The Challenge

A major US tech company provides smart home devices to households and businesses. The company has a big customer base with a growing demand for its products. However, they recently has recorded an increasing trend in attrition among tenured customers The company needed rapid support from Pathrithm to build a proactive solution that will identify key risk factors and select customers for an outreach campaign to prevent attrition.

The Solution


First, we connected distinct customer and product data into a unified 360-customer view. Data is updated real-time: Customer demographics, geographic, financial, orders data Product, App and Web usage logs Customer service interactions Other offline and online data
Then we applied machine learning models to identify key drivers of attrition. Top risk factors include: Recent drop in product and app usage Issues with the product navigation and connectivity poor customer service response
We build an early warning system to detect real-time trend changes in key risk factors. We developed a suite of live visualization to track attrition drivers.
Finally, we implemented a machine learning model that automatically identifies current customers with a high risk of attrition. Each customer is matched with an attrition prevention policy, based on their unique risk factors. A personalized message is triggered through an optimal communication channel e.g. text, email, or dispatched for an outbound call.

The Impact

We tested the attrition prevention policy over the course of a few months. We observed an annualized reduction in attrition of up to 10% among active customers. The attrition prevention policy is being continuously tested with a closed-loop measurement and updated through an iterative learning