Minimizing Excess Inventory and Its Impact by Enhancing Forecasting Accuracy

Abstract

This study examines how PT. TRUCKSTORS LANCAR ABADI can improve inventory management by addressing recurring sales forecasting issues. Despite extensive research on forecasting in manufacturing, there is a notable gap in understanding how to manage seasonal demand patterns in Indonesia's heavy equipment sector using integrated forecasting methods. This gap is especially clear in the lack of validated models tailored to capital-intensive industries with volatile demand cycles. Using both qualitative and quantitative approaches, the study began by gathering insights from key departments and applying the Analytic Hierarchy Process (AHP). The analysis identified inaccurate forecasts as the primary cause of inventory discrepancies. To address this, several forecasting models were tested. SARIMA emerged as the most effective, capturing seasonal demand fluctuations with a Mean Absolute Percentage Error (MAPE) of 15.3% and a Mean Absolute Error (MAE) of 257 units. This outperformed alternatives such as Multi-Linear Regression (MAPE: 18.7%), Neural Networks (MAPE: 19.4%), and XGBoost (MAPE: 20.1%). Implementing SARIMA is projected to reduce interest loss by 88% (from USD 3.32 million to USD 0.39 million) and inventory holding costs by the same margin (from USD 9.64 million to USD 1.13 million annually). This model is expected to improve planning accuracy and better align projected inventory with actual levels. The study highlights the importance of integrated planning, cross-functional collaboration, and data-driven forecasting for Minimizing Excess Inventory and Its Impact by Enhancing Forecasting Accuracy in volatile markets

Affiliated Institutions

• Bandung Institute of Technology ID

Related Publications