Case Study: Redbrick database analysis

In 1993, food giant, General Mills, decided to invest in data analysis and management tools. By 1999 their sales had overtaken...

In 1993, food giant, General Mills, decided to invest in data analysis and management tools. By 1999 their sales had overtaken those of their archrivals, Kellogg's

In 1993, General Mills decided to move its Nielsen database from an ageing Britton-Lee/Metaphor decision support system to one offering higher speed query capability. They considered moving it into a DB2 database environment on the corporation's mainframes. Management also investigated placing their Nielsen data on mid-range Unix computers to obtain higher performance for the dollars invested in hardware.

General Mills' systems management reviewed several hardware and software combinations before deciding on the best decision support solution for its marketing professionals. The goal was to find a system as fast as possible for pulling information out of large databases to increase the speed with which marketing management could make decisions. Management wanted the highest processor price/performance and software offering the quickest database loading and query times.

To choose such a system, General Mills investigated several database alternatives. Management concluded that Red Brick Warehouse was the solution best optimised for decision support environments. Red Brick Warehouse, running on a Unix processor, offered higher performance at a lower cost than a DB2/mainframe choice. The Red Brick Systems - focused on high-speed data loading and retrieval - shone during General Mills' comparative performance evaluations. The software was twice as fast as any alternative compared against it. General Mills chose to acquire Red Brick Warehouse hosted on a Hewlett-Packard 9000 Model G70 RISC server.

Because of Hewlett-Packard's track record at General Mills, it was easy for the packaged foods giant to select the systems new hardware component. Hewlett-Packard has been the strategic mid-tier computer vendor at General Mills since 1977. Mike Meinz, Principal technical consultant at General Mills said, "HP systems work well. They are really reliable and easy to install. HP provides good support and has good people to work with."

In reviewing the software alternatives, support was also an important consideration. "Red Brick delivered on its promises," recalls Mike Rueswald, manager of Sales and Marketing Decision Support Systems. "When we were installing Red Brick, we also brought in a new decision support tool that had never worked against Red Brick. The tool had requirements Red Brick did not have in place then. Red Brick and this vendor worked together to make the data access tool and the Red Brick technology work together. This was completed under a tight schedule."

Red Brick Warehouse has reliably supported General Mills' needs using minimal systems support resources. "We put it up and it worked well. It's been a low maintenance application for us," Rueswald observes.

Today, General Mills uses Red Brick Warehouse to sift through its Nielsen data. Marketing staff use the system to uncover General Mills' market share for various products in different markets. Red Brick Warehouse manipulates data on the buying activities of millions of consumers. It tracks the impact of sales and marketing campaigns in different markets on consumer purchases.

By combining shipment data with Nielsen sales history, General Mills' account representatives obtain a better view of inventory build-up in their accounts. Marketing managers use the data warehouse to see which accounts are moving closer to just-in-time inventory replenishment. With these insights, marketing can devise programs to increase General Mills' sales, and increase turns and profitability for its customers.

An important factor in reaching insight is having enough system speed to support repetitive queries. The high-performance Red Brick Warehouse supports insight-seeking query activity by changing the dynamics of how staff query databases. According to Rueswald, "When the system moves fast enough to approximate a person's thinking and response time, he or she starts interacting with the data. The more interactive a person can become with data, the better analysis you get."

For many years, General Mills has expected its marketing people to analyse their businesses and be facile with relevant tools and information. "You want to empower marketing people with good query tools. Although you can ask an MIS person to create a report or get data, a marketing or sales professional is likely to see relationships that an MIS professional wouldn't", explains Rueswald. He sees opportunities for the corporation to digest and distribute to its salespeople high-value insights derived from the data warehouse and elsewhere. "The key is to deliver the intellectual capacity of the company to the sales representative in the field."

Enhancing the creativity of marketing staff and delivering more knowledge to the field is merely the beginning of General Mills' journey with data warehouse technology. " End users will be pulling up commercials on what we've done on Cheerios and looking at all the Cheerios commercials at their workstation rather than going to screening rooms. Maybe they will deliver commercials on CD-ROM or other mechanisms. Users will employ a host of multimedia technology in making decisions. We will engage the challenge of harnessing the volumes of data within our corporation. Making sense of all this information presents the business opportunity of the future."

Compiled by Will Garside from

Read more on Business applications