The Business Intelligence Guide
   BI Strategy | BI Program | BI Projects | BI Data | BI Infrastructure | BI User Tools | BI Vendors | Resource Guides | Articles | BI Blog | BIG Bookstore

Get a FREE Sample of the
The Total BI Guide

and receive the
Just enter your details below

Business Intelligence
BI Strategy
BI Program Guide
BI Tools
- Dashboards
- Scorecards
- Operational BI
- Analytics
BI Software Solutions
Data Management
Decision Support
Marketing Tools
Industry Solutions
Case Studies
BI Surveys & Awards

About the Author

View Gail La Grouw's profile on LinkedIn

Google+ Gail La Grouw

Bookmark and Share

Reducing Customer Churn

Case Study - Connecting with Customers

Company - EarthLink

Solution – SAS Enterprise Miner, complemented with Base SAS with SAS Enterprise Guide for gathering and manipulating large data sets, and performing complex analysis.

Business Requirement – Earthlink wanted to expand from a traditional ISP into new lines of business. In doing so, it wanted to ensure it kept its existing base of customers engaged. The challenge was to associate customer attributes and product usage with subsequent churn. Although the company employed descriptive analytics to track monthly service disconnects, as well as traditional baseline reporting, EarthLink lacked a holistic, real-time depth of intelligence into who its customers were, what they were using and what programs the company could craft to encourage desired customer behaviors.
Goal – to be able to uncovers the right marketing messages and product offers to keep customers coming back


Solution Outcome

EarthLink is retaining customers, increasing product adoption and reaping US$50 million more in total value. The SAS solution has provided the capability to identify customers who might be thinking of leaving and create tailored marketing messages and right product offers to convince customers to stay . The cost of retaining these customers is significantly less than acquiring new ones.

Solution Summary

  1. The team responsible for reducing churn and increasing product adoption modeled customer engagement based on a combination of customer attributes and usage of different products. For example:
    1. Some products were free, others were revenue-generating.
    2. Some products are more closely tied to customer retention than others.
  2. Create initial models and road maps, with input from product managers all the way up to executives. The models aim to identify the strongest drivers to keep customers engaged, to effectively manage churn.
  3. Model usage patterns to design optimal marketing programs.
  4. Using SAS Enterprise Miner, uncover a robust variety of different types of analysis to test different methods to provide better answers about what drives customer churn.
  5. Gain high level of customer insight by examining linear and nonlinear models of several different groups of customers within different usage scenarios.



Measurable results were visible within five weeks of implementing SAS Enterprise Miner.
Reviewing data from multiple dimensions, report on it and drill down into root cause for churn enabled EarthLink to rapidly identify which products are most retentive to different customer segments. They couldn’t do this in the past.

Additionally, the success of the modeling efforts has been easy to promote within the company - loyalty and retention campaign response rates have skyrocketed from 5 percent to 20 percent, and the company has achieved an ROI of 268 percent.
Combining the depth and dimensionality to analyses with the ability to interpret results and communicate to other stakeholders for rapid action has been a significant competitive advantage.

Exciting VPs to talk about the ability to predict churn with high accuracy was powerful and has driven other teams across EarthLink to see how SAS might be able to improve their own results.

Key Wins

  1. Being able to leverage every possible aspect of SAS in a very short time.
  2. The complete modeling platform - automate data cleansing, test different transformations on many evolving models, and quickly create and evaluate different types of models.
  3. Automation of many complex tasks
  4. Being able to answer complex questions quickly through a flexible GUI interface allowed business users to quickly build and transform data sets quickly into meaningful information.
  5. The way SAS seamlessly integrates data from numerous servers, as well as my own data, made it the only option for analytics.
  6. The options presented by SAS provided a variety of ways to develop real-time analytics and identify the drivers, then proactively deliver messages to customers.
    Not being restricted to providing a ‘one size fits all’ solution, EarthLink were able to take a much more segmented approach, which has translated into significant cost savings as well as reduced churn.
  7. In addition to reducing churn, the solution outputs are “helping customers feel more connected to its products through targeted messages about usage, product features and pricing options, giving EarthLink confidence that customer loyalty and product adoption will continue to grow”.


Next at EarthLink

The company is currently on track to move from an all-time high churn rate in 2005 for its dial-up customers to an all-time low by the beginning of 2008 – a 19 percent reduction in churn that has added an incremental value of $50 million – value derived from greater revenue due to extended customer life as well as savings in operational and acquisition costs.

Back to: CSP Case Study Index

Back To Top

BI Case Study Index | Bank EBM | CSP's | Education Forecasting | Local Government | Manufacturer | Port Profitability | Postal Logistics | Retail DCM | Telco Churn | Telecom Data Environment | Telecom Data Retention | Telecom RA

Bookmark and Share


A Strategic Guide To Corporate Performance Using Business Intelligence


The Logical Organization Book Cover