Since the BERT Google update in October 2019, Google has looked to incorporate NLP (natural language processing) into their search algorithm to deliver more accurate search results for users.
Once that happened we have worked to harness Google’s NLP API to provide a tool that allows our clients to optimize their content according to how Google is analyzing content.
The following questions are the most frequently asked questions we get asked about NLP. If you have a question that is not covered here, please contact us, our client success team will be happy to help.
What is NLP
How Surfer uses NLP
True Density vs. True Density + NLP
Which Languages Are Supported in NLP?
How To Check Sentiment in SERP Analyzer?
NLP in Content Editor
Pricing and Plans for NLP Analysis
NLP stands for Natural Language Processing. During the BERT algorithm update Google increased their understanding of content by creating new metrics to analyze the content on the page, through machine learning.
BERT analyzes the context, entities and sentiment of the page. Which means that Google is much more accurate at understanding content context, user intent and the sentiment of the content.
Read an in-depth article that walks you through all the specifics to NLP to find out more.
Google has a NLP tool that analyzes content
We are using it to extract the data about entities and sentiment
Then we cross that data with Surfer’s True Density calculation
We follow a similar process with IBM Watson to measure sentiment of the overall page content.
If you’re interested in finding out more about our decision to use IBM Watson for page sentiment, Slawek explains it in our article on NLP and onpage SEO.
Standard True Density
We crawl top10 results for given query
Then we extract the data about popularity of common words and phrases
Based on the competitor’s content, we determine the density of each word and phrase from the list.
Next, Surfer analyzes your content in terms of the length and use of individual phrases and words. At this point, you already know which terms you should add and which should be reduced, as well as some specific actions recommended for optimal optimization.
True Density + NLP
We send the content from top10 results for given query to the Google NLP API
We cross the data from Google with Surfer’s calculation
We compare your site content to the calculation result
The ultimate goal for implementing the NLP is to provide you even more relevant terms to use within your content.
As the NLP features run in connection with Google’s NLP API we are only able to support the languages that the API supports, which are:
After logging in, you will arrive at the main dashboard. Here you are to create a query in the SERP Analyzer.
If you want to include the NLP feature, simply click “Additional Options” (as illustrated in the image below) and you will be presented with the NLP options to select for this query.
Once you’ve run the query you’ll be presented with the standard query result page, with some added features
It’s important that you understand the overall sentiment of the page. Is it positive neutral or negative. and if so, how strongly is it?
You can access this data from the left sidebar.
Once you’ve opted to see sentiment by ticking the box, the graph will be updated to include the sentiment score.
The sentiment score ranges from 1 to -1. The higher or lower the score the more strongly that sentiment is expressed.
0.25 to 1 means the page has positive sentiment
0.25 to -0.25 means the page has neutral sentiment
0.25 to -1.0 means the page has negative sentiment
The sentiment score is displayed in the graph.
NLP entities are identified by the NLP badge next to the phrase, which can be seen as you open the audit and scroll to the True Density section.
On the True Density section of a query audit you’ll find a positive, negative or neutral sentiment grade.
You’ll find the relevance score on the True Density section of a query audit. It’s scored as a percentage with 100% being the most relevant term.
First thing you need to do is make sure it’s switched on. This works similar to using NLP on the SERP query.
Once you’ve done that, the rest of the process is automated. It simply incorporates NLP into the True Density analysis and search term suggestions. Simply optimize the content as you would using Surfer normally and you are simultaneously optimizing for NLP as you are standard True Density.
Each sentiment and entities analysis costs one credit. You can customize your query within the new input in SERP Analyzer and Content Editor.
The most “expensive query in SERP Analyzer costs you 3 NLP credits. In such a case, you get NLP Entities, NLP Sentiment Analysis, and NLP Sentiment for Entities.
In Content Editor, there’s only one option for NLP Analysis (entities).
Check out the latest Surfer pricing and packages