EXPRESS: Execution Failures in Retail Supply Chains – A Virtual Reality Experiment

Abstract

Retail store employees are increasingly being asked to pick orders from inventory. These tasks are performed under intense conditions and are often made more difficult because of high product variety and high degrees of product similarity. We conduct a real-effort task in a virtual environment where subjects must sort cubes into bins. We study task complexity by varying the degree of similarity between the cubes and task intensity by varying the arrival pace of the cube. Beyond traditional descriptive performance analysis, we also analyze subjects’ movements. We focus on four performance metrics: Throughput (number of cubes sorted per minute), Accuracy (percentage of correctly sorted cubes), Productivity (number of correctly sorted cubes per minute), and Error Rate (number of incorrectly sorted cubes per minute). When task complexity is reduced, productivity rises by as much as 38.2%, and the error rate falls by as much as 93.6%. It also leads to more efficient movements. Increasing task intensity improves throughput but decreases accuracy slightly while varying task intensity improves performance via faster learning. Subjects tend to cut corners when the task is more complex or more intense.