Satellite television dishes dot the American landscape like many mushrooms after a rainy spring. For EchoStar Communications Corp., commonly known as "Dish Network," this market saturation is both a blessing and a curse.

George Vigil, manager of enterprise analytics for EchoStar Communications Corp., says his company needed—and found—help managing the data generated by EchoStar's huge growth.

With more than 13.1 million subscribers and 14 satellites that deliver more than 2,600 channels, EchoStar's $8.4 billion in revenues make the company the second largest direct broadcast satellite (DBS) television (TV) provider in the United States. Twenty years after filing for its first DBS license with the Federal Communications Commission from its headquarters in Englewood, Colo., EchoStar stands face-to-face with the challenges of a mature industry: growing the subscriber base, improving customer retention and reducing costs while increasing efficiencies. Co-founder and CEO Charlie Ergen may have built the company by giving away free dishes to rural homeowners, but maintaining success in today's DBS market is a whole new game.

Increased price competition and converged media offerings from Internet, cable and telecommunications companies threaten to siphon customers from traditional DBS providers. Simultaneously, skyrocketing costs jeopardize profitability. The expense of supporting a vast array of programming is rising along with the need to operate retail initiatives and staff call and service centers. Additionally, there is the expense for equipment: the cost of launching satellites, purchasing frequency rights, and acquiring ground facilities and related assets is staggering.

Information overload
With a reputation for innovation and technical excellence, EchoStar understood the need to leverage enterprise information as a business asset. Yet optimizing the use of corporate data was not always easy.

Early on, the company built an Oracle data warehouse to store customer information and transactional data. However, the discrete databases included in the data warehouse were simple replications of the company's operational systems—creating a "vault" where the data could be interrogated by IT-written queries. "As the customer base began to grow, the Oracle solutions couldn't handle the volume of data," explains George Vigil, manager of enterprise analytics for EchoStar. "It took too long to process, and it was hard to use the data to meet the needs of internal customers."

In 2001, the company implemented a Teradata Warehouse to conduct customer relationship management (CRM) campaigns. "Performance was the main reason we selected Teradata," explains Vigil. "With Teradata, we reduced processing time from our original application—in which we calculated customer churn with an Oracle application—from nine days to one."

Yet EchoStar still struggled to optimize its enterprise data assets. By early 2005, the company's data warehouse had become a "data garage," says Vigil. "We had all sorts of information in the [data] warehouse, but it was in islands that were not integrated. The large growth of the company created tremendous demand in the business for information. IT spent too much time responding to similar requests, compromising our ability to perform more valuable services." Moreover, IT was unable to provide the rapid turnaround needed for effective decision support.

A goldmine of insight
Eager to use the data to EchoStar's best advantage, IT Vice President Mark Veyette spearheaded an initiative to use the data warehouse to gain a single view of the business and to create a foundation for operational initiatives. The primary objectives of the enterprise data warehouse (EDW) were to enable speed, managerial control and the ability to leverage the company's strategic data assets to drive sustainable competitive advantage. Interviews with stakeholders in the business revealed the most desired capabilities, and discussions with IT helped whittle down the options for meeting those needs.

IT created key application prototypes, which were shared with the business to show how the data warehouse could be used to better leverage enterprise information for decision support. By rallying their peers in the business units and creating an appetite for the value that an EDW could deliver, Veyette and Vigil secured the funding needed to support a common vision: the roadmap to "Project Goldmine."

EchoStar conducted data warehouse assessments, choosing Teradata again to provide the technology and expertise that would create a new EDW. Based on the Teradata Communications Logical Data Model (CLDM), work on the first integrated data model began in April 2006, and the model went live in September. To provide key decision support capabilities, EchoStar decided to take advantage of the Teradata partnership with Hyperion to deliver ad hoc managed reports and analytics.

Using solutions such as Hyperion Performance Suite for business intelligence (BI) and reporting and Hyperion Essbase for analytics in combination with the Teradata Warehouse, the first phase of Project Goldmine aimed to help the call center meet an immediate need: improving operations. "We wanted to find out why customers called when they did," says Vigil. "Were there installation or usage issues being missed in the field? Were there service problems?"

Two call center processes were launched. The first sought to determine why so many calls required handling by multiple agents. The agent transfer application revealed that agents, responding to a rule requiring that customer calls be handled within a maximum time, tended to transfer calls that would exceed that limit. The second agent would begin with the caller anew. "With this analysis, we realized that the metric was causing the problem," says Vigil. "Once we understood, we began to change the business process to a 'one-and-done' approach to help ensure customers get their problem solved in one call."

A second process examined why certain customers called repeatedly—sometimes as many as four or five times a day. In examining the data, EchoStar realized that some agents were not adequately trained to solve certain customer problems. Retraining these agents helped the company reduce the number of repeat calls and enhance customer service.

The EchoStar Data Warehouse and Analytics team helped the company implement its joint Teradata and Hyperion solution. Pictured, from left: Thomas Ryan, Jiten Kelshiker, Richard Stamp, George Vigil, Sean Thorne, Dustin Montoya, Kiran Chilukuri and Baskaran Swamiappan.

Empowering users to serve themselves
EchoStar continues developing experience using its Teradata Warehouse to support analysis and the Hyperion BI capabilities to expose the "goldmine" of insight. Today the company is using a customer churn application that leverages 31 unique data dimensions (compared with the two dimensions available with the Oracle CRM system) to better understand customer retention issues.

"Teradata provides better insight into why customers are churning," says Vigil. "Having more dimensions available in the applications allows business users to derive better insights from the data than they could before. They can use this information to create campaigns to stop the churn and increase our customer retention rate."

In addition, reporting designed to support business and legal requirements has streamlined business processes. For example, the accounting department has used the data warehouse to reduce IT processing by one full day for each close cycle. Reports needed to address legal requests can be produced within hours instead of days or weeks.

To help change the corporate culture from a focus on data gathering to an emphasis on data analysis, EchoStar is using its Teradata and Hyperion tools to drive the organization toward a new self-service model. Supported by IT and executed by the business, the company is developing a variety of self-service analysis and reporting applications to help each department access and use information. Supporting this effort, subject matter experts (SMEs) from each of the business units meet regularly to discuss their data warehouse requirements. "The business users thought in silo mode," says Vigil. "Each group wanted the data in its own way. They didn't realize that other groups wanted the same thing and that they were creating many redundancies in their requests."

By having the SMEs work together to determine group requirements, the IT group has been able to tackle tasks such as building a data glossary and defining metadata. With this consolidated information and the power of the Hyperion BI tools, the team built an optimized analytical area where users can get the information they need.

Early self-service applications provide data on equipment and subscribers. Call center data was added to the reporting applications in March 2006. Today, the subscriber table includes 22 million active and disconnected customers with 117 columns of data. The equipment tables include 20 million pieces of active, serialized equipment.

The self-service reporting capabilities were enhanced by using pre-joined data that reduces query complexity and improves performance. Using Teradata features such as data compression and join indexes, the team was able to enhance processing time by translating codes into meaningful text values and including hierarchies in tables. In addition, some fact tables and dimensions were merged into one or two large subject-specific tables. Most queries return a result in less than 10 seconds during periods when the system is not extremely busy.

The team also created an easy-to-use interface that allows users to query the database directly. "Users can better understand the information now," Vigil adds. "Now IT only has to go the last mile; the business units can build 60 percent of their simple information requests." More than 600 active users run nearly 300 jobs each day using the self-service capabilities.

For example, marketing uses the reports to measure campaign effectiveness and more efficiently apply the lessons learned from previous campaigns to improve results. The sales department draws information about customer churn, and the management team is beginning to use a new application that automatically pushes high-level corporate information—such as subscriber counts, forecasts and budget numbers—to executives' BlackBerry hand-helds each day.

The new capabilities reduced by more than 80% the number of ad hoc requests to IT from some groups and slashed the reporting backlog in some departments by as much as 60%. "In the past, IT struggled to get information out to the business," remembers Vigil. "Now business users can work in their own Teradata and Hyperion-supported 'sandbox,' getting their information without compromising system performance. It's a powerful way to empower users and give them control over the information."

Vision for new value
Although EchoStar is implementing its Project Goldmine vision incrementally because of budgetary constraints, the company never loses sight of its enterprise-wide vision for the data warehouse. In the future, the company may build applications that use the Hyperion data mining tools to deliver active enterprise intelligence—helping to spot customers likely to churn or detecting fraudulent activity and taking quick action to protect revenues or avoid loss.

For now, the value delivered by Teradata power and performance and Hyperion BI is clear. "Our Teradata and Hyperion partnership has given us the tools and technology to take us to the next level and deliver on our Project Goldmine vision," says Vigil. T

Behind the solution: EchoStar Teradata Database: Teradata Database V2R5.1 Server: 12-node 5400 Teradata Server production system Users: 500 + DBAs: 2 Data Model: Teradata Communications Logical Data Model (CLDM) Operating System: UNIX MP-RAS Storage: 12.5TB Teradata Utilities: Teradata Load Utilities 7.1, Teradata Manager, Teradata Priority Scheduler, Teradata Metadata Services, Teradata Analyst Pak - Visual Explain, Teradata Utility Pak: ODBC Driver, SQL Assistant, Administrator, BETQ and CLI Tools/Applications: Teradata Customer Relationship Management, Teradata Warehouse Miner, Teradata Value Analyzer; Hyperion Performance Suite Version 8.3, Hyperion Essbase, Hyperion Intelligence; and products from Brio, Informatica and WebMethods

Cheryl D. Krivda has written for more than 20 years about the intersection of high technology and business practices for publications and corporations around the world.

Photography by John Marmaras

Teradata Magazine-June 2007