Content Strategy

Where do the validation metrics in the content strategy tool come from?

Last updated: April 18, 2018

Available For:

Marketing: Basic, Pro, Enterprise
Sales: N/A
Service: N/A

In the content strategy tool, there are four metrics that contribute to the topic validation for your subtopics and core topics. These metrics come from a combination of various sources:

  • Domain authority (core topic only) is calculated using Moz and is a prediction of how well your content will rank on search engines. It's measured on a logarithmic scale of 100 points.
  • Monthly search volume is calculated using SEMrush and is the average monthly searches for a keyword in search engines.
  • Core topic similarity (subtopics only) is calculated using HubSpot propriety machine learning and measures the similarity of your subtopic to your core topic on a scale of 0 to 100%. The higher your percentage, the more closely related your subtopic is to your core topic.

 

To find your validation metrics in HubSpot: 
  • In your HubSpot Marketing BasicProfessional, or Enterprise account, navigate to Content > Strategy.

  • Click the name of a topic cluster. 
  • Click into a subtopic bubble or your core topic bubble to open the sidebar on the right. Your metrics will appear at the top.

Why do my validation metrics read No data

If your subtopic validation metrics don’t pull in any data, it could be due to the following:
  • Moz (domain authority) or SEMRush (monthly search volume) do not have any data to pass for this metric.
  • The locale of your topic cluster is set to a location that does not have any search volume for your topic. For example, if you have a language-specific topic but your locale is set to be a country that does not speak that language, you may not have any search data. You can see locales available for monthly search volume and subtopic suggestions from SEMrush databases here
  • Your topic is specialized enough that our machine learning model does not yet have data available for relevancy and core topic similarity. This is more common for non-English topics that are not easily translated, as our core machine learning models are based in English. See this article for more information.

Was this article helpful?