Order Ahead for Pickup: Promise or Peril?

• 成果类型: Article
• 署名作者: Sun, Ke; Liu, Yunan; Yang, Luyi
• 署名单位: McGill University; Koc University; North Carolina State University; University of California System; University of California Berkeley
• 刊物名称: M&SOM-MANUFACTURING & SERVICE OPERATIONS MANAGEMENT
• ISSN/ISSBN: 1523-4614
• DOI: 10.1287/msom.2024.0865
• 发表日期: 2026
• 关键词: Queues

摘要

Problem definition: Mobile technologies have increasingly enabled remote customers to order ahead at quick-service restaurants. As customers travel to the service facility to pick up their orders, their orders also advance in the food preparation queue. It is widely believed that the ability to order ahead reduces customers' total delay and, therefore, allows restaurants to attract more orders and achieve higher throughput than if customers must order onsite. Methodology/results: We build a queueing game-theoretic model to study a mixed ordering scheme in which some customers order ahead and some order on-site. Our analysis shows that the common practice of accepting all orders as they come in and requiring all orders to be irrevocable can cause a mixed ordering scheme to surprisingly achieve lower throughput than an on-site-only ordering scheme. The throughput shortfall can persist even when the service provider freely chooses whether to share queue-length information with remote customers. However, if remote customers who order ahead are allowed to cancel unprepared orders when they arrive at the service facility, then such a mixed ordering scheme with cancellation achieves higher throughput than the on-site-only ordering scheme even though it does not uniformly dominate the original mixed ordering scheme (without cancellations). We then study a capped ordering scheme in which the service provider stops accepting new remote orders if the number of outstanding orders reaches a certain threshold. When the service provider optimally sets the cap, the mixed-capped ordering scheme outperforms both the mixed ordering scheme without capping and the on-site-only ordering scheme in throughput but not necessarily the cancellation scheme. Finally, we propose an optimal mechanism in which the service provider determines both the capping and cancellation thresholds subject to customers' individual rationality constraints. Managerial implications: Our paper highlights the unintended consequences of ordering ahead and provides prescriptive guidance for managing such a service system.