Scenario Modeling Tip #5: Use dummy communications to enforce other types of dependencies between communications.
Suppose your Communications table includes an estimate of COMMUNICATION_UNIT_COST for each communication. However, some of your communications might use the direct mail channel and you want to model the cost structure that allows a customer to receive multiple communications within the same direct mail piece for a single cost. You might create a dummy communication to represent the vehicle through which direct mail communications are made. This is not a real communication; think of it as an empty envelope that incurs the mailing cost. Therefore, the objective measure of this dummy communication must be 0, and the COMMUNICATION_UNIT_COST must represent the estimated combined cost of mailing one or more direct mail communications. The COMMUNICATION_UNIT_COST of the real direct mail communications must be changed to 0 or some small number that represents the incremental cost of including the additional communication in the combined direct mail piece.
Then, you must enforce a dependency between each real direct mail communication and the dummy direct mail communication: a customer who receives Communication A through direct mail must also receive the dummy communication through direct mail. Furthermore, a customer who receives the dummy communication through direct mail must also receive at least one of the real communications through direct mail.
But be careful. When using dummy communications, carefully examine the communication filters of your constraints and contact policies to verify that they model the problem as you had intended. For example, a contact policy on the total number of contacts should not include both the dummy communication through direct mail and all the real communications through direct mail. If you want a single mail piece to be counted as one contact, then your communication filter should include the dummy communication through direct mail but exclude the real direct mail communications. On the other hand, if you want a single mail piece to be counted as multiple contacts, based on the number of communications in the mail piece, then your communication filter should include the real direct mail communications but exclude the dummy communication through direct mail.
Note: This post was adapted from Michelle Opp’s “Scenario Modeling Tips” document. For more information, see SAS Marketing Optimization User’s Guide.