Abstract: Method and system for producing custom behavior scores for use in credit decisioning. A custom credit score takes into account credit bureau attributes and relatively detailed transaction attributes as well as traditional masterfile information for a customer applying for credit on an account. To determine the custom score, accounts at the financial institution are segmented. Segments can be created by grouping accounts into a plurality of clusters based on Euclidean distances between points in a multi-dimensional space, as well as in other ways. The customer's credit score is determined according to a behavior model created for a particular segment of accounts into which the customer's account falls.