Banks: Big Money in Data Warehouses
"[Direct marketing] is the fastest growing area of marketing expenditures in the bank," said Randy Grossman, senior vice president and director of Fleet's customer data management and analysis division. "We want to be able to improve the efficiency of our target marketing. Like most banks, Fleet has a customer base of which 40 percent are unprofitable, and half of the customer product sales we make every day never become profitable."
Fleet's data warehouse, which cost $37.7 million, maintains information for 36 months on the bank's more than 6 million retail and small business customers, Grossman said during a luncheon here sponsored by the Direct Marketing Club of New York. Fleet, Boston, plans to add data on another 7 million customers coming from recent acquisitions by year's end and information on larger companies in 1999.
The data warehouse captures information from 33 sources, including deposit accounts, mortgage products, commercial loans and ATMs and houses demographic information bought from outside vendors. The data passes through a server that cleans and standardizes the information before it reaches the warehouse, which uses technology from Informix Corp., Menlo Park, CA. Reports are pumped out weekly into data marts, smaller subsets of the warehouse that let employees perform analysis and segmentation.
Fleet intends to connect its 1,200 branches to the data warehouse by the end of the year, allowing branch managers to view the entire relationship between the bank and each of its customers.
"If a branch manager is about to pick out a customer and make a phone call, that manager knows if that customer received a mailing last week or is on the list to be called by the call center next week," Grossman said. "Vice-versa, we can also pluck out a customer from the calling list and see if that person was called by the branch manager and exclude them from the list if the manager did call."
First Union's data warehouse, which also runs on database technology from Informix, works much like Fleet's, but maintains data on its 16 million customers for 24 months and receives information from 20 sources. First Union, Charlotte, NC, which declined to disclose the cost of the project, implemented the data warehouse for its retail consumer banking and plans to add the commercial side of the business.
"To maintain our business advantage, we need to differentiate ourselves from other institutions through service, and service comes from knowing your customers," said Naras Eechambadi, senior vice president of First Union's Knowledge-Based Marketing, an organization formed in 1995 to develop the company's data warehouse.
The two companies have had to hire many new employees -- ranging from technicians, analysts and database marketers -- to keep the data warehouses running smoothly. Both had been relying on Marketing Customer Information Files (MCIFs) to keep tabs on their customers, but noted the need for more advanced technology as competition for market share intensifies.
"Our MCIF provided us with snapshots of data every month -- and it was very hard to go back and tie things up," Eechambadi said. "Now the data will be much more current. With the MCIF process, it takes us almost 30 to 40 days to turn around information before we have access to it for analysis. The data warehouse gives it to us within a week."
Before the data warehouse was running, Fleet conducted an analysis of 30,000 of its most profitable clients. The analysis took 10 weeks. Grossman's staff spent eight of those weeks accessing the data but only two weeks analyzing the information.
"We hope to turn that ratio around -- to get our hands on the data in a very short time and spend most of the time analyzing the information," he said.
Under its old system, Fleet didn't even have the ability to determine how much it was charging customers in fees. The prior technology tracked accounts and assigned fees, but the system didn't record the amounts after the billing cycle had passed.
"MCIFs are no longer enough," Grossman said. "They don't have the ability to analyze and understand customer data."