Sales Forecasting of Perishable Products: A Case Study of a Perishable Orange Drink

Abstract

The primary goal of any organization involved in trading business is to maximize profits while
keeping costs to a bare minimum. Sales forecasting is an inexpensive way to achieve the aforementioned
goal. Sales forecasting frequently leads to improved customer service, lower product returns, lower deadstock,
and efficient production planning. Because of short shelf life of food products and importance of
product quality, which is of concern to human health, successful sales forecasting systems are critical for the
food industry. The ARIMA model is used to forecast sales of a perishable orange drink in this paper. The
methodology is applied successfully. ARIMA (0, 1, 1)(0, 1, 1)12 was concluded as the appropriate model.
Model diagnostics were done; results showed that no model assumption was violated. Fitted values were
regressed against observed values. A very strong linear relationship was evident with an R2 value of over
90% which is very plausible.

Keywords: Sales forecasting, ARIMA, Model Diagnostics, R2- value.