15% increase in Net Promoter Score by implementing a customer churn prediction model for a leading safety and HVAC equipment solution provider

About the client

A leading product and solution provider for building safety and HVAC equipment.

Problems:

Our client was struggling with a high customer churn rate and reached out to Innover with the following business objectives:

• Development of an analytical model to predict the propensity of churn for B2B customers
• Creation of a customer churn risk score and identification of necessary steps to mitigate the churn of high value customers

Solution:

Innover developed an Analytics and Machine Learning powered customer churn prediction model which included:

• Unification of structured and unstructured historical data from various heterogeneous data sources including - customer information, churn data, products, contracts, service desk etc
• Analysis of this data to create new features and define dependent variables. Then, categorize and cluster this data and deploy different ensemble models for each cluster
• Next step was to perform a multivariate analysis to identify key causal factors that are driving customer churn
• Finally, generation of customer churn risk score as an output of the ensemble models to create risk bins for each customer