Top-Down Versus Bottom-Up Demand Forecasts: The Value of Shared Point-of-Sale Data in the Retail Supply Chain

Abstract

Accurate demand forecasts are critical to maintaining customer service levels and minimizing total costs, yet increasingly difficult to achieve. Using weekly point-of-sale (POS) and order data for 10 ready-to-eat cereal stock-keeping units from 18 regional U.S. grocery distribution centers, this research empirically investigates two demand forecasting issues: (1) the accuracy of top-down versus bottom-up demand forecasts; and (2) whether shared POS data improve demand forecast accuracy. The results reveal a previously unexplored relationship between demand forecast methodology and the use of shared POS data. We find that the superiority of the top-down or bottom-up forecasting as the more accurate demand forecast method depends on whether shared POS data are used.