Statistical Model For Short-Term Forecasting Sparkling Beverage Sales In Southern Thailand

Abstract

This study developed a simple statistical model for forecasting a companys sparkling beverage sales in the 14 provinces of Southern Thailand. Data comprised sales revenue from January 2000 to December 2006 obtained from the company. We fitted an observation-driven multiple regression model to log-transformed monthly revenue containing season of year (month), location and beverage flavour as factors, as well as lagged observations for the preceding four months. The model gave a r-squared of 0.95 and was effective for forecasting revenues for up to 12 future months. Using such models for forecasting sales revenue can assist company managers with planning more effectively.