•  
  •  
 

Abstract

Companies can build a data warehouse using a top-down or a bottom-up approach, and each has its advantages and disadvantages. With the top-down approach, a project team creates an enterprise data warehouse that combines data from across the organization, and end-user applications are developed after the warehouse is in place. This strategy is likely to result in a scaleable data warehouse, but like most large IT projects, it is time consuming, expensive, and may fail to deliver benefits within a reasonable timeframe. With the bottom-up approach, a project team begins by creating a data mart that has a limited set of data sources and that meets very specific user requirements. After the data mart is complete, subsequent marts are developed, and they are conformed to data structures and processes that are already in place. The data marts are incrementally architected into an enterprise data warehouse that meets the needs of users across the organization. The appeal of the data mart strategy is that a mart can be built quickly, at relatively little cost and risk, while providing a proof of concept for data warehousing. The risk is that the initial data mart will not scale into an enterprise data warehouse, and what has been built will have to be scrapped and redone. This article provides a case study of Sherwin-Williams' successful use of the bottom-up, data mart strategy. It provides background information on Sherwin-Williams, the data warehousing project, the benefits being realized from the warehouse, and the lessons learned. The case is a "textbook example" of how to successfully execute a data mart strategy. Video clips of interviews with key individuals at Sherwin-Williams help bring the case alive.

Share

COinS