Case Study

How added keyword analysis to save their users tons of time processing NPS feedback

About has built one of the best tools to measure loyalty and customer sentiment using the Net Promoter Score (NPS) system. Their tool helps companies understand their customers while driving organic growth and proactively reducing churn.

The story

NPS surveys are the most reliable and effective method for collecting actionable customer feedback, as shown by the many clients of who use the service for this purpose. The challenge with NPS surveys is that the verbatim feedback provided by end users is completely unstructured.

To gain further insights into user’s comments in these surveys, it is necessary to go through every single response tagging each one to identify trends and themes. This is a time consuming, tedious process, and the consequence for many of’s end users is that quite often this rich feedback is left unanalyzed.

"Going through and tagging NPS results with trends/themes can be a manual and time consuming process. It’s also one of the most important differences between the companies getting the most out of NPS and the ones who just want the score to report to their boss." is a big advocate of how NPS can offer so much more than just a score, and wanted to find an easy way that their customers could get richer feedback from the unstructured text responses that many were leaving aside.

If their users could get deeper, more actionable insights delivered to their dashboards automatically, it would be a great way to add more value to their customers, while further demonstrating the power of NPS surveys.

The Challenge

The team would need to add a new feature to their platform. This feature would leverage machine learning, would need to be able to work with a large and complex data set, and it would have to produce accurate results.

  • The technology also needed to be flexible enough for the platform to produce visualizations and reporting from the results.

  • Having realized what this new feature required, considered either building this feature in-house, or using an external provider of machine learning models.

  • They determined that it would be too expensive and time consuming to dedicate internal resources to this feature, so they began to assess other providers, eventually looking at the following:

    • IBM Watson

    • Amazon Comprehend

    • MonkeyLearn

For, though the other options has strong offerings, MonkeyLearn ended up being the right solution.


  • The team at found that they could leverage the Keyword Extractor model provided by MonkeyLearn, and then create a customized version of this model with high accuracy. Furthermore, they could have the feature in production in a matter of a few weeks time.


“MonkeyLearn hits that sweet spot of being easy to use AND being extremely powerful. Out of all the platforms I’ve tried out, MonkeyLearn has provided the best results by far.”

Chad Keck

Co-Founder & CEO @


MonkeyLearn Inc. All rights reserved 2024