Knowledge that Transforms

Recent innovations have driven a steady increase in online marketplace transactions. To remain competitive, numerous marketplace platforms and independent data providers offer Competitive Intelligence Services (CIS), enabling sellers to explore not only their market potential but also that of their competitors. In this article, we employ a two-period game-theoretic approach to competitive learning to analyze the impact of CIS on participants with varying market shares in an online marketplace. In the presence of noisy demand signals, the online platform benefits from offering CIS to both sellers. This is because high demand noise makes demand exploration difficult for each seller in the first period. Consequently, price competition under poor knowledge of the price-demand relationship in the second period leads to a lower payoff for each seller as well as the platform. However, as demand uncertainty decreases, the platform prefers to induce CIS subscription exclusively for the seller with the larger market share. This scenario leads to signal-jamming behavior between the sellers, which results in a win-win-win situation for both sellers and the platform provider. Finally, we consider various model extensions and discuss the managerial implications for the design and regulation of competitive intelligence services in online marketplaces.

Retail recommendation systems increasingly operate as real-time decision engines that must personalize suggestions while respecting operational constraints such as inventory availability, category rules, and promotion policies. This is especially challenging in basket-based retail because transactions are set-valued and the observed checkout order is operational rather than behavioral. We study a governed agentic AI system for basket recommendation in which a Bayesian consumer world model serves as the agent’s internal state. The model maintains calibrated beliefs over latent shopping profiles and updates them online as items are observed while representing basket context in an order-invariant way. The agent then selects recommendation slates under explicit governance, combining interpretable control levers (e.g., profile-context trade-off, bounded exploration) with operational guardrails and feasibility masking (e.g., in-stock status, category, promotion eligibility). Using large-scale grocery transaction data, we evaluate a framework in both offline next-item prediction and an operations-coupled simulator with inventory and promotion dynamics. The agent achieves a higher hit rate than a nonagentic variant as well as various strong item–item baselines under a common holdout protocol. In the simulation, ranking accuracy changes little, yet the agent delivers substantial gains in revenue and inventory productivity by steering demand toward feasible complements and enabling controlled substitution when constraints bind. This highlights an operations insight: under binding feasibility, decision quality can improve through constraint-aware substitutions and inventory coupling even when conventional ranking metrics remain unchanged. Overall, the results show how basket-aware demand models can be deployed as governed, agentic policies that coordinate personalization with operational objectives.

Competitive intelligence is essential for operations management decision-making. Beyond traditional offline information channels, firms increasingly gather online data and resources to generate comprehensive competitive intelligence. This study derives competitive intelligence in large markets by developing an interpretable machine learning framework that integrates multifaceted user behavior data, including user favorites, user-commented products, and user textual comments. Considering the complementary nature of these data sources, we first combine latent features derived from user favorites and user-commented products to improve submarket inference. Using these inferred submarkets as supervised signals, we connect user-commented products and associated textual comments to uncover consumer perceptions. We estimate the model using multifaceted data on online user behavior in the automotive domain. The results demonstrate that our model effectively improves submarket identification, captures consumer perceptions, and predicts competitive positions for new entrants. The derived competitive intelligence helps managers make more informed decisions in product operations and marketing strategies.

Generative artificial intelligence (GenAI) is shifting from conversational assistants toward agentic systems—autonomous decision-making systems that sense, decide, and act within operational workflows. This shift creates an autonomy paradox: as GenAI systems are granted greater operational autonomy, they should, by design, embody more formal structure, more explicit constraints, and stronger tail-risk discipline. We argue that stochastic generative models can be fragile in operational domains unless paired with mechanisms that provide verifiable feasibility, robustness to distribution shift, and stress testing under high-consequence scenarios. To address this challenge, we develop a conceptual framework for assured autonomy grounded in operations research (OR), built on two complementary approaches. First, flow-based generative models frame generation as deterministic transport characterized by an ordinary differential equation, enabling auditability, constraint-aware generation, and connections to optimal transport, robust optimization, and sequential decision control. Second, operational safety is formulated through an adversarial robustness lens: decision rules are evaluated against worst-case perturbations within uncertainty or ambiguity sets, making unmodeled risks part of the design. This framework clarifies how increasing autonomy shifts OR’s role from solver to guardrail to system architect, with responsibility for control logic, incentive protocols, monitoring regimes, and safety boundaries. These elements define a research agenda for assured autonomy in safety-critical, reliability-sensitive operational domains.

For omnichannel multi-brand retailers, store openings are consequential strategic decisions. Beyond whether to open a store, firms must choose a format, select product categories, and decide how to support them in-store. To inform these decisions, we study the openings of two large experience-centric stores and one small convenience-centric store operated by an omnichannel consumer electronics retailer. Using a staggered difference-in-differences model, we estimate the impact of store openings on retailer performance while accounting for product-category heterogeneity. We find that all three openings reduce online net revenue (online purchases minus returns), contrary to halo-effect expectations. However, only the large experience-centric stores offset these losses, increasing total net revenue (online plus in-store purchases, net of returns) by 21% to 23% in the short term, with effects that grow over time. The small convenience-centric store does not generate such gains. Total net revenue uplift also varies substantially across product categories within experience-centric stores. To explain this heterogeneity, we combine category-level sales data with survey-based measures of perceived in-store utility across three customer journey stages: information search, fulfillment, and returns. Using the retailer's classification, we distinguish between destination categories—higher-priced, complex products that motivate store visits (e.g., TVs)—and accessory categories—lower-priced complementary products (e.g., earbuds). For destination categories, variation in perceived in-store utility, especially at the information-search and fulfillment stages, explains differences in total net revenue uplift. For accessory categories, such variation is not statistically associated with revenue gains. These findings show that store format alone is not sufficient: the effectiveness of store openings depends on how well in-store capabilities align with category-specific customer needs across the customer journey, highlighting the importance of prioritizing destination categories while reconsidering the role of physical stores for accessory categories.

Consumers often purchase access to a digital service by paying an upfront fee, and then consume the service over a period of time. In this article, we examine the implications of such temporal separation of purchase and consumption on a user’s consumption choices and on the firm’s optimal demand management strategy. Relying on behavioral economics and consumer behavior literature, we develop a formal microfounded model of a user’s decision calculus, and use it to derive the implied demand function and thus analyze the firm’s optimal decisions. In contrast to the classical recommendation to pursue admission control through higher prices as a means to manage demand for the digital service, we find that when mental accounting bias is a key driver of consumer choices, it might be optimal for a firm to pursue consumption control through lower prices. These results are robust when quality is endogenized, capacity is constrained, subscription duration is finite, and in the presence of a two-part tariff. We translate our findings into a conceptual framework for digital service management that characterizes the optimal demand management strategy along two key dimensions: the strength of the consumer bias and the cost of servicing demand. When these factors are significant, firms need to employ a combination of admission control and consumption control so as to manage congestion and maintain profitability.

Follow-up appointments are crucial for maintaining continuity of care, yet patients often encounter various barriers to accessing these services. In this study, we examine the potential of telemedicine applications for follow-up care (tele-follow-up), focusing on its impact on care access, service efficiency, and care quality. By collaborating with a large hospital that implemented tele-follow-up services across departments over time, we employ a staggered difference-in-differences design to identify the causal effects of tele-follow-up services. Our findings indicate that the adoption of tele-follow-up services increases total follow-up volume by 48.14%. Notably, we identify positive spillover effects on traditional onsite care, with onsite follow-up visits and initial visits increasing by 14.11% and 7.90%, respectively. We further investigate the mechanism underlying these results. On the demand side, patients with higher costs of accessing onsite follow-up care, such as those living in rural areas or with comorbidities, exhibit greater demand elasticity following the availability of the tele-follow-up channel. On the supply side, dedicating the telemedicine channel exclusively to follow-up care enhances physicians’ efficiency by enabling more focused practice across online and in-person work shifts. The increased access to follow-up services enabled by telemedicine further translates into better patient outcomes, as evidenced by a significant reduction in readmission rates. Our study demonstrates the value of tele-follow-up services and offers practical insights for healthcare decision-makers seeking to leverage digital health to enhance continuity of care.

A fundamental principle in operations management holds that increasing the number of servers reduces delays in service systems. To date, no mechanism has been identified that could reverse this effect. We propose, however, that risk aversion can cause increased service capacity to intensify congestion. We study an unobservable <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline" overflow="scroll"> <mml:mi>M</mml:mi> <mml:mo>/</mml:mo> <mml:mi>M</mml:mi> <mml:mo>/</mml:mo> <mml:mi>s</mml:mi> </mml:math> queue where risk-averse customers choose whether to join based on their anticipated waiting time. We show that, in equilibrium, demand, expected waiting time, expected sojourn time, and the probability of waiting all increase with the number of servers, and that these effects are stronger for more risk-averse customers. We further uncover the mechanism behind this phenomenon: adding servers makes delays less risky (in the sense of second-order stochastic dominance), which increases the sensitivity of demand to capacity as customers become more risk-averse. These patterns are more prevalent in small systems and fade as the system grows. They can also persist when customers differ in their degree of risk aversion, when capacity is increased by raising service speed, and when the system is observable. Our findings reveal a novel trade-off created by customer risk aversion: expanding capacity attracts more customers, but also exacerbates congestion. A manager aiming to reduce waiting times may therefore prefer to de-pool service capacity instead of following the standard approach of pooling, while increasing capacity in the resulting smaller systems to preserve total throughput. When the objective is to maximize profitability, our results further suggest that the cost of additional servers may be offset by the associated increase in revenue when customers are sufficiently risk-averse.

Motivated by ticket pricing challenges facing live event managers, we study how to maximize revenue when setting prices for multiple ticket categories with interdependent demand, realistic capacity constraints, and pricing restrictions. We define this decision problem as the Event Ticket Pricing (ETP) problem and formulate it as a constrained nonlinear optimization model. To examine how the nature of product differentiation (i.e., differentiation across ticket categories) affects seat portfolio pricing and revenue outcomes, we analyze the ETP problem under two demand specifications: vertically and horizontally differentiated ticket categories. For each case, we isolate the role of different constraints by developing and solving a sequence of problems with progressively added restrictions. This approach enables us to identify structural properties that either fully characterize the optimal solution or guide the development of efficient algorithms. Under vertical differentiation, we prove that the optimal solution has a sold-out threshold structure , with only high-quality categories sold out. The treatment of partially sold products depends on the active constraints. Sales vanish under capacity constraints alone but become positive once pricing restrictions are introduced, producing solutions distinct from the unconstrained benchmark. Under horizontal differentiation, the threshold structure persists, but the addition of constraints breaks the well-known “equal markup” rule, yielding prices that are non-increasing in quality. Finally, we study how flexible adjustments in seat capacities reshape optimal outcomes. These results underscore the importance of incorporating realistic constraints and offer practical guidance for live event planners.

This work addresses the operational conflicts between visibility-driven mobilization and cost efficiency in disaster management scenarios involving wildfires. Using official wildfire reporting on the social media platform Twitter (now X), we develop a temporal gravity model to extract a signal of public attention for California wildfires (2007–2021) without the “noise” of spurious content. Interpreting this signal through the lens of behavioral disaster management operations, our analysis finds a “Visibility-Efficiency Paradox.” This paradox shows that while social media visibility functions as a potent mobilization signal to the general public during wildfires and is associated with greater resource deployment, it simultaneously correlates with reduced cost efficiency under high resource use loads. We identify resource saturation as a boundary condition where heuristic signals appear to shift from valuable inputs to potential stressors. These findings challenge the assumption that high visibility of responders in a wildfire emergency is a direct proxy for operational urgency and effectiveness. We propose actionable strategies, including reverse audits and decoupling, in order to help counteract salience bias; thus, highlighting the potential for algorithmic governance to align public attention with sustainable resource management.