Arnd Huchzermeier

Most prior academic research examines how firms respond to a single tariff episode. Yet firms increasingly operate in an environment characterized by recurring tariff shocks. Whether successive tariff episodes lead firms to continue diversifying their supply chains or instead prompt a different mode of adaptation remains largely unknown. We study how U.S. importing firms reconfigure their global supply chains across two major tariff episodes: Trade War 1.0 in 2018 and Trade War 2.0 in 2025. Using firm-product-level import transactions from Panjiva matched with Compustat, we track changes in sourcing locations, supplier networks, and shipment flows. We first show that both tariff episodes divert sourcing away from China toward a similar set of alternative countries. However, the magnitude of reallocation is substantially weaker during the second episode, and the nearshoring shift toward North America that emerged in the first episode is largely absent. We then document a fundamental shift in firms’ adjustment behavior across episodes. The initial tariff episode induces broad-based diversification: firms expand the number of sourcing countries and suppliers while reducing average shipment volumes per supplier, consistent with an explore-and-test response under uncertainty. In contrast, the repeated tariff episode does not further broaden supplier networks. Instead, firms reduce the number of active suppliers and concentrate sourcing within networks established during the first episode. Moreover, supply chain reconfiguration following the second episode occurs more rapidly and is driven primarily by supplier exits rather than new supplier entry. This pattern suggests that firms draw on knowledge accumulated during the first trade war to make more targeted and decisive adjustments. The consolidation effect is particularly pronounced for differentiated products, where coordination-intensive inputs are less amenable to fragmented supplier networks. Repeated tariff episodes do not lead to further diversification. Rather, an initial episode prompts firms to build supply chain optionality by expanding their sourcing networks, whereas subsequent episodes trigger selective pruning and consolidation within those expanded networks. These findings offer guidance for managers seeking to design resilient global supply chains in an era of recurring trade-policy uncertainty.

This case study on the MEDICON Pharmacy Consortium is embedded in a historical background of a founding family in the pharmacy retail sector that faces intense competition from online channels. Since customers may not actually need medication and marketing measures are strictly regulated, managing customers' loyalty is a challenge. The questions involved include: How do I keep customers active in local on-site pharmacies? When should a customer be declared inactive? How do I reactivate elapsed customers within a limited budget? The goal of this case study-and pioneering work, being the first of its kind in the German market-is to provide students with the decision-making skills required for sound, data-driven investments in direct marketing based on mathematical simulations. In particular, students are taught: (1) about the characteristics and unique circumstances of this highly contested and growing market; (2) how to evaluate marketing activities based on data science techniques as compared with past purchase observations; (3) how to identify the adjustments needed to implement successfully the standard Elsner et al. (2003) model in a given case; and (4) how to use simulation and optimization techniques to make the first-best decision about efficient use of a marketing budget. After discussing the literature on customer reactivation and churn management, we describe the adaptations necessary for successfully transferring approaches to the given practice case. This case is recommended for students taking courses in business administration, marketing, retail operations, strategic management, or business analytics that involve methods relying on data analysis.

This data-driven paper is a comprehensive analysis of the post-purchase behavior of individual activated and reactivated customers. The analysis is based on nearly 8 million instances of customer-specific transaction data that have been enriched with marketing performance measures, and it is made feasible by highly scalable software solutions. A brief overview of the test and control group design (32 focus groups were used) is followed by accounts of the database and the key performance indicators formed dynamically at the customer level with the aid of a modern data platform. Descriptive analysis of post-purchase behavior yields initial tendencies. However, varying group compositions preclude relying on such trends if the goal is to make valid management decisions in the spirit of customer centricity. Covariate balancing is used to reduce selection bias; altogether, 24 propensity score matchings (each based on 12 covariates) were successfully implemented for the different group combinations. This matching procedure "relativizes" the previously observed tendencies and enriches them with further insights. Thus difference-indifferences analyses were used to examine the balanced data sets for significant changes in post-purchase behavior. It was then possible to show that, contrary to common research assumptions, post-purchase behavior is not constant and varies from customer to customer. In fact, post-purchase sales may be increased by activating active customers and reactivating inactive customers. Finally, this study reveals-in the context of designing a resource-efficient marketing scheme-the possibility of its negative effects even on active, valuable customer groups.

This paper describes the journey of MEDICON, a medium-sized pharmacy consortium, to become a data-driven company. Although the pharmacy market in Germany is growing steadily, the existing business of established pharmacies has become increasingly threatened. Even this partially protected market is not immune to digitalization: encroachment by online pharmacies is underwritten by large investor funds, and legislation has encouraged people to start using e-prescriptions. The real-world example of MEDICON shows that it is possible to compete successfully with the pharmacy market’s big players by making comparatively small but smart investments. We outline the path from setting up a modern data platform all the way through simulations that lead to optimizing for a holistic marketing strategy. This pioneering work describes how a firm can identify its most valuable customers—and how it can activate (or re-activate) its customer base via efficient marketing measures that are based on a formalized customer migration model. The key insights detailed here are as follows. First, data enables proper targeting. A pharmacy must not only keep customers active; it must also activate the right customers in a targeted manner—especially in times of limited budgets. To do nothing is not an option, for then one quickly runs out of customers. Second, data enables rapid decision making. We implemented several test runs, in the spirit of “fail fast, learn fast”, in order to find out why half of the marketing budget is wasted and how best to orchestrate marketing measures. Our results show the advantages of eschewing gut decisions or proxy planning and of embracing simulations of marketing strategies with a clear, optimized target variable (viz., active customers). Third, post-purchase behavior matters, and it differs among customer groups. Implementing a modern data platform facilitates the long-term analysis of individual buying behavior, which has revealed that the post-purchase behavior of activated and re-activated customers differs from their previous buying behavior. We show how this success-relevant factor can be integrated into the underlying Elsner et al. (2003) model, and in a realistic scenario we find that such integration increases the (simulated) total gross profit by a factor of 7 or more.

This case study examines how a pharmacy consortium can create economic value by systematically reactivating inactive customers through data-driven direct marketing. The case is set in the MEDICON Pharmacy Consortium, a network of independent pharmacies operating in the highly regulated and increasingly competitive German pharmacy market. Facing rising costs, declining margins, and growing competition from online pharmacies, MEDICON’s management is challenged to justify marketing expenditures and shift decision-making from intuition to analytical control. The case describes the institutional environment of the German pharmacy sector, the strategic role of consortiums, and the managerial tension between professional self-perception as health counselors and the need for retail and marketing competence. It introduces a customer migration framework and a simulation-based marketing optimizer that allows managers to evaluate alternative direct marketing campaigns under budget constraints. Particular emphasis is placed on customer reactivation, heterogeneous post-reactivation behavior, and the profitability of coordinated portfolios of targeted campaigns. Designed for use in graduate-level courses in operations management, marketing analytics, and data-driven decision-making, the case enables students to analyze customer segmentation, assess marketing effectiveness, and explore the managerial implications of analytics-based control systems in a traditional service industry. The case also provides a platform for discussing the transfer of academic marketing models into real-world managerial practice.

In this study, we demonstrate the ability of a multimodal LLM to manage a virtual end-to-end diagnostic workflow. We develop and evaluate an autonomous agent in a high-fidelity medical simulation across four acute care scenarios. We compare this agent’s performance and behavior against over 14,000 simulation runs by medical students and an expert emergency room physician. We find that (1) a multimodal LLM can function as a competent virtual physician, successfully stabilizing patients and solving complex cases that require interpreting text, audio, and imaging in real time, closely mirroring most of the actions of the expert; (2) the AI agent matches or exceeds medical students in case completion rates and secondary outcomes such as time and diagnostic accuracy, though it engages in less patient communication than both students and the expert; and (3) the agent's evolving diagnostic beliefs exhibit value-of-information properties—front-loading high-yield tests, experiencing diminishing belief revisions over time, and producing properly-calibrated confidence estimates. These findings suggest that LLMs can orchestrate complex clinical workflows rather than merely execute isolated tasks. They also offer design implications for human-AI collaboration and diagnostic pathway design in low-resource and time-critical healthcare operations. Supplemental Video: https://tinyurl.com/llm-dynamic-medical-eval

We study a supply chain distribution system and investigate experimentally operations of markets where retailers can trade digital claims (tokens) on the supplier’s capacity. Subjects play the role of retailers, have heterogeneous valuations of goods, face random demands, and buy tokens on the supplier’s capacity. Following demand realization, retailers trade tokens with each other in markets implemented as double-sided, single-price, blind, batch auctions. We compare six behavioral treatments, featuring two wholesale prices and three market sizes. As expected, markets reduce leftovers and shortages. Interestingly, market-clearing prices are anchored to wholesale prices and do not signal the value of goods in large markets. Players deploy novel ordering and trading strategies that differ from the transshipment literature. We identify strategies by applying unsupervised machine learning algorithms. In one strategy, players buy a few claims and, after demand realization, use the market to satisfy it. Other players buy more claims than the maximum demand and, once demand is known, sell their excess on the market. Both strategies reduce costs from demand uncertainty but expose players to liquidity and mistakes risks. A third strategy, in which players order from the supplier initially as if expecting the market to be cleared cooperatively, is more profitable. This strategy diversifies demand and market risks. The introduction of markets causes the “pull-to-the-mean” effect and increases order variability. Thus, markets can cause the Bullwhip Effect. Retailers’ and the supply chain’s average profits are higher with markets, but suppliers with low wholesale prices suffer from lower revenues because of the pull-to-the-mean effect. This paper was accepted by Elena Katok, operations management. Supplemental Material: The data files are available at https://doi.org/10.1287/mnsc.2023.03771 .

Problem definition: The COVID-19 pandemic imposed unprecedented stresses on global supply chains (GSCs), compelling companies to reassess their supply chain structures and strategies. This crisis has also heightened awareness among businesses, consumers, and policymakers about the critical importance and far-reaching implications of GSC design and management. This unique moment presents a generational opportunity for Operations Management (OM) researchers to document and understand the ongoing restructuring of GSCs. Methodology/results: By analyzing microlevel data on U.S. customs import shipments (2019–2021), we uncover shifts in GSC strategies during the COVID-19 pandemic. Firms diversified suppliers within existing sourcing locations and reallocated volumes among them. Whereas dependence on China decreased, imports from other Asian nations like India and Vietnam, as well as North American countries like Canada and Mexico, increased. Industry-specific differences were pronounced, and a notable shift toward lower-frequency, higher-quantity shipments was also observed. Managerial implications: Beyond the challenges of COVID-19, recent years have witnessed other major supply chain disruptions, due to causes such as geopolitical tensions, natural disasters, and port worker strikes. We offer actionable insights for executives designing supply chain strategies to prepare for similar disruptions as they increase in frequency and severity. We identify future research avenues aimed at enhancing the resilience and adaptability of GSCs in a continuously evolving environment. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2024.0879 .