, to filter through synonyms, etc. BERT provides a framework for understanding what words mean within the context of a sentence, phrase, or search query.

BERT and Google Search

There’s much, much more detail out there about the ins and outs of BERT, but by now you can probably get a good sense of why Google might have an interest in this kind of natural language processing framework. Search engine users are increasingly prone to longer, more circuitous, and less obvious search queries, which makes it harder for Google’s algorithms to discern the user’s intent. BERT helps with that problem.

Also note that BERT is intended to play a big role in helping Google’s algorithms interpret voice searches, which are often more conversational and less formal than a written search query.

But what does this mean for search marketers? Is there a way to optimize your content for BERT?

To answer that second question, Google has indicated that it’s basically a no. BERT isn’t meant to judge or rank content so much as to interpret it, so there’s no sense in trying to “impress” BERT.

With that said, it’s never been more important for content to be written in a way that’s natural, conversational, and easy to understand. Writing with clarity and precision is probably the best thing you can do to make sure that your content connects with the people who might be trying to find it.

We’d add that Google’s enthusiasm for natural language processing, and its big investment in the BERT framework, indicates that they perceive voice search as a pretty big deal, one that’s probably just going to become more significant in the years to come. That’s certainly something to keep in mind when doing SEO work; it’s important pay attention to not only what people are searching for, but how they’re searching for it.

Stay in the Loop

The BERT update is just one example of how, in the world of search marketing, things can change dramatically, and in a short span of time. At enCOMPASS, we ensure that our clients are kept up to speed on new developments that might impact their marketing activity. If you have questions about any of this, feel free to reach out to our team any time.