Panelists at an executive summit in New York Tuesday said they have used enterprise-level analysis of data with BI tools when deciding on strategic initiatives such as deciding where to build new manufacturing plants, how to bolster customer loyalty or how to dramatically increase market share. But doing so successfully has meant moving analytics from scattered pockets throughout a company to embedding it in front-line business processes, executives said at the Competing on Analytics symposium. The event was sponsored by Harvard Business School Publishing, Intel Corp. and SAS Institute.
Keith Coulter, managing director of consumer cards and loans at London-based Barclays PLC, said the use of analytics was core to a five-year plan launched in 2000 to reverse a decline in revenue by winning new customers against emerging competitors MBNA Corp. and Capital One Financial Corp. in the U.K. Barclays infused analytics into its supply chain systems as well as the systems used by call center representatives to interact with customers. The result: Barclays acquired 1.5 million accounts through direct marketing from 2003 to 2004, three times the number of new accounts added between 2001 and 2003, he said.
This year, the company plans to combine its unsecured loans business with its second mortgage business and to integrate analytics into those combined systems, Coulter said.
The initiative would not have been successful without a philosophy that relies on analytics, he said, adding that most employees in marketing now are programmers and analysts. "The people who run the business ... this is the way they are expected to work," Coulter said. "This is now embedded in our business, embedded in how we work and how we compete."
Thomas Davenport, a professor of IT and management at Babson College, has been surveying companies that use analytics to compete over the past year. He said analytics can help companies optimize key business processes such as setting prices for products and identifying most valuable customers. But companies can't optimize these core processes without taking an large-scale approach, he said.