Google updates their search engine all the time, so it’s hardly news when there’s an update. But every once in a while, Google makes a change that alters the SEO landscape significantly.
When that happens, those who adapt first win.
NLP, which is part of Google’s BERT update, is one of those changes that is going to matter. Let’s rewind to the end of 2019. Google announced the BERT algorithm release.
Here’s something important you need to know about Google. They only tell you about it when it’s a biggy, so the fact they announced this update should have everyone nervous.
So, the SEO community tested, analyzed, and tested some more, we listened to Google, and we figured out that BERT focused on content, in terms of quality and context. And to do that, they are using NLP – Natural Language Processing.
So, Google think NLP is a big deal, and they think BERT is a big deal. Who are we to argue?
What is BERT?
Without getting into the weeds, BERT stands for Bidirectional Encoder Representations from Transformers.
Without going too far into the weeds, BERT collects a whole bunch of data points on the page, and then NLP acts like the brain to analyze and understand those data points. It tries to understand the words and phrases, to understand the power of the content, and how it matches the search query made.
Google looks at all that data and compares it against other pages for the same search and then determines which offers the most relevant and best content for the user. This is how Google now understands content.
Understanding Google’s motives
Google was first to use backlinks as a way of determining which pages and sites offer better value than others. Which pages should be presented to users.
That’s how Google was different from any other search engine. It’s how they became so dominant.
But over the last five years, Google has shifted significantly. Their ultimate goal is to be the search engine best at delivering relevant and high-quality search results. Users want to find the pages that best provide the information they are looking for, as quickly as possible.
Google knows that if they don’t continue to be the best at doing that, their grip on power will be diluted.
But now Google has moved away from backlinks a little and is looking to achieve this by better understanding what the user is looking for. Are they looking to buy something? Are they looking for a review of a product, some information only?
Traditionally, they would have analyzed previous searches. They would see what pages other users clicked on to determine the type of page they should be displaying the user.
However, according to Google, 15% of searches are new queries. That means they can’t look at historical information to understand what the user wants. They need a new plan.
So the geniuses at Google put their heads together and came up with using NLP. Natural Language Processing. An AI learning tool to better understand the content on the page.
What does this mean?
It means that Google has made a huge shift towards understanding content with NLP. And if Google is heading that way, it makes sense that we all start moving towards NLP too.
NLP Sentiment and NLP Entities
There are two main areas to focus on when looking at optimizing your pages to take into account NLP. These are optimizing for sentiment and entities.
Every page, section paragraph, and even words are given a sentiment score. It is a range but it can also be classified as a positive word, neutral, or negative word.
Google actually scores them. The range is between 1 and minus 1
Positive sentiment is scored between 0.25 and 1
Neutral sentiment is scored 0.25 and -0.25
Negative sentiment is scored lower than -0.25 to -1.
Why does sentiment matter?
Studies have shown that if the top 10 pages in the SERPs have a positive sentiment score, there is little to no chance your page will be ranked in the top 10 with a negative sentiment score.
It’s that simple.
Entities are probably the most critical component of NLP from an SEO perspective because they are the most actionable data point we can use to improve our rankings.
The entity is a word or phrase that represents an object. Some examples of objects are:
- consumer goods
NLP entities are the words or phrases that help to place and describe the object.
For example, if a page was about breakfast, then words like “morning,” “coffee,” and “toast” could be words that Google sees as important words on the page that they use to understand the content.
What is the Impact of NLP on SEO?
This is the most important question. How does NLP change the way we optimize our websites for Google?
It actually can impact many aspects of SEO, from content optimization to competitor research. Let’s walk through some of the main changes you can make to take into account NLP.
There is a well-proven and used theory that says you look at what is ranking at the top of Google’s search results, and you mimic it. If Google likes this page and trusts it to rank it #1, then it must be doing something right.
In terms of content, it’s called TF IDF – term frequency-inverse document frequency. It’s a calculation that looks at which words are most commonly used on the top pages and either comparing them to your page or using that data to create a new page.
With NLP, you can go one step further.
You can use Google’s NLP API to find out which words Google identifies as NLP entities, and then you can analyze how often those NLP entities are used by the top-ranking pages.
We know Google thinks these terms are important, and we know that they like the top pages, so it makes sense to mimic them in terms of NLP.
NLP and Keyword Research
The great thing about the BERT update is that it allows us to have a “before” and “after” snapshot of how Google sees the content.
Some pages saw an increase after the BERT update; others saw a drop in the rankings.
If you can track the changes and then analyze those pages that saw an increase and compare with those that saw a decrease, you’ll be left with a list of keywords that Google appears to deem important and potentially how often to use them.
You can take keyword research and flip it to look at when you are creating new content.
You want to be able to determine how strong the competitors are and how that impacts your SEO strategy.
Using the same theory again, you can make the assumption that all things being equal, what Google likes from the top pages and sites, they’ll like from your own pages and sites. That includes NLP analysis.
NLP is a big deal. Google has signaled that they are trying to understand content more effectively. The result of this was the BERT algorithm update, which Google has told us it a big deal.
So we know it’s important and history tells us, those SEO strategies that adapt quickest to Google’s changes get to dominate the SERPs.
NLP is about the words that Google uses to better understand the objects on the page. And the overall sentiment of the page too.
We can use Google’s NLP API to understand specifically which terms Google deems are important; then, we can compare the top-ranking pages to understand how often we should use those terms to be optimized.
There are over 240 ranking factors for SEO. It’s tough to know what to focus on. But, Google has given us a massive signal that NLP and content are important. SEOs who follow their lead are getting results. Maybe you should too?