Predictive modeling helps spot unhappy customers and retain them too

How a LATAM based communications provider employed Machine Learning (ML) technology to reduce churn and increase marketing ROI


Customer churn is one of the primary concerns for all service providers. And, as it directly affects revenues, they seek ways to align themselves to the pulse of the customer.

Determining the churn propensity is paramount for service providers with a subscription model, as it helps them target the right set of customers. In recent times, they have increased their investment in strategic marketing to retain their customers. However, it is critical to justify investments by targeting customers for maximum engagements.

How can a service provider prevent its customers from churning?


38% more accuracy in prediction of the campaign respondents

5X increase in the campaign ROI

32% improvement in the customer retention rate

Client Situation

Our client, a leading communications company in LATAM, was challenged with voluntary churn, which had a direct impact on their top line. They took a lot of initiatives to address churn and improve customer retention. Despite those initiatives, they had to deal with several challenges like a leaky customer base and negative word of mouth, resulting in high retention costs and lower ROI on marketing campaigns.

Beyond predicting potential churners, Prodapt’s solution identified customers who respond positively to marketing campaigns, thus saving millions of marketing dollars.


The client’s existing analytical models couldn’t give them a clearer picture of the potential churners. In addition, there was no way to predict who would respond positively to marketing campaigns. As a result, they ended up targeting an inaccurate customer base.

Solving It

Prodapt’s data scientists collaborated with domain experts to develop an AI/ML engine that identified potential churners and predicted churn propensity.

Further, by leveraging Google Cloud’s AI platform capabilities, we built an ML-uplift model, which helped identify customers who are expected to respond positively to marketing campaigns.

The client could achieve 35%-40% more accuracy in predicting the prospective respondents of marketing campaigns. They increased their marketing campaign ROI by 5X and improved the customer retention rate by 32%.

Let’s connect

How can we help?

We'd love to hear from you.

Talk to a consultant