Data Warehousing and Business Intelligence for e-Commerce

Most of the major IT occurrences of the 1990s that we discussed in Chapter 1 data warehousing, CRM, and ERP in particular have been, as the saying goes, a "mixed blessing." On the positive side, the ever-increasing availability of packaged software has brought about dramatic increases in productivity as compared to similar environments from the "prepackaged" era that were custom developed.
The downside, however, is that the availability of and reliance upon those packages has left many IT practitioners without adequate knowledge of the many pieces that must fit together to build a real-world, sustainable computer system. Consider the typical data warehousing "architecture diagram": A handful of rectangles are placed on the left side of the piece of paper and labeled "source applications." They are then connected via lines with a single arrow at the right to another larger rectangle that might be labeled "operational data store (ODS)" or "staging area." That ODS (or equivalent) rectangle is then connected to another equivalently sized rectangle labeled "data warehouse," which in turn has a handful of lines with arrows on the right side connected to a set of smaller rectangles, each labeled some variant of "Department XYZ Data Mart." Each data mart will have either (1) a stick person, (2) a better representation of a person courtesy of the clip art from a graphics package, or (3) a graphical representation of a desktop computer, indicating that users will access data from those particular components in the overall environment, but not any...