Optimization of Warehouse Inventory Policy Using ABC–XYZ Analysis and the (Q,R) Model to Reduce Total Inventory Cost and Stockouts

Abstract

Warehouses are pivotal nodes in supply chains, yet many distribution centers still apply uniform replenishment rules that ignore differences in item value and demand uncertainty. This practice often leads to simultaneous overstock (high holding cost) and stockouts (service failure), reducing overall operational efficiency. Objective: This study aims to optimize finished-goods inventory control in the distribution warehouse of Manufacturing Company X by integrating ABC–XYZ classification with class-based replenishment policies to improve the cost–service trade-off and service reliability. Methodology: The research uses a quantitative applied case-study design. Data were collected from secondary operational records (SKU demand history, lead time, inventory transactions, and cost parameters) and supported by primary inputs through observation and interviews to confirm replenishment constraints and routines. Analysis was conducted by (1) classifying items using ABC (annual usage value) and XYZ (demand variability), (2) translating classes into differentiated service targets and continuous-review policy parameters (order quantity, reorder point, and safety stock), and (3) evaluating performance using a before–after KPI comparison between the baseline and proposed policies. Findings: The results show strong value concentration and heterogeneous demand variability across SKUs, supporting differentiated control. The proposed policy reduces total inventory cost from IDR 8.75 billion to IDR 8.10 billion (−7.4%), decreases stockout incidents from 96 to 52 (−45.8%), and increases service level from 92.4% to 96.1% (+4.0 percentage points). Improvements are most pronounced in high-priority and high-uncertainty groups. Implications: The findings suggest that managers can improve warehouse service reliability while lowering costs by allocating buffers and control intensity according to item priority and uncertainty, supported by periodic class refresh, master-data governance, and inventory record-accuracy improvement (e.g., class-based cycle counting). Originality: This study contributes an end-to-end, implementable pipeline from ABC–XYZ segmentation to differentiated service targets, policy parameterization under practical constraints (e.g., MOQ/pack sizes), and KPI-based validation demonstrating measurable operational benefits beyond classification-only approaches.