The US population is changing dramatically, e.g. aging, migrating, becoming more racially diverse and overweight, etc. We examine how these major shifts affect snack consumption and, therefore, how Consumer Packaged Goods (CPG) manufacturers should adjust their business-as-usual practices. We present the development and implementation of DemoImpact, a consumption forecasting model and decision support system for all snacks available on the US market (900+). It captures the effects of 48 demographic variables (age, gender, marital status, obesity, education, income
), and accounts for the effects of significant holidays and seasons. DemoImpact is built on 6 years of historical snack consumption and uses the US Census demographic predictions to forecast consumption frequency and volume. The model was implemented at a leading CPG company and forecasted consumption well. The DSS enables top and mid-level managers first to understand how the consumer is changing, and then how to prepare for the change in tastes and how to develop accurate forecasts. DemoImpact has been used, validated and extended since 2006. We show how the CPG manufacturer embraced the results of the model, adjusted its business strategy and impacted retailers in its sales channel.