When your company is creating a data warehouse, it is important to create a logical design for your database. This is important for several reasons. First, you want to be able to communicate openly about what will be in the data warehouse. The user community will want to know which elements are going to be included. If there are multiple implementations, they will want to know which elements will be available for each implementation. Next, a logical data model facilitates the creation of a low maintenance physical database design that will require only minor modifications after implementation. Assuming your business does not change dramatically, the physical implementation based on the logical model should last for many years without modification. Finally, the data model is an excellent way to contain the scope of a data warehouse project. To contain the scope of the data warehouse using the logical data model, you would select subject areas that are included in the first and later implementations. Then you would build a timeline based on the creation of those subject areas.

The following data model has very clear subject areas: time hierarchy, operational hierarchy, merchandise hierarchy, point-of-sale (POS), inventory, planned inventory, markdown markup, merchandise summaries, operational summaries, and orders. To prioritize the implementation, the first implementation could be the POS subject area plus the supporting time hierarchy, operational hierarchy, and merchandise hierarchy. The second data warehouse implementation could be another subject area, like inventory. Of course, you will want to implement each subject area...