Forecasting Item Movement with Scan Data: Box-Jenkins Results

Preliminary forecasts using the Box-Jenkins methodology for supermarket scan data for ground beef and roast item movement are described. The functional form and the accuracy of the forecasts vary by product. Results suggest that further analyses incorporating price and advertising may increase the accuracy of the forecasts. Accurate forecasts of sales can be a key determi-nant of the economic viability of any business. This is especially important in the highly competitive supermarket industry. Low profit margins neces-sitate very careful management of inventories, scheduling of labor, and timing of shipments. The introduction of scanners was heralded as a tech-nology that could generate sales data for use in managerial decisions in these areas, thereby re-ducing costs. While some progress has been made, the super