Complex customer behavior can sometimes be discovered only by running lengthy iterative analyses. In these cases, it is impractical to embed the behavior analyses inside every BI application that wants to constrain all the members of the customer dimension who exhibit the complex behavior. The results of the complex behavior analyses, however, can be captured in a simple table, called a study group, consisting only of the customers’ durable keys. This static table can then be used as a kind of ﬁlter on any dimensional schema with a customer dimension by constraining the study group column to the customer dimension’s durable key in the target schema at query time. Multiple study groups can be deﬁned and derivative study groups can be created with intersections, unions, and set differences.