Winners: Best Practices Awards 2004

Business Intelligence Journal, Fall 2004

Innovators in Business Intelligence and Data Warehousing

Each year The Data Warehousing Institute (TDWI) asks practitioners to share their unique and innovative business intelligence and data warehousing solutions during the annual Best Practices Awards competition. This year's competition recognized 12 industry leaders, illustrating a wide range of best practices that business intelligence and data warehousing professionals can learn from and implement within their own environments.

Any company that has designed and/or implemented a data warehouse or business intelligence solution to enhance corporate decision making is eligible to compete in the Best Practices Awards program. In addition, solutions sponsors may nominate their most compelling customer's story for consideration for a Best Practices Award. This year we received 85 submissions from companies representing diverse industries.

Summaries for 11 of the 12 winners are featured here. L'Oreal Paris Consumer Products Division also won in the Data Warehousing on a Limited Budget category, sponsored by Intelligent Solutions, Inc.

For additional information on these award-winning implementations, please visit http://www.dw-institute.com/display.asp?id=7145.

Category: Predictive Analytics

Winner: Absa Bank

Absa is a leading bank in South Africa formed in 1992 through the amalgamation of four major banking groups. Absa offers a complete range of banking, insurance, financial, and property products and services to some 5.9 million customers in the personal, commercial, and corporate market segments. Absa's vision is to position itself in targeted market segments. The core-operating model is based on segmenting the client base.

The Absa data warehouse (DW) has a high degree of integration across products, customers, and business units providing a single view of the customer. Therefore, Absa's DW unit can assist business in its main focus areas: customer management, product management, channel management, operational management, and capital risk management.

The current ABSA enterprise data warehouse (EDW) is based on IBM RS/6000 RISC hardware architecture, Oracle Enterprise RDBMS and AIX operating system software. The back-end storage architecture for the DW complex has recently been moved from standalone IBM ESS (Shark) storage islands to a flexible SAN architecture.

The main data sources reside in the production mainframe areas. COBOL programs extract data in text file format at various intervals. The text files are placed in a structured landing area on a dedicated staging server. An Oracle tool (SQL*Loader) was used to load these files into a staging database. The entire process of loading flat files with SQL*Loader to the EDW is controlled with an eventbased system built using Oracle Workflow. As soon as the EDW is populated with the latest data, events trigger further transformation processes, this time in the EDW, that populate tables in several enterprise data marts.

The staging database is divided into a front end and a back end. The "raw" data is loaded into the front-end area. Business rules, transformations, normalization, and enriching are then applied to the data through a set of sophisticated PL*SQL programs, the results are loaded into the back-end staging system.

All production database instances are physically separated from each other, each residing on a dedicated host. This ensures that the ETL process does not impact any user and that a user query against business data has the least possible effect on the ETL process.

The entire DW and statistical modeling team is part of a specialized business unit that focuses specifically on "information management," i.e., the enabling of information and knowledge-based strategy formulation and decision making. The information management business unit is a distinct entity positioned between the business and information technology divisions. It thus has a distinct focus on leveraging information for improved performance. The effectiveness of the division is enabled through the data warehouse, one of the largest in South Africa.

Category: Radical Data Warehousing/Business Intelligence

Winner: Acxioni Corporation

Acxiom uses data, technology, and services to create customer and information management solutions to help companies build better customer relationships and market to prospects. Acxiom is headquartered in Little Rock, AR, with offices in 10 countries and annual revenues of more than $1 billion.

To provide these information management solutions, Acxiom must process more data than ever before. The huge computer servers needed to store and integrate massive amounts of data are extremely expensive and slower than client-centric businesses demand. Acxiom needed a system that would allow the company to update its multi-billion record databases faster, more economically, and more reliably.

Acxiom began investigating the potential of grid computing: the sharing of high-end PCs across a network so they function as one large, faster, cheaper, more powerful supercomputer.