Why build a Data Warehouse?

There are three major reasons for building a data warehouse.

1 - To achieve greater performance by separating the Decision Support and OLAP activities from the day to day OLTP activities. For instance, the star schema design with integer surrogate (synthetic) keys is faster than a third normal form relational structure.
2 - To integrate data from different source systems by standardizing and transforming the data to common dimensions and measures. The banking industry, in particular, has seen considerable consolidation over the past decade.
3 - To allow greater user interaction (self service) by providing the data in structures and terminology understandable by the user community. This includes the ability to access the data using off the shelf query, statistical, predictive, and other OLAP softwares depending on the users' skills and responsibilities.