Modeling & Forecasting

Decision Sciences Research Associates (DSRA) research team has extensive experience in developing statistical models which are used to predict customer behaviors. Models are used to predict outcomes in complex situations. Models can include the use of summarized data about groups (e.g., zip code level demographics, or detailed weather patterns at small localized areas) and/or about individuals (e.g., previous purchase history, house characteristics, or gender). The models yield a score that is customized to the exact behavior being predicted. It uses information about past events to determine what characteristics are important in predicting future behaviors.

Response modeling uses past campaign results to forecast future performance as it relates to demographic, socioeconomic, or other available characteristics. The forecast is used to select higher performance candidates for the campaign or reject lower performance candidates that are below breakeven. Econometric methods estimate the relationships among complex sets of actions and can be most useful for evaluation of program interventions or business policy. Site selection and location modeling helps our customers to determine the optimum placement of resources. By looking at the spatial and behavioral elements involved in customer decision making, DSRA maps out optimum site selection criteria and projections.