Most people assume that the starting point for any business intelligence (BI) project is a data warehouse. In reality, while data warehouses are important for many types of analytical systems, they aren't always necessary. Building a data warehouse can dramatically increase the cost of a BI project. It can also reduce the value of the information by taking timely operational data and making it dated or even irrelevant. Many BI projects can realize benefits from alternate data integration scenarios.

Data warehouses themselves are not the problem. The problem arises when a data warehouse is viewed as a solution to all BI deployments, or there is an expectation that simply building a data warehouse will drive users to information. Data warehouses should not be implemented without a clear understanding of the business challenges they will solve. Before building a data warehouse, you should also carefully research potential data-access architectures to make sure you have devised the best method for connecting your BI tools with your data.

The intended audience for this paper is project, business, and IT managers who have responsibility for BI activities. This paper describes data warehouses along with many other options for placing relevant, timely information in the hands of business users. While a data warehouse is a good solution in some instances, many BI applications are better served with integration and portal technologies that simply pull data into reports on an as-needed basis.