Forecasting department store sales using a dummy variable model

Abstract

The problem of forecasting department store sales has existed for quite some time. Sales do fluctuate during certain seasons of the year. The major objective of this study was to investigate why department store sales fluctuate and to develop a prediction model to help forecast sales taking seasonal variation into account.Specific objectives of the study were to: 1) determine the trend in monthly sales for the Muncie/Anderson area by drawing a freehand curve; 2) determine the trend in monthly sales to discern the national pattern by drawing a freehand curve; 3) determine the months during which variations in sales occur both at the local and national levels; 4) determine the causes that might he responsible for variations in monthly sales; 5) determine the trend in monthly sales using the method of least squares both for the local and the national data; and 6) develop a statistical model to predict monthly sales.Monthly sales data obtained from various department stores in the Muncie/Anderson area were analized using an IBM 360/50 computer available at the Ball State University. The total number of observations included for the regression analysis was 120. A Dummy Variable model which takes into account the seasonal variations in sales was used to develop the prediction equation. The analysis showed a high degree of correlation between the monthly sales and the independent variables and the t and F values were highly significant.