Demand variability in engineer-to-order supply chains: insights from a DDMRP case study

Engineer-to-Order (ETO) supply chains with modular T-plant structures face critical inventory challenges: shared components must buffer against demand variability across multiple product variants, while planners struggle to balance responsiveness and cost efficiency under stochastic demand. Demand-Driven Material Requirements Planning (DDMRP) has been proposed to enhance resilience and inventory efficiency, yet its effectiveness in such contexts remains underexplored. This study simulates a two-echelon industrial piping chain, where shared components are managed as Make-to-Stock and final assemblies as ETO. Using real purchase order data, DDMRP is compared with the current push-based method at the supplier level, tracking inventory levels, fill rates, and order counts. Results show that DDMRP reduced weighted average on-hand inventory by 21.4%, improved fill rates by 7.5 percentage points, and cut backorders by 76%, while increasing order frequency by 1.5%. Sensitivity analysis highlights the Average Daily Usage (ADU) window as the primary lever for balancing service and inventory. The study demonstrates that resilience can be achieved through adaptive response rather than static redundancy alone, identifies signal averaging horizon as a key parameter affecting the balance between responsiveness and efficiency, and supports selective use of DDMRP, suggesting particular effectiveness under high uncertainty in demand timing, or both timing and quantity.