The type of information that a direct marketing professional may want to retain about a customer or prospect may best be maintained in a relational database. The concept behind a relational database is that any particular data element is maintained in one place, rather than repeated in multiple places or multiple records.

Below is a very simple example of how a relational structure makes maintaining the data easier. Suppose we wanted to maintain the purchases of customers of the XYZ company sells golfing accessories through catalog sales. Typical data records for XYZ company might be:

• Record1: Debbie Andrews, 123 Main St, 1 package golf balls, 1/2/2001
• Record2: Debbie Andrews, 123 Main St, 1 pair golf shoes, 1/2/2001
• Record3: Debbie Andrews, 123 Main St, 1 golf sweater, 11/12/2001

Alternatively, the above information could be stored in a relational database in two separate tables that are linked by unique id:

Customer Segment:

200145, Debbie Andrews, 123 Main St

Purchase Segment:

• 200145, 1 package golf balls, 1/2/2001
• 200145, 1 pair golf shoes, 1/2/2001
• 200145, 1 golf sweater, 11/12/2001

In this simplified example, clearly the individual name and address is stored in only one place on the customer segment and not replicated for each product purchase. The benefit of this design is that the potential errors in maintaining the name and address information decreases. For instance, if Debbie Andrews moved, then only one record in the relational database needs to be found and changed. Originally, multiple records would need to be updated for an address change, and they may be updated differently, or some records may be skipped altogether in error.

In addition, information about the purchase history of Debbie Andrews can be rolled up to the customer segment record as follows:

Customer Segment:

200145, Debbie Andrews, 123 Main St, Last Purchase Date 11/12/2001

Now the purchase information is usable in the form of a data element that can be selected on. The customer segment contains one record per individual, with date of last purchase included on the record. This data storage format makes it much easier to know quickly how many individuals have purchased from XYZ company, and to be able to segment (or count) the individuals based on last purchase date.

In addition, because data elements are stored once, the absolute size of the database is smaller under a relational database. In the above illustration, the name and address is repeated three times in the non-relational example. This data storage problem becomes magnified as the number of records increases. The more data on a system, the slower it becomes to search for records.