EarthLink Connects to a World of Loyalty

Article Featured Image

It’s always cheaper to retain customers than acquire new ones, something that a tough economic climate has made increasingly clear to many companies. EarthLink, one of the largest Internet service providers (ISPs), managed to connect with this reality, but Stuart Roesel, the company’s director of customer insights, analytics, and strategy, reports that it’s no easy feat—in fact, Roesel says, it’s nearly impossible without three key moves. (See box, “EarthLink’s Links to Success,” right.) 

EarthLink has been a SAS customer since 2006—using the SAS Enterprise Miner solution, Base SAS, SAS Enterprise Guide, and, more recently, SAS Campaign Management—but the ISP is continually updating and expanding its predictive modeling programs. 

Complicated analysis is great—for analysts and statisticians. “When you take that out to product management organizations and the people with marketing degrees, they have a difficult time with what that means,” Roesel says. The company was in search of a cohesive analytics solution that would provide a single view of the customer. More specifically, EarthLink needed a better mechanism for associating customer churn with specific customer attributes and products. 

In a program begun earlier this year to identify and further understand the intricacies and nuances of customer churn, EarthLink has applied modeling and segmentation to connect with customers and keep them aboard. The collaboration—between EarthLink’s customer insights, analytics, and strategy team and its product marketing, merchandising, and loyalty team—has uncovered interesting statistics in the areas of product marketing and loyalty. The real goal in the analysis is to build engaging life-cycle experiences, Roesel says. Those experiences have enabled EarthLink to not only lower churn rates to record levels, he says, but also raise profitability and customer satisfaction to all-time highs.

The combined SAS toolset helped EarthLink develop what it calls its Loyalty 2.0 blueprint, part of which involves analyzing churn. Traditionally, EarthLink would score the entire base on a customer’s propensity to churn in the next 60 days. “That drove a very broad strategy and one lifecycle,” Roesel says. “Today, we do our propensity to churn modeling within each segment and have created unique lifecycle experiences across identified key engagement and touch points.” 

Now EarthLink focuses more on the most-profitable customers and spends less time on customers who are bound to churn regardless of intervention. “With this added focus, the company has significantly reduced the number of overall programs, [is] sending more timely and relevant messaging, and [is] focused on extending the life of its most valuable customers,” Roesel says, adding that the intensified segmentation strategy has led to a 55 percent reduction in overall marketing expenses. 

“We now know exactly who they are and how to treat them,” says Chris Schmidt, EarthLink’s director of product marketing, merchandising, and loyalty. EarthLink segments customers based on their levels of engagement, filtering results of programs by “takers” and “nontakers.” 

Roesel says that not only has EarthLink been able to better identify customers more likely to churn, it has been able to run more relevant marketing programs to save those customers. Based on year-over-year measurement of a three-month period, EarthLink has reduced churn by nearly 18 percent. The reduction in churn has had a huge impact on the company’s increased revenue, Roesel says. 

In the past year, EarthLink has begun passing along intelligence from the predictive models to its customer care organization. A service representative gets what Roesel calls a “live whisper” that the customer is at a high propensity to churn. 

Through this identification process, control group testing has demonstrated a 25 percent reduction in churn. Since bringing such modeling to the contact center, EarthLink has added a program in which agents issue a trouble-ticket gift card to those who qualify based on profitability metrics and other data. According to Roesel, customers receiving an agent-issued gift card have a 57 percent lower churn rate than do those in a control group.  


Since deploying SAS Enterprise Miner and its complementary products, EarthLink has witnessed:

  • 40 percent higher customer retention;
  • a reduction in operational marketing expenses by 55 percent;
  • a drop in churn of 18 percent over a five-month period; and  
  • a rise in loyalty and an improvement in retention-campaign responses from 5 percent to 20 percent.
Stuart Roesel, the Internet service provider’s director of customer insights, analytics, and strategy, identifies the three key steps:
  • Employee engagement and collaboration
  • Simplifying analytics for mass consumption in the organization 
  • Tapping into the skill set that turns analytics into strategy.

CRM Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues

Related Articles

SAS Takes Care of the Conversation

With the launch of Conversation Center, the company adds another module to its Social Media Analytics platform.