Peter Drucker saw the potential for business intelligence when he wrote in Post-Capitalist Society (1993, HarperBusiness): "The basic economic resource is no longer capital, nor natural resources, nor labor. It is and will be knowledge. Value is now created by productivity and innovation, both applications of knowledge at work." In the early 1980s data warehousing was conceived as a feasible way of transforming a jumble of operational data into a tool for supporting intelligent decision-making. That was a time when a computer-initiated avalanche was beginning to inundate businesses with a never-ending flow of numbers on sales transactions, inventory records, and statistics on production and quality. Bill Inmon, journalist turned information-technology (IT) consultant, observed that these silos of information could be organized into a corporate asset that he called a data warehouse. (Today, his business base is Pine Cone Systems Inc., Englewood, Colo., a company that also offers software tools for data-warehouse administration and management.) Since then, data warehousing has become a mainstream technology, notes Wayne Eckerson, vice president of technology services for the Data Warehousing Institute, Gaithersburg, Md. A market study from Frost & Sullivan, Mountain View, Calif., says that the U.S. data-warehousing-software market accounted for $2.1 billion in revenues in 1997, up 23.7% from the previous year. Hundreds of software vendors are offering compelling solutions that make having a data-warehouse infrastructure a requirement for doing business, adds Eckerson. "First," Inmon says, "a data warehouse causes data to be integrated across the corporation, so instead of con-fronting lots of applications in the search for business intelligence, you are able to focus analytic tools and techniques on an integrated repository of business-transaction data. "The second major characteristic of data warehousing is its historical nature. . . . Armed with historical information, we are enabled to predict the future." Third, the data warehouse offers a combination of both detailed and summarized data. The early adopters of this concept were in consumer goods, companies with thousands of stockkeeping units (SKUs) that were trying to access and analyze summary and detail data. They were looking for an enabling information infrastructure for discerning trends by product, by region, and by retailer, notes Dallas-based James D. Murphy, vice president, National Accounts Div., NCR Corp. "By using a data warehouse, applying analytical tools and techniques becomes a far more doable process, especially in terms of achieving valid results quickly." "For retailers the key to business performance is to establish data warehouses to improve and manage customer relationships," adds Murphy. For instance, an NCR data warehouse gives Sears Roebuck & Co. the ability to monitor sales by store and is enabling Sears to fulfill its strategic goal of creating a sharp local market focus. Others striving for market leadership were quick to discover that data warehouses can turn information into advantage in virtually any industry. For example at Robert Bosch Corp., a data warehouse was implemented to facilitate access to quality-control data at its Anderson, S.C., facility. With daily production of 5,000 antilock brake control units, the plant needed to improve access and analysis of the 450 quality tests performed on each unit. As many as 1.5 million data values are collected daily. Unfortunately, until recently, most of the collected data was not available to the technical staff. The solution proposed by senior test engineer K.C. Podd is basically a quality-control data warehouse that provides plant personnel with the ability to quickly retrieve and graphically see test results for all products being manufactured. In addition to tracking control parameters in the manufacturing process, it also highlights where potential cost savings might be gained, says Podd. On ROI, he notes that one good "save" can pay for the system, which is exactly what happened. "With high quality standards, the scrapping of parts is a large expense, so scrap reduction is one of our major goals," Podd says. "A few months ago we discovered an issue localized with our previous system to one pallet of units. We knew the issue occurred around a certain time, but nothing more. Without our new system we would have scrapped the entire pallet. Instead, we were able to identify the affected parts and save about 80% of the product on the pallet." Using an Oracle database, the Bosch solution was built using data collection, analysis, and reporting tools from the SAS Institute Inc., Cary, N.C. Web-enabled solutions are beginning to amplify the benefits of data warehouses. One example is Lockheed Martin Astronautics Space Launch Systems (SLS), Denver, where the focus is on achieving manufacturing-schedule dependability and project management for building Titan Rockets. The SAS-based Warehouse On the Web (WOW) extracts and transforms data from relational database management systems and other proprietary applications and data stores, and presents business intelligence in a flexible and user-friendly environment, says Art Rerecich, project group leader for Decision Support Systems. "Prior to WOW, it took days and weeks for managers to get information they needed, and then the information was only at summary level," Rerecich says. Web enablement has proven to be a cost-effective distribution strategy. "Managers are able to hit the Web, pull up any vehicle they want to look at, and see what work remains to be completed at any given level of the process." One of the driving forces for data warehouses is the rapid emergence of analytic applications. These are the business-intelligence and knowledge-management solutions that can reach into the data warehouse to gain competitive advantage in one or more of the following ways: by increasing the speed and flexibility of business analysis; by enabling the improvement or reinvention of business processes; and by understanding and managing customer behavior. The selection process for choosing analytic tools is a key strategic step after the implementation of a data warehouse, advises Stephen Cole, assistant director, research and development, American Century Investments, Kansas City. "Not only does it translate the data warehouse's capability into performance, it determines the return on investment." American Century, a diversified provider of financial services, worked with Exchange Applications Inc., Boston, to develop and implement a data warehouse running an Oracle database in a Hewlett-Packard environment. The company had extensive and detailed information about its more than 2 million customers yet it had no way to use the information for decision-making purposes." Cole believes that the decision to implement a data warehouse was only the beginning of his business-intelligence solution. Part two was devising a strategy on how to approach the search for analytic tools for use in a data-warehouse environment. "Begin and stay with the presumption that the search is about a business solution, not about a technology solution," he advises. "We look for tools that put the power of analysis in the hands of the person with business savvy and business needs." He also says the search should involve more than finding flexible tools that meet the business needs with minimum support requirements. "Consider the vendor. Does he listen to you? A lot of vendors try to change your business problem into their business perspective." Cole also urges benchmarking. "Find out how your competitors and IT professionals in other industries are solving their business-intelligence needs. Don't forget to study failures as well as successes." He says the journey to business intelligence is about constant change. "You make the investment to instigate change by obtaining knowledge you never had about your business before and the infrastructure that you build must itself be able to adapt to change. It is never over. Make sure that you have a process for an annual review of your business intelligence infrastructure and be prepared to reappraise your strategy as vendors develop new capability." For companies choosing to follow a new enterprise resource planning (ERP) implementation with a data warehouse, recent signs of converging interests may complicate the de-cision-making process, notes Henry Morris, director for data-warehousing research, International Data Corp., Framingham, Mass. He is re-ferring to the announcement by SAP AG, the leading ERP provider of its Business Information Warehouse at its annual Sapphire Users Group meeting in Los Angeles last September. SAP seems to be crossing the market divide between online transaction processing (OLTP) and online analytical processing. In October PeopleSoft Inc. announced the expansion of its analytic applications with Performance Measurement, a solution intended to help people make better decisions. In November a collaboration with KPMG Peat Marwick LLP revealed the enterprise-data-warehouse part of that strategy. Tom Patterson, product manager, describes the concept as an integrated management system where "you take all of your information out of your ERP system or wherever it resides in your organization and put it into the right analytic environment." In November Computer Associates International Inc.'s Acacia Technologies Div., Islandia, N.Y., a leader in manufacturing and distribution enterprise software, announced the availability of its new business-intelligence product branded ClearView. "There's a tremendous amount of valuable business information hidden within the voluminous amounts of raw data captured by companies each day," says Ken Ramoutar, vice president of worldwide marketing. "ClearView will help turn that raw data into meaningful, decision-support information that managers can use." Features include predefined analytical views and measurements that will augment the management information and analysis reports already available in PRMS and KBM (its ERP offerings). Also in November, American Software Inc., Atlanta, introduced Intelliprise, an ERP/business intelligence solution for its middle-market constituents. Karin Bursa, vice president-marketing, describes Intelliprise as providing out-of-the-box data marts with decision-support tools that have the flexibility to quickly respond to changing business environments. What differentiates Intelliprise, says Bursa, is a data-mart approach that provides a way of evaluating a business based on a preset grouping of key performance indicators. Another November announcement from Macola Software Inc., Marion, Ohio, expressed the need to fulfill business intelligence needs in a different way. It plans to bundle Seagate Crystal Reports 7 with its Progressive Series 7.5 product line. Crystal Reports 7 is a new release from Macola partner, Seagate Software, a subsidiary of Seagate Technology Inc. "Macola's customers have come to rely on the abilities of Seagate Crystal Reports over the years," says Macola President and CEO Bruce A. Hellinger. "By matching this version with our Progression Series 7.5 we can provide a very comprehensive solution for small to mid-sized enterprises. At Symix Systems Inc., Columbus, the business-partner approach is also used to bring business-intelligence capability to its ERP customers. "In 1996 we started supplying our midmarket customers with a business-intelligence option supplied by Cognos," explains Krista Endsley, product manager. The Symix approach is to provide 10 data-mart models that can be run against a customer's OLTP database. As ERP vendors seek to expand their market footprint with business-intelligence tools such as data warehouses and data marts, users should be asking more questions, says Joshua Greenbaum, principal, Enterprise Applications Consulting, Berkeley, Calif. "For example, a vendor's business-intelligence offerings may fit very well with its own ERP system, but how will it deal with data from other sources?" IDC's Morris sees the data-warehouse interest of enterprise-systems vendors as a force that will shape their future ERP offerings. "In other words, some people will say: 'How come SAP is not collecting this kind of information that we need for our analysis?' And if SAP didn't design their application with the idea of supporting those data warehouse needs, then changes will happen. Even more significant will be the fundamental change that will emerge when users start asking: 'Now that we've developed this information in our data warehouse and analyzed it, how can we incorporate the results of that analysis better inside the operational system?' That will lead to a faster and broader trend to closed-loop systems."