DW, THE CATALYST FOR CHANGE
Three years ago, Dow Chemical Co. underwent a radical metamorphosis, reorganizing business units and shutting down or selling nonstrategic businesses and underperforming assets. A linchpin in this global overhaul was a new corporate data warehouse.
"What began as an organizational change quickly became an opportunity to rethink the way we were providing business information," explains Mike Costa, client manager of IT at Dow, in Midland, Mich.
The warehouse project, which has yielded a 43 percent gain in reporting productivity since 1993 as well as 70 percent faster report generation, recently earned Dow the top award for ROI (return on investment) in the Technology Managers Forum Best Practices Awards, sponsored by PC Week. The warehouse was an award winner not only because it shows a solid ROI but because it carries out closely the business mission of Dow's reorganization, says TMF contest judge Anne Kane, senior systems manager at Boston-based BankBoston Corp. Kane praised the way Dow's IT group worked with the business units. "Partnering with your business is really important. And [IT] had a lot of checkpoints to meet with the business to talk about where they were -- they had good project management," she explains.
Analysts say a data warehouse is a natural tool to help companies of all kinds restructure and re-engineer business processes. "We see this across all kinds of different industries," notes Douglas Hackney, president of the Enterprise Group Ltd., a data warehousing consultancy in Hudson, Wis. "A warehouse lets you look at an organization from any number of points of view. "
At Dow -- which provides materials such as polyethylene and emulsion polymers to manufacturers for use in their products--the corporate reorganization replaced a number of geographical units with 15 worldwide operating units. Two of the new corporate units are, for example, Chemicals and Metals and Specialty Chemicals.
Also overhauled was the way that performance data was reported. Replacing traditional profit-and-loss statements was a new way of looking at financial performance, called VBM (value-based measurement). The tool for implementing VBM was a new VBIS (value-based information system) that relies heavily on the data warehouse. The VBIS provides information to managers about total shareholder return on products and services, rather than simply earnings per share.
"Our objective was to generate an income statement that takes into account recovering investment capital," Costa says. To do this, company managers needed to drill down and see exactly what products are generating profits and costs. "At the end of the day, our goal was to provide mechanisms for measuring profit and growth specifically," he explains.
Before the advent of data warehousing and OLAP (online analytical processing) technology, it was impossible to view this data quickly and cost-effectively. And because Dow was organized by geographic region, its information systems were also limited geographically and could not generate reports on worldwide performance.
Warehouse operations
The VBIS data warehouse is built from transactional data from Dow's enterprise SAP AG R/2 installation, says Ron Johnson, IS program manager for finance. The R/2 system captures and cleanses the data and feeds it to an Oracle Corp. database housed on two Digital Equipment Corp. Alpha 8400 servers running OpenVMS. One terabyte of data, consisting of 30 million records, is added to the warehouse every three months.
From there, Johnson says, Cognos Corp.'s PowerPlay software running on 10 Windows NT application servers transforms the data into OLAP cubes. The cubes are then stored on another 20 Windows NT file servers. Users running PowerPlay on Windows 95 desktop PCs access those file servers for reports.
A number of teams worked on the project, according to Johnson, each with its own checkpoints. In turn, the overall project had checkpoints as well. Johnson says the teams could meet checkpoints because of aggressive problem resolution as obstacles were encountered. "We didn't let an issue sit 24 hours before we addressed it; [we] had action plans in place," he explains.
Looking ahead, Costa says the next step will be to re-engineer Dow's forecasting and planning processes. "We are going to try to use the same approach we used for VBIS, using data mining tools to help [the] forecasting process," he says.
These efforts are all aimed at carrying out the mission of giving managers the best tools to make strategic decisions. Says Costa: "Our goal is to provide the businesses as much information as possible."