What Happens when Internal IT Manipulates a Marketing Database

Last month we noted that, when it is time to build a Marketing Database, a company's internal IT group almost always lobbies tenaciously for the assignment. This is because IT professionals deeply associate the word "database" with nuts-and-bolts computer science. We also observed that many internal Marketing Database builds are unsuccessful because much of what determines success has nothing to do with database technology. After providing seven examples, we concluded that:

"Clearly, the database developers who can best navigate such complex issues are those with many years of hands-on data mining experience. As seasoned power users of data, such individuals possess the knowledge required to create Best-Practices Marketing Database ContentSM, and avoid the missteps that compromise the usefulness of so many databases."

Just as internal IT groups want to build Marketing Databases, so too do they want to be the ones who manipulate them once they are up and running. As with internal builds, delegating ongoing Marketing Database manipulation to IT professionals often leads to unfortunate outcomes. The reason is the same as with database builds: a deep, hands-on understanding of data mining is extremely helpful when navigating the inherent complexities of database manipulation. The following are just three of the tasks that must be performed:

  • Apply dynamic business rules to create the multiple past-point-in-time ("time-0") views required to develop sophisticated scoring and decision models. An example of the former is specialized models to predict customer behavior resulting from different types of promotional efforts such as outbound telemarketing, and where each analysis file contains time-0 views corresponding to several recent campaigns. An example of the latter is quantification of the over-time effects of page counts on the performance of direct mail catalogs.
  • Apply dynamic business rules to multiple time-0 views, in order to generate ongoing reports showing: 1) historical trends in customer segment counts and performance, and 2) channel performance (such as direct mail, e-commerce and retail) corresponding to specific lines of merchandise.
  • Apply dynamic business rules to perform the production scoring of multiple specialized predictive models, and integrate them with homogeneous classification ("segmentation") systems, in order to execute selections for multi-channel promotional campaigns.

The bottom line is that ongoing Marketing Database manipulation requires much more than IT skills. Therefore, regardless of whether the manipulation is done in house or by a third party, make sure that the maintenance team is composed of more than IT professionals!