Intelligent Manufacturing € September € 1996 € Vol. 2 € No. 9

IBM Opens a Data Warehousing Center

Computer manufacturer IBM Corp. (Armonk, N.Y.) has opened a $47 million Teraplex Complex, a scalable, "real world" center for testing live customer business intelligence systems, software and applications that turn information into insight. The complex is the first in a number of planned new IBM initiatives that will help businesses use business intelligence systems, software and services to analyze information, gain market share, and determine profitability.

These business intelligence systems, as IBM defines them, perform complex analysis and data mining on vast amounts of factual information in order to drive business insight and improve decisions. With the amount of information collected by businesses doubling every five years, finding ways to use that data in ways that make the organization more responsive to customer and market changes continues to be a growing challenge.

Business intelligence applications are viewed by many consultants as being one of the first key business applications for networked computing in the emerging knowledge-based economy. Information resulting from these applications increasingly is being shared by employees, customers, suppliers and partners, across intranets and the Internet.

The test centers that comprise the IBM Teraplex Complex will accommodate IBM's largest and most ambitious data warehouse, decision support and data mining projects, such as understanding customer relationships in order to enhance "customer intimacy"; predicting worldwide market share; and optimizing data mining on multi-terabyte data warehouses.

The complex will accommodate UNIX and non-UNIX applications, and offer IBM workstations with a combined storage capacity of 12 terabytes. The systems will also offer support on IBM database management software, a full suite of IBM data warehouse software products, and third-party software products.

"The world is moving quickly toward a digital, knowledge-based economy, where information and knowledge is a primary asset of every business," said Ben C. Barnes, general manager of IBM's Worldwide Decision Support Solutions organization. "We're working with hundreds of customers to install the latest in business intelligence systems and software -- for data warehouses, statistical analysis, multidimensional analysis, optimization analysis and data mining. The objective of these customers is to build what eventually will become 'closed loop' business intelligence systems that use information to drive mission-critical applications."

According to David Wells, a researcher with information specialists Ovum Ltd. (London, England), "Data warehousing is not just a fashionable technology." It will become increasingly popular over the next few years as more companies build small stand-alone data warehouses (known as "datamarts") rather than huge complex corporate data warehouses. "Smaller companies will also increasingly be able to afford data warehouses as tool support and integration improves," he added.

As lead author of a new Ovum report on data warehouses, Wells surveyed the rising interest worldwide in this new business intelligence process. According to the report, "Warehouse data is taken from one or more operational systems, modified and combined to make it suitable for analysis by business-oriented users."

A key point for manufacturers and other corporate users to keep in mind is that data warehouses are built, rather than purchased. "The most difficult aspect of building a data warehouse is building the mechanisms which populate the warehouse with data," Wells noted. "This can be very costly in terms of the effort involved, especially where tools are not used. Many of the tools available to reduce that cost are themselves expensive."

It has been estimated that organizations currently analyze less than 7% of in-house data. Business intelligence systems provide organizations with the capability to not only store trillions of bytes of data in a warehouse, but also use data mining tools. These tools can extract, for example, information that provides a better understanding of customer buying patterns and product preferences (market-basket analysis), enhanced customer intimacy and loyalty (customer relationship management), or the detection of fraudulent behavior (fraud and abuse management).

According to Wells, the main risks involved in the development of a data warehouse are: a project which never delivers; warehousing the wrong data; building a warehouse which is too expensive to run; failing to manage organizational change; user apathy; unclear ownership of the warehouse; inflaming existing organizational disputes; and scalability disaster.

These risks can be minimized by the use of an incremental development methodology, as well as tools that can automate the management of the data.

Ultimately, the proper use of business intelligence will help manufacturing organizations measure and manage themselves more effectively. The emerging solutions of data mining and data warehouses can certainly contribute to increasing a company's competitive edge, but keep in mind: Data warehouses are not a quick fix. They take a lot of time, money and effort, but early returns indicate that with proper implementation, their contribution to the bottom line can be considerable.

Click here to return to Table of Contents for the Intelligent Manufacturing September issue.