Supply Chain Inventory Replenishment: The Debiasing Effect of Declarative Knowledge

Abstract

ABSTRACT Previous experimental research demonstrates that inefficient replenishment decision making in the supply chain can be caused by specific judgment and decision biases. Based on the literature we use controlled experiments involving both student subjects and supply chain managers to test debiasing interventions that provide declarative knowledge, which is theorized to enhance the acquisition of procedural knowledge. We first investigate the effects of three debiasing components in a single‐echelon setting: knowledge of bullwhip, inventory position (IP), and use of a target order‐up‐to quantity. Experiment 1 (N = 1,608 decisions by 67 student subjects) using a 2 × 2 × 2 factorial design for the three components finds that the conceptual understanding of IP is salient for efficient replenishment decisions. We next examine the effects of the components in a simulated, multi‐echelon, serial supply chain, which introduces the additional complexity of coordination risk. Experiment 2 (N = 3,072 decisions by 128 student subjects) using a 2 × 2 × 2 factorial design finds that although subjects benefit from training components, there is evidence of cognitive overload with an increased quantity of information. Finally we test whether these debiasing components may be an effective training program for practicing supply chain managers who can be expected to have higher levels of procedural knowledge through experience gained in the field. Experiment 3 (N = 864 decisions by 36 supply chain managers) using a 2 × 1 design investigates the effects of an instructional training intervention which includes all three debiasing components and finds the intervention to reduce costs by 14%. We provide avenues for future research and successful practice.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 50 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.