Targeting / Statistical Modeling Overview: Maximize objective (i.e. response rates, incremental sales, marketing mix) using regression analysis, target models, segmentation using cluster and factor analysis, forecasting and tracking using predictive models, significance testing. Maximize Objective Using Regression Modeling Improve response rates for acquisitions and renewals, increase sales to target audiences, and optimize marketing mix. Example: built models for several of HachetteFilipacchi magazine titles identifying subscribers most likely to renew. In this case, models were coupled with an innovative mail timing solution and mail selection strategy. More about response modeling (PowerPoint). Target Models Identify brand preference, affinity points programs redeemers, likely program defectors, product switchers & migrators within program, the timing of market entry, and heavy users. Example: built a GM Card redeemer model identifying the likelihood each customer would redeem their rewards and how soon for millions of cardholders. Segmentation Using Cluster and Factor Analysis Identify audience types such as defectors, in market buyers, ethnic groups, military/college buyers, employees, owners/loyalty, competitors/conquest, and unique segmentation schemes using multivariate analysis. Example: Segmented GM’s hourly, salaried and retired employees into research defined “attitude and energy” dimensions for inclusion in their Ambassador Program, designed to encourage and empower employees to sell cars. More about segmentation (PowerPoint). Forecasting and Tracking Forecasting; including seasonality and other patterns using previous months or years of data and assumptions about future trends. Results are tracked and compared to forecasts. Example: dynamic, interactive sales and revenue forecasts for the GM Card. Significance Testing Distinguish between results that are directionally different and those that are statistically significant.