Ready-Made Data Warehouse Use on the Rise                          11.11.97
ACTION ITEMS                                                          D S *

San Diego, CA -- As Information Week reported, rather than designing a customized system, ReliaStar Life Insurance Co. chose IBM's fraud-management package to establish a data warehouse for detecting false insurance claims. ReliStar estimates it saved half a million dollars and six months of development time. "The IBM package contained the industry experience we needed, while providing a financially viable option to in-house development," says Joel Portice, director of the Minneapolis company's Integrity Plus Services division.

The number of companies opting for ready-made data warehouses is on the rise. A reason for the rise may be the tendency of these "out of the box" warehouses to be targeted at specific industry needs. These packages can save both time and money. For some small and midsize companies, they are the only affordable option. These packaged data warehouses however, are not as scalable or flexible as custom solutions.

The most recent round of vertical data warehouse products are tightly focused on applications for addressing specific business problems common to a particular industry, such as customer retention for phone companies or fraud detection for credit-card issuers, says Richard Rist, VP of research and education at the Data Warehousing Institute in Gaithersburg, MD. The products target data-intensive industries such as insurance, telecommunications, health care, manufacturing, and energy.

This second generation of vertically oriented solutions contrasts with the packaged "data models" of a few years ago, which tried to address all the warehousing needs of companies in a given industry. These flopped, Rist says, because they were too generic.

The experience of MPM Corp., a Franklin, MA, manufacturer of stencil printers for electronic assembly, is indicative of just how effective the new round of products is. MPM needed a data warehouse to analyze production patterns to improve inventory and pricing practices. The company didn't want the risk involved in developing a custom data warehouse, says Kevin Brassard, a project manager for the 450-employee company. "As a small company, we could not afford to take that kind of risk," he says.

Instead, MPM bought QAD Inc.'s Enterprise Data Warehouse for manufacturers, which employs cross-plant and cross-product line analysis to optimize processes. The total cost of the project: $200,000 or about half what custom development would have cost, Brassard says. Another bonus for MPM: The warehouse only took two months to develop.

QAD, a Carpinteria, Calif., supplier of integrated business software and services, developed Enterprise Data Warehouse in conjunction with Data General Corp. and Oracle. Enterprise Data Warehouse is designed to work with QAD's enterprise resource planning software, MFG/Pro, also designed specifically for the manufacturing industry.

Both products helped MPM reduce costs by furnishing detailed information on inventory levels for more accurate demand forecasting. They also let MPM analyze product pricing to set margins and discounts. MPM expects to recoup its investment within a year, Brassard says. QAD's Enterprise Data Warehouse is also extensible, so in the future MPM can start drawing data from subsidiaries in other parts of the world, Brassard adds.

ReliaStar had a similar experience with IBM's Fraud and Abuse Management System (FAMS). In FAMS's first full year, ReliaStar's Integrity Plus Services division cut insurance payouts in half, identifying more false claims.

Developed with Cigna Corp. and Pennsylvania Blue Shield, FAMS came ready-made to meet the Integrity Plus division's needs. FAMS examines insurance claims based on geography and the specialty of the health-care provider submitting the claim. It uses IBM's fuzzy logic technology to score the claims, indicating the likelihood that a provider is submitting a false claim.

The rapid development time of these packaged solutions also helps sell data warehousing to company management, the Data Warehousing Institute's Rist says. Custom warehouses can take years to build, but IS departments can set up vertical off-the-shelf packages in just a few months. That, says Rist, can make the difference between success or failure. "If a data warehouse project drags on too long, you lose support. There's always someone else fighting for the dollars."

Rist, also commented that the ready-to-go vertical data warehouse offerings are just the start. "In a year or so application-specific data warehousing solutions will really take off," he says. "Right now, however, the market is still immature."

Among the other vendors providing these packaged solutions are major companies like Hewlett-Packard, Informix, Oracle, Sun Microsystems, and Sybase. Niche players provide specialized products as well. For example, Magnify Inc. of Chicago has a handful of products, including Pattern: Profit, which provides the financial industry with a tool for optimizing pricing models for complicated financial instruments. In addition, a number of the bigger companies are partnering with smaller ones that have data warehousing experience in specific industries.