In the traditional approach to retention, a customer retention analyst "mines" the customer information file (CIF) to find customers who have switched to the competition. The analyst then tries to find characteristics of those who have defected. At its simplest, this is an exercise in finding demographics, life or service events in common among those who defected to the competition. At the other end of the spectrum, this analysis uses multivariate predictive models that score all customers with a defection likelihood score. Decision makers can then try to "head the customer off at the pass" by offering reward or incentive programs to their most highly valued customers.