This observation analyzes Minyakita sales forecasting at PT Java Agri Sukses Makmur using the POM-QM application. Accurate sales forecasting is important for stock planning and management. Sales data were collected through interviews conducted over three months (March-May 2025). Three forecasting methods were employed, namely linear trend analysis, Single Exponential Smoothing, and Simple Moving Average. The results of Trend Analysis Linear show a sales pattern that tends to increase. The regression equation is Demand (y) = 11, 267. 68 + 881. 664xTime. Although the Bias value (0. 001) is close to zero, error values such as MAD (4, 663, 778), MSE (28, 397, 000), and Standard Error (5, 837. 5) show a fairly high error rate. However, the MAPE value (30. 745%) is considered feasible for medium-term forecasting. The correlation coefficient of 0. 496 and R² of 0. 246 indicate a moderate relationship, with the variation in sales explained by time accounting for only 24. 6%. Single Exponential Smoothing (=0. 1) resulted in a Bias of 1, 684. 15 (tends to underestimate) and MAPE of 32. 77%, with a forecast of 17, 307. 56 units for the next period. Simple Moving Average shows a MAPE of 27. 191% (feasible), with a forecast of 19, 616. 25 cartons. Overall, the linear trend model is suitable for identifying long-term patterns, but less responsive to short-term fluctuations. These forecasting results are expected to support PT Java Agri Sukses Makmur in production and marketing strategies. Contribution to Sustainable Development Goals (SDGs): SDG 2: Zero HungerSDG 8: Decent Work and Economic GrowthSDG 9: Industry, Innovation, and InfrastructureSDG 12: Responsible Consumption and Production
Zahara et al. (Fri,) studied this question.