Research Article on:WushWush Tea Plantation Forecasting Sales Using the Box- Jenkins Approach

Abstract

This study is conducted to facilitate the decision making process of Wushwush Tea Planation Factory by developing an appropriate forecasting model. It aims at providing accurate sales forecasts for future sales, which is a vital input in decision-making. Particularly, the focus is given for forecasting monthly Tea sold by the factory. Towards achieving its objective the study considered the Box-Jenkins approach to time series analysis. A total of 60 monthly sales data has been taken for analysis or model building purpose. Moreover, additional 5 months sales data has been used for forecasting purpose. The analysis of the data, which is carried out using S-Plus 2000 package, suggested that ARIMA (3,1,0) model represents the pattern of monthly Tea sales data. According to this model, forecasting current sales essentially requires the inclusion or consideration of the previous four consecutive sales data occurring at the four successive lags. Moreover, it is found that the sales data recorded in the first lag has greater influence or contribution in forecasting current sales volume. On the other hand, it is observed that the sales data involves a seasonal component that turns out to affect the sales volume approximately in 3.5 months. In other words, the analysis indicated that there is a seasonal component that occurs with 3 or 4 months periodicity. This in turn resulted in attaching high importance to the third and fourth lag coefficients as compared to the coefficient of the third lag. The results obtained led to the conclusion that the time factor is the major but not the only relevant factor in forecasting sales. Other considerations in relation to promotional activities, competitors’ action, seasonal factors, etc should be kept in mind.