Case Study
Client
Local Telco Service Provider
Project
Reduce DSL Churn
Background
DataCo has built many anti-attrition and/or churn models for various industries. One of our clients, a local Telco service provider, asked us to build a model to help them route inbound calls to a special agent group within customer service. The goal was to proactively treat and save Telco customers who were at risk of closing their DSL accounts with the company.
Objective
Build a statistical model that proactively detects characteristics of DSL customers who are most likely to close their accounts within 90 days. Validate that the model rank-orders the customer file based on actual churn rates.
Execution
DataCo built a churn model and scored the Telco's DSL accounts to determine each customer's likelihood to close the DSL service within 90 days. Churn rates were projected at each 10% increment of the file (i.e., at each decile). Those scoring in decile 1 were most likely to churn the DSL account. After 90 days, the Telco provided a then-current DSL customer file, from which actual churn rates were calculated for each decile.
Results
When the actual churn rates were calculated, 82% of DSL customers who closed had scored in decile 1 ninety days earlier. 94% of the customers who had closed the DSL account had scored in deciles 1 or 2. That means that marketing efforts could be focused on 20% of the DSL customers to almost entirely address the churn threat.
|
