The construction of a data warehouse for a large enterprise can be a lengthy task. To
exploit the benefits as soon as possible, the data warehouse is often built iteratively,
one subject area at a time. As subject areas are added to the data warehouse, they may
be used to load data marts (see Figure 16.1). A data mart is a smaller “departmental
warehouse” focused on one subject area, often containing more summarized and less
detailed data than the data warehouse. For end users who perform their analysis within
one subject area, a data mart provides a simpler model adapted to their needs, and the
smaller data volume often leads to better performance.