Abstract: System and method for managing debt collection using clustering. Debt collection treatment for a delinquent account can be determined in a reliable and consistent way by examining historical credit and demographic data for the debtor. In some embodiments, a plurality of behavior clusters within a multi-dimensional space for modeling debtor behavior are established using representative portfolio data. Result probabilities for various stimuli to be applied to the customer can be represented by a difference score, which indicates a difference in result flowing from the presence of particular stimuli versus the absence of the stimuli. The difference scores for the various stimuli, when normalized, can be used to assign the delinquent account to a cluster, which determines the treatment to be applied to the account.