The Data-Warehouse Advantage

Following the 1998 merger of NationsBank and Bank of America, the new Bank of America Corp. found itself with two data warehouses, each based on different technologies and containing separate customer lists. As the banks' operations were consolidated, the need to integrate the data warehouses grew. While links were built between the two data warehouses, the performance of the system degraded as its cost and complexity increased. Some departments were even building their own data marts to work around the problem.

It was clear the company needed a consolidated data warehouse. That project, initiated in November 2001 and completed six months later, won the Data Warehousing Institute's best-practice award for data-warehouse integration.

How did the company do it? Data-warehouse managers looking for magic solutions or breakthrough techniques won't find them here. "It's just doing the basics well," says Cheryl Keller, the bank's CIO for corporate staff support groups. "It was good blocking and tackling."

Support from top leadership, which designated the data warehouse as one of the bank's top three IT priorities, was key, Keller says. Pre-project planning, including researching user requirements, was extensive. And the development team was given clear goals for the project and the resources needed to do the job.

Keller's team assembled the system using hardware and data-warehouse software from Teradata (the old data warehouses used Teradata and IBM products) and data-management and transformation tools developed in-house. The system has 25 terabytes of raw data about its customers--Bank of America serves more than 28 million households and 2 million businesses--pulled from more than 190 operational systems.

Today, several thousand analysts throughout the company have direct access to the data warehouse, while employees in call centers, bank branches, and other customer-facing operations use applications that tap into it. The system is more efficient: An E-commerce-division chore that took 22 hours on the old system now takes 15 minutes, Keller says. The bank has realized significant operational savings with the new data warehouse, and several CRM and risk-management projects that will use the system are expected to add to the bank's revenue stream.

At International Truck and Engine Corp., closing the books each month was taking too long--as long as two weeks using manually intensive processes. The truck, diesel-engine, and truck-parts maker already had a data warehouse assembled in 1996. The challenge was finding a way to leverage that data to provide financial managers and other executives with quick access to business-performance information.

In late 2001, the company built what it called its Key Business Indicators analytic information portal, which today gives 450 top executives, managers, analysts, and other knowledge workers customized views of the key performance metrics for their jobs. The system won the Data Warehousing Institute's award for business-performance management.

International's data warehouse uses an IBM Informix database running on Unix-based Hewlett-Packard servers. Ascential Software Corp.'s DataStage software handles the data collection and transformation chores. An Essbase online analytical processing engine from Hyperion Solutions Corp. running on a Windows 2000 server provides access to the warehouse.

The Key Business Indicators portal was one of several initiatives International undertook to speed up the financial-closing process. And it's working, providing an efficiency gain of 10% to 12% in the monthly close process, says Jim Rappé, the company's enterprise data-warehouse group leader. Today the portal has been expanded beyond finance to give managers and employees in International's corporate, truck, engine, and service-parts operations the performance metrics they need to make day-to-day decisions and detect trends.

Among the more than 100 performance metrics the system provides are production volume, factory shipments, reliability and quality, and safety. "These are metrics we want to have in front of management all the time," Rappé says.

International laid the foundation for the system in 2000 by building its own portal framework. Portal technology at the time was in its infancy, Rappé says. Software from Alphablox Corp. was used to build the front-end analysis applications that give users the data. A Microsoft SQL Server database stores information about users' portal and report preferences. Before the portal was built, all this critical performance data was communicated via paper reports.

By speeding up decision making and lowering operating costs, the portal is critical to International because of the cost-competitive nature of the truck-manufacturing industry. "We try to project to customers that we have the lowest total cost of ownership," Rappé says. "It's all about competitive advantage."

Data-warehouse systems are still complex and costly. But as International's Rappé says, it's all about competitive advantage. All the Data Warehousing Institute Award winners are proving that data warehouses, when designed, built, and operated the right way, can deliver that advantage.

Continue on to: The Data-Warehouse Advantage, Part II