Case Study

Grab uses sentiment modeling to measure customer satisfaction and increase retention


Grab has over half a million drivers across South-East Asia. 

In the past, the company would send NPS (Net Promoter Score) surveys to each driver asking, “Would you recommend Grab to a Friend”?. 

When asking drivers this question, the response rate was quite low and insights weren’t actionable.

Grab needed a new scalable way to measure their customer satisfaction and loyalty and take more targeted action to keep customers happy.



Instead of measuring satisfaction and loyalty and NPS, a natural language processing model was developed, assessing customer satisfaction and loyalty based on the large pool of driver behaviors and interactions tracked and available.

By using embedded transformer models, Grab translated the free text into a scale similar to NPS, but updated in real-time. This allowed them to both have a constant pulse of customer satisfaction, as well as better target drivers to prevent churn.


When combined with Grabs retention model, Grab was able to proactively reduce churn through targeted retention campaigns which triggered automatically based on customer sentiment and predicted churn.

Our data scientists have worked with companies like

Why Measurly

Gain your competitive edge

Tailored to your business

Data models fully customized to your business and unique challenges

Best-in-class methodology

Advanced ML techniques coupled with powerful human guidance

Built for scale & automation

Fully automated models that allow you to improve engagement at scale

Increase your customer lifetime value with predictive ML personalization

Book a free tailored demo to learn how Measurly can help your business grow.