Knowledge that Transforms

The continued success of Manufacturing & Service Operations Management (M&SOM) depends on the volunteer work of many professionals who take their precious time to provide careful and constructive reviews of the manuscripts submitted to the journal in a timely manner. On behalf of M&SOM, Editor-in-Chief Georgia Perakis expresses her deepest gratitude to all those who served as reviewers for the journal in 2025. Among all reviewers, some individuals have distinguished themselves by reviewing several manuscripts and, with each manuscript, by writing a fair, critical, and constructive review in a timely fashion. In recognition of their outstanding service provided to support the journal’s scholarly mission, M&SOM grants the 2025 Meritorious Service Award to…

Problem definition: Faced with the growing need to improve access to care, hospitals have been operating at capacity in recent times. We argue, however, that the pursuit of utilization beyond a threshold might come at a cost. We propose that there exists a zone of efficiency within which the costs of care reach the minimum and the pursuit of utilization outside of this zone may be counterproductive. Our second objective is to explicitly incorporate the complexity of care requirements in understanding the relationship between utilization and costs of care. We propose that complexity exacerbates the demands on resources during care delivery and shifts the zone of efficient utilization lower. Methodology/results: Using patient-level data from over 325,338 inpatient discharges in 156 hospitals, we empirically test these propositions. Our findings indicate that the marginal cost curve is nonlinear, revealing a zone of efficiency between 75%–85% utilization, within which costs per patient day are minimized. Departing from this zone (either above or below it) incurs substantial cost penalties. The results also reveal heterogeneity in the cost implications of utilization across departments with varied complexity regimes. Managerial implications: The study findings carry key implications for managers in improving access to care in an efficient manner. Our study informs decision making in this regard by highlighting that more (utilization) is not always better. Furthermore, a one-size-fits-all policy across all departments may be counterproductive. Whereas managers may pursue high utilization in some departments, it may not be prudent in others.

Problem definition: Assortment selection and marketing mix allocation are critical decisions for retailers, directly influencing consumer choices. In this paper, we propose a multinomial logit (MNL) choice model in which consumer utility is influenced by marketing decisions such as advertising and promotions: a model widely utilized in empirical marketing literature. We then study the joint assortment and marketing mix allocation problem subject to either cardinality constraints or knapsack constraints on marketing mix allocations. Methodologies/results: We prove that the problem under cardinality constraints is already strongly NP-hard and does not admit constant ratio approximation. For the model with cardinality constraints, we provide an optimal ratio approximation algorithm and polynomial-time algorithms for special cases. With a constant number of marketing mix decisions, the problem can be solved using a linear program of polynomial size. Under knapsack constraints, we also provide an optimal ratio approximation algorithm and a fully polynomial-time approximation scheme (FPTAS) for special cases. With a constant number of marketing mixes, the problem admits an optimal polynomial-time approximation scheme (PTAS). Computational experiments with real-world NielsenIQ retail data show significant 2.05% revenue increases using our method over a two-stage “assortment-then-marketing mix allocation” heuristic approach. A complementary experiment using online transaction data from JD.com is also conducted to demonstrate the applicability of our method in online settings. Managerial implications: Our comprehensive numerical experiments across various scenarios demonstrate that neglecting the impact of marketing decisions in assortment selection can lead to a significant decline in profitability. This finding demonstrates the importance of jointly optimizing assortment and marketing mix allocation, particularly when the number of marketing mix decisions significantly exceeds the number of products and when resources for marketing mix allocation are limited.

Problem definition: Emotional labor is increasingly demanded in service operations, placing tremendous psychological strain on human employees and posing challenges to scalability and sustainability. Our study scrutinizes whether artificial intelligence (AI) service bots may tackle these challenges by examining how and when AI’s engagement in emotional labor enhances economic performance in service operations. Methodology/results: We provide causal evidence from a pair of randomized field experiments conducted in partnership with a firm for loan collection service. Results suggest that, compared with human employees, undisclosed AI service agents display the required emotions (both positive and negative) more accurately. However, AI’s advantage of higher emotion display accuracy does not always guarantee better economic performance. For AI to collect more payments from borrowers than human workers, the displayed emotion must be contextually appropriate. Specifically, AI substantially outperforms human workers in debt collection by 49%–94% when the emotion display instructions are suitable for the collection task (i.e., displaying positive emotions to borrowers with minor delinquency but negative emotions to borrowers with repeated delays). However, when the displayed emotions are unsuitable, AI backfires and performs worse than human employees because of its unwavering adherence to inappropriate emotional display instructions. Further, AI’s performance advantages over human agents are amplified when the suitable emotions involve negative (versus positive) valence. We also leverage the machine learning method causal forest to explore heterogeneous treatment effects across customer segments. Managerial implications: Our research suggests that operations managers should deploy AI to reduce frontline employee emotional burnout, develop explicit emotional labor guidelines for AI, and use negative-emotion AI strategically to boost compliance and efficiency. It is also important to identify emotion-intense operational tasks, target AI when it has clear advantages, and set up continuous monitoring and quality control for AI emotional performance.

Problem definition: Our research investigates how preference satisfaction, particularly intrinsic values such as psychological comfort, can improve workers’ service efficiency and quality. Methodology/results: We utilize a unique setting to examine solely the intrinsic values driving workers’ service performance. In this setting, surgeons have operating room preferences that result only from intrinsic values because these operating rooms are instrumentally identical. Examining a comprehensive data set linking surgeons’ performances to their preferences for operating rooms, we confirm and quantify the intrinsic benefits of preference satisfaction on service efficiency and quality. We also find that compared with workers without preferences, workers with preferences perform better if satisfied, but worse if unsatisfied. Moreover, we find that when workers are under heavy workloads or performing complex tasks, these preference effects are more pronounced. Based on our findings, we update the surgery scheduling framework by incorporating surgeons’ preferences. Our counterfactual analysis shows that, for the operations in our sample, the upper-bound healthcare cost savings from satisfying surgeons’ preferences can exceed $3.5 million while also improving patients’ and surgeons’ welfare. Managerial implications: Our findings suggest that firms can utilize the intrinsic benefits of preference satisfaction as a lever to improve service performance without incurring the cost of instrumental changes. Moreover, firms can consider cultivating workers’ preferences if their systems have enough flexibility to satisfy them. When firms cannot satisfy the preferences of all workers, they can consider prioritizing workers with heavy workloads or complex tasks to maximize the improvement in service performance. Specifically for the healthcare industry, the scheduling system can integrate surgeons’ preferences for operating rooms into the optimization framework to achieve huge benefits in operation cost saving and patient welfare improvement at little expense.

Problem definition: Despite rapid advances in artificial intelligence, the adoption and effective use of customer service chatbots remain slow relative to their capabilities. This paper explores the behavioral reasons for these adoption hurdles. Methodology/results: We use incentivized online experiments to study chatbot uptake. The results of these experiments are threefold. First, people respond positively to improvements in chatbot performance; however, the chatbot channel is used less frequently than expected-time minimization would predict. A key driver of this underutilization is reluctance to engage with a gatekeeper process (i.e., a process with an imperfect initial service stage and possible transfer to a second expert service stage—a behavior that we term gatekeeper aversion). Second, we find that gatekeeper aversion can be further amplified by an additional hurdle—algorithm aversion. Third, we find that chatbot adoption decreases when stakes are higher and when the human/algorithmic nature of the server is manipulated with more realism. Managerial implications: We use an illustrative case to show how the behaviors identified in our experiments affect optimal technology investment and staffing levels and how failing to anticipate these behaviors can lead to suboptimal decisions and higher realized costs. More broadly, our results suggest that adding a chatbot to a service system requires rethinking the entire service process, including technology investment, staffing, and queueing policies.

Problem definition: Probabilistic selling (PS) is a novel selling strategy whereby consumers only know the exact product identity after the payment. By studying PS for vertically differentiated products in a supply chain consisting of one supplier and one retailer, we examine who should assemble the probabilistic products. We explore two cases, retailer assembly (RA) and supplier assembly (SA), analyzing the dynamics and implications within each. Methodology/results: Using game theory methodology, we characterize the best selling strategies for both RA and SA cases, and obtain the equilibrium strategy. Generally, PS is optimal when the associated transaction cost is below a threshold; in this case, PS can alleviate the double marginalization problem in distribution channels. Interestingly, in comparison with a centralized channel, PS adoption is less probable in the RA case but can be more so in the SA case. When PS emerges as the optimal strategy for both cases, the supplier under the SA case mixes a smaller fraction of high-quality products than the retailer would do under the RA case. Moreover, at equilibrium, a relatively powerful supplier should assemble probabilistic products, whereas a relatively powerful retailer may prefer not to have the stewardship of assembling probabilistic products. Notably, PS can serve as a mutually beneficial strategy for the supplier, retailer, and consumers, offering a “win–win–win” outcome. Managerial implications: The supplier should assemble probabilistic products targeting low-type consumers and offer them alongside high-quality (and low-quality) products when the capacity of high-quality products is abundant. Conversely, the supplier should assemble probabilistic products targeting high-type consumers and offer them alongside low-quality products when the capacity of high-quality products is limited. Surprisingly, it may be advantageous for the retailer to delegate the stewardship/assembly of probabilistic products to the supplier, which also benefits the supplier and consumers. PS can create value for all stakeholders by effectively mitigating supply chain inefficiencies.

Problem definition: Healthcare resources are often distributed unevenly, posing major challenges to the equitable provision of medical services. Governments and nonprofit organizations have initiated programs to support underresourced hospitals, but their effectiveness in improving hospital and physician performance remains unclear. This study addresses this gap by evaluating the impact of China’s medical poverty alleviation program (MPAP) on service efficiency and quality at an underresourced hospital. Methodology/results: We employ a staggered difference-in-differences design with matching to reduce imbalance between treatment and control departments. We find that MPAP significantly improves both efficiency and quality, reducing service time by 7.22% and the seven-day revisit by 3.76%. These gains are especially pronounced among less experienced physicians, physicians with prior experience serving as assistants, and physician-expert pairs with larger experience gaps. Managerial implications: Taken together, our analysis indicates that the benefits of MPAP vary across physicians and departments. These findings have practical implications for the design and implementation of MPAP. Specifically, the program should prioritize pairing external experts with less experienced physicians, assigning local physicians to assistant roles to facilitate hands-on learning, and targeting pairings with sufficiently large experience gaps. In addition, the program should focus on departments with higher technological complexity, greater illness severity, and more demanding diagnostic equipment.

Problem definition: Some governments have toughened traffic penalties for meal-delivery drivers, yet the number of meal delivery–related traffic incidents has not abated. These observations prompted two questions: Under a given penalty scheme, how should a profit-maximizing platform determine the delivery fee to charge its customers and the commission to pay its drivers? How should a welfare-maximizing government determine its penalty scheme to curb traffic incidents and improve the social surplus of all stakeholders? Methodology/results: We represent the system dynamics using a three-stage game-theoretic model, from which our equilibrium analysis yields three key insights. First, penalizing the platform for delivery-related incidents is more effective at reducing risky driving than penalizing drivers. Higher platform penalties discourage unrealistic delivery-time promises, allowing drivers to lower their driving speeds. By contrast, increasing driver penalties does not consistently curb risky driving, because the platform may offset penalties by raising driver commissions, which further incentivizes speeding. Second, lowering driver penalties expands the platform’s service area by encouraging drivers to accept longer-distance or lower-paying orders. Third, penalizing only the platform—rather than the drivers—maximizes total social surplus, provided the penalty is not set excessively high, which could undermine the balance between safety and market coverage. Our findings remain robust even when key modeling assumptions are relaxed separately. We illustrate our findings using data collected from a Chinese meal-delivery platform. Managerial implications: Governments should shift from penalizing drivers to targeting platforms, acknowledging their pivotal role in shaping driver behavior. This policy change would incentivize platforms to promote safe driving practices over speedy deliveries.

Problem definition: Data-driven models in machine learning have enabled efficient management of production systems. However, a majority of machine learning models are devoted to modeling the mean response or average pattern, and this is inappropriate for studying abnormal extreme events that are often of primary interest in aircraft manufacturing. Because extreme events from heavy-tailed distributions give rise to prohibitive expenditures in system management, sophisticated extreme models are urgently needed to analyze complex extreme risks. Engineering applications of extreme models usually focus on individual extreme events, and this is insufficient for complex systems with correlations. Methodology/results: We introduce an extreme spatial model for multioutput response control systems that efficiently captures the dynamics using a bilinear function on two spatial domains for control variables and measurement locations. Marginal parameter modeling and extremal dependence have been investigated. In addition, an efficient graph-assisted composite likelihood estimation and corresponding computational algorithms are developed to cope with high-dimensional outputs. The application to composite aircraft production shows that the proposed model enables comprehensive analyses with superior predictive performance on extreme events compared with canonical methods. Managerial implications: Our method shows how to use an extreme spatial model for predicting extreme events and managing extreme risks in complex production systems, such as aircraft. This can help achieve better quality management and operation safety in aircraft production systems and beyond.