Improving a frontier approach to selecting suppliers for a make-to-order supply chain

Purpose Make-to-order (MTO) manufacturers often receive lead time-sensitive orders requiring multiple custom parts not commonly in stock. These irregular orders can be challenging to manage across numerous suppliers, resulting in high costs and associated risks in completing orders on time. Many MTO manufacturers develop relationships with multiple sourcing suppliers, spreading risk and ensuring delivery performance at a reasonable cost. These relationships, however, result in an order allocation problem. The purpose of this study is to consider previously proposed frontier portfolio solution methodologies used in sourcing selection and improve upon an identified weakness to improve the evaluation and selection of order quantities across multiple suppliers. Design/methodology/approach This paper analyzes an existing frontier portfolio selection model where potentially inefficient portfolios are selected and considered. The authors propose introducing a data envelopment analysis (DEA) technique paired with the frontier portfolio selection model to eliminate inefficient portfolios from being considered to help MTO manufacturers determine allocations by calculating and evaluating two factors: on-time probability and associated cost. Findings While previously proposed frontier models provide mechanisms to allocate orders across different suppliers to meet variable service levels and costs, this study finds model improvements to eliminate inefficient portfolios to facilitate better decisions for order allocation. This study finds a more efficient potential solution using a frontier method paired with DEA to help the MTO manufacturer select sourcing partners and determine the associated quantity allocation for each chosen supplier. The technique provides flexibility, resulting in a range of options to solve the dilemma with differing emphasis on due dates and costs. Practical implications The proposed method provides a pool of more efficient frontier portfolios ranging from higher on-time probability and higher cost to less on-time likelihood and lower cost, resulting in less chance of an inefficient portfolio choice. Such a design ensures that regardless of the preference and specific need criteria (cost versus on-time delivery), MTO manufacturers have flexibility in choosing an optimal solution. Originality/value This paper extends an existing model of supplier selection to improve selection quality based on a DEA technique to remove inefficient portfolios.