Analysis Of Product Demand Forecasting Using Time Series Methods

This research is motivated by the uncertainty of product demand at the Si Geboy Tea Cracker Factory, which causes both overproduction and underproduction. This study aims to assist the Hikmat Si Geboy Tea Cracker Factory in forecasting demand for the January 2026 period. The analytical tools used were the Moving Average, Weighted Moving Average, and Exponential Smoothing methods, with forecast accuracy tested using MAD, MSE, and MAPE. The data. The results showed that the MA 3-Period method predicted a demand of 7,932 dozen, with a MAD of 452.26, MSE of 288.050, and MAPE of 5.54%. The WMA 3-Period method predicted a demand of 7,724 dozen, with a MAD of 425.74, MSE of 266.599, and MAPE of 5.19%. ES α = 0.1 predicted a demand of 8,168 dozen, with a MAD of 418.22, MSE of 252.198, and MAPE of 5.24%. ES α = 0.5 predicted a demand of 7,810 dozen, with a MAD of 420.95, MSE of 264.257, and MAPE of 5.25%. ES α = 0.9 predicted a demand of 7,460 dozen, with a MAD of 398.99, MSE of 259.721, and MAPE of 4.93%. The conclusion from the three methods is that the most suitable method for forecasting cracker demand for January 2026 is the Exponential Smoothing method with α = 0.9 because it has the best accuracy compared with the other methods, as indicated by its lower MAD, MSE, and MAPE values.