85% accuracy in Customer Churn Prediction for a leading Bakery Ingredients Manufacturer

About the client

A leading bakery ingredients manufacturer based in the US

Problems:

• Our client, a leading bakery ingredient manufacturer was catering to bakeries, supermarkets or specialty stores. However, a high majority of their customers were going inactive or churning just after a few orders
• The client wanted to understand the reason behind high inactivity/churn risk propensity and take proactive measures to eradicate customer churn

Solution:

Innover leveraged its proprietary Analytics Platform - LEAP^TM ready-to-use algorithms for customer churn prediction to resolve client issues with speed and precision

• Our experts analyzed historical data (customer profiles, product data, contract data, churn/ inactivity related data, miscellaneous data like rejected orders, repeat cancellations etc.) from heterogenous sources to create new features and defined dependent variables to predict the churn risk score
• The team also leveraged Model Generation methods – Decision Tree classifier, Random Forest Classifier and others to derive customer risk scores as output, along with key factors for churn/inactivity