Abstract: In algorithm-augmented service contexts where workers have decision authority, they face two decisions about the algorithm: whether to follow its advice, and how quickly to do so. The pressure to work quickly increases with the speed of arriving customers. In this paper, we ask: how do workers use algorithms to manage system loads? With a laboratory experiment, we find that superior algorithm quality and high system loads increase participants' willingness to use their algorithm's advice. Consequently, participants with the superior algorithm make higher-quality recommendations than those with no algorithm (participants with the inferior algorithm make slightly lower-quality recommendations than those without). However, participants do not necessarily speed up by using algorithms' advice; their throughput times only decrease compared to the no-algorithm baseline when the system load is high and algorithm quality is superior, although participants would benefit from working faster in all treatments. This happens in part because participants in the high-load, superior-algorithm treatment serve customers more quickly than participants in the other treatments, conditional on using the algorithm. Participants in the high-load, superior-algorithm treatment work especially quickly in later periods as they increasingly default to their algorithm's advice. Our findings show that algorithms can have benefits for both decision quality and speed. Quality benefits come from workers' decision to use their algorithms' advice, while speed benefits depend on workers' algorithm use and the time they spend deliberating about their algorithm use. Ultimately, algorithm quality and system load are mutually reinforcing factors that influence both service quality and especially speed.

Abstract: We experimentally study dynamic resource allocation decisions using product development as the context. A product manager must accept or reject a series of design improvement opportunities, given a limited budget. Human subjects perform well when the cost-to-implement is fixed throughout the project. However, in a more complex setting where the cost increases for the latter half of the project, subjects' performance worsens substantially. We use the Strategy Frequency Estimation Method to analyze subjects' decision mechanisms, and find that many subjects are (a) mis-weighting future periods (underweighting in the simple case, overweighting in the complex) and (b) focusing on only the highest value opportunities. These heuristics perform poorly in the complex setting, leading to excess savings and are a counterproductive reaction to the cost increase. Top performers in the complex setting do well by decomposing the problem into two sub-problems resembling the simpler setting, which they can handle nearly optimally. In a second study, we test managerial interventions based on prompting this decomposition approach to improve performance in the complex setting. Merely prompting subjects to consider problem decomposition is largely ineffective. However, additionally sharing a "best practice" budget plan that gives information about how and why top performers decompose the problem significantly improves performance. Our results highlight when decisionmakers will perform well or poorly in a dynamic resource allocation problem, and shows effective ways to re-frame the problem and improve their performance. 

Abstract: Companies often use enterprise social media platforms (ESMPs) to promote helping behavior among knowledge workers and to improve their productivity. However, many ESMP implementations have had mixed success due to employee reluctance to engage with the platform (e.g. by requesting and giving help). Prominent ESMPs have added varying design features to promote helping behavior, but the effectiveness of individual design features remains unclear. Our paper studies this using a laboratory experiment that consists of two studies. In Study 1, we explore the effect of individual design features on a participant’s decision to make and answer help requests, and the impact that has on their performance. Our first two treatments vary the level of platform helping behavior visible to participants. Our third treatment sets helping goals that participants can complete by making and answering help requests and rewards them with private badges for doing so. We find that only the treatment with helping goals and badges results in a significant 5.5% increase in participant performance. It achieves this primarily by increasing a participant’s propensity to request help and, to a lesser degree, their propensity to answer help requests. In Study 2, we show that similar improvements can be achieved with just a helping goal that asks participants to help others (without a badge) . This speaks to the effectiveness of using goal setting motivation to leverage an individual’s intrinsic motivation to help others, even when it comes at a cost to themselves. Interestingly, we further show that helping goals that explicitly ask participants to help others also encourage participants to request more help, increasing the overall use of the platform. Based on our findings, we encourage companies to adopt ESMPs with badges for the helping goals they have. We also encourage ESMPs that do not have badges to consider adding them natively to their platforms.

MS Replication Project webpage

Abstract: Over the last two decades, researchers in operations management have increasingly leveraged laboratory experiments to identify key behavioral insights. These experiments inform behavioral theories of operations management, impacting domains including inventory, supply chain management, queuing, forecasting, and sourcing. Yet, until now, the replicability of most behavioral insights from these laboratory experiments has been untested. We remedy this with the first large-scale replication study in operations management. With the input of the wider operations management community, we identify 10 prominent experimental operations management papers published in Management Science, which span a variety of domains, to be the focus of our replication effort. For each paper, we conduct a high-powered replication study of the main results across multiple locations using original materials (when available and suitable). In addition, our study tests replicability in multiple modalities (in-person and online) due to laboratory closures during the COVID-19 pandemic. Our replication study contributes new knowledge about the robustness of several key behavioral theories in operations management and contributes more broadly to efforts in the operations management field to improve research transparency and reliability.

Abstract: Although innovation sharing between a buyer and a supplier—a common practice in the automotive industry—can increase the efficiency and total profit of a supply chain, suppliers are often reluctant to do so. Sharing innovations would leave the supplier in a vulnerable position if the buyer were to exploit the information (e.g., by resharing the supplier’s innovation with competing suppliers). Anecdotal evidence from automotive suppliers tells us that the nature of a supplier-buyer relationship for the most part depends on who (e.g., a short-run-focused procurement manager or a long-term-focused engineer) manages the relationship. In this paper, we examine how the allocation of decision rights to short-run- and long-run-focused employees affects collaboration between the firms. To accomplish this, we model a relationship between a supplier and a buyer where the buyer is a dual decision maker, consisting of long-run- and/or short-run-focused employees. We characterize the equilibrium of this model and show that the frequency of collaborative outcomes is lowest when the procurement manager has full control and highest when the engineer has full control or in a setup where employees can unilaterally enforce collaboration. A laboratory experiment confirms that collaborations occur more frequently when a manager with a long-term focus is in charge of or actively involved in decision making. Depending on how joint control is structured, joint control can help or hurt collaboration: Collaboration is highest when employees can unilaterally enforce it, whereas the random joint-control case results in collaboration that is as low as in the procurement manager control case.

Abstract:
Problem definition: Giving out a symbolic “supplier of the year” or “outstanding supplier” award can be beneficial for a buyer as it may incentivize a supplier to exert higher efforts. However, when a good supplier is scarce, the award announces which supplier is particularly good and may increase the cost of building and maintaining the relationship. This paper studies both positive and negative effects of a symbolic award and offers explanations on underlying behavioral mechanisms.
Academic/practical relevance: We show that symbolic awards can effectively incentivize suppliers to provide high effort, improving a buyer’s bottom line. This is particularly relevant in cases in which certain aspects of a buyer–supplier relationship are not contractible and suppliers have discretion over the quality provided. The award format significantly influences the award’s effectiveness.
Methodology: We develop a game-theoretical model that captures a supplier’s utility for the award in a competitive setting and test the predictions of the model with laboratory experiments.
Results: Our experimental results confirm that private symbolic awards have motivating effects and lead to higher buyer profits. When the awards are public, this profit premium diminishes as buyers pay higher prices to get the good suppliers. When the buyer is given the option to make the award public or private, buyers prefer that awards are public over private, anticipating a negative supplier response to their choice of the private award format.
Managerial implications: Expressing praise or gratitude for a supplier’s efforts can be highly beneficial for a buyer. However, when there is scarcity of good suppliers, buyers should expect increased competition and accompany the award with efforts to preserve the relationship. Finally, if buyers choose to offer a distinctive award format, private recognitions may be perceived as greedy or self-interested and backfire. 

Abstract: We report the results of a novel protocol for running online experiments using a combination of an online experimental platform in parallel with web-conferencing software in two formats—with and without subject webcams—to improve subjects’ attention and engagement. We compare the results between our online sessions with the offline (lab) sessions of the same experiment. We find that both online formats lead to comparable subject characteristics and performance as the offline (lab) experiment. However, the webcam-on protocol has less noisy data, and hence better statistical power, than the protocol without a webcam. The webcam-on protocol can detect reasonable effect sizes with a comparable sample size as in the offline (lab) protocol. 

Abstract: Entrepreneurial teams assign equity positions in their start-ups using a term sheet that details equity splits and the conditions for being granted those splits. The design of equity split agreements has attracted considerable attention in the entrepreneurial community, with no convergence on a single preferred contract form. This paper experimentally examines the effectiveness of different contractual arrangements, focusing in particular on the effects of contract form and contracting timing on founder effort and on the value of the venture. Our results suggest that performance improves with the incentive strength of the contract, but they question the conventional logic that this effect is causal. Instead, we suggest a novel causal sequence. Rather than the contract form being the primitive and the behavior the derived consequence, our results suggest the reverse. The differences in contract performance are driven primarily by the sorting of high contributors into nonequal contracts and of low contributors into equal contracts; that is, equal contracts are bad for team performance not because of their incentive strength but because of the founder types that adopt them. Taken together, these results suggest that both investors and founders should pay as much (or more) attention to personality type as they do to contract form.

Abstract: We conduct a field experiment where we vary both the presence of a gift-exchange wage and the effect of the worker’s effort on the manager’s payoff. Results indicate a strong complementarity between the initial wage-gift and the agent’s ability to “repay the gift”. We control for differences in ability and reciprocal inclination and show that gift-exchange is more effective with more reciprocal agents. We present a principal-agent model with reciprocal subjects that motivates our findings. Our results help to reconcile the conflicting evidence on the efficacy of gift-exchange outside the lab.

Abstract: We report the results of a randomized field experiment in the Philippines on the effects of two common anti-vote-selling strategies involving eliciting promises from voters. An invitation to promise not to vote-sell is taken up by most respondents, reduces vote-selling, and has a larger effect in races with smaller vote-buying payments. The treatment reduces vote-selling in the smallest-stakes election by 10.9 percentage points. Inviting voters to promise to “vote your conscience” despite accepting money is significantly less effective. The results are consistent with a behavioral model in which voters are only partially sophisticated about their vote-selling temptation.

Abstract: Suppliers are often reluctant to invest in capacity if they believe that they will be unable to recover their investment costs in subsequent transactions with buyers. In theory, a number of different contracts can solve this issue and induce first-best investment levels by the supplier. In this study, we investigate the performance of these contracts in a two-tier supply chain. We develop an experimental design where retailers and suppliers bargain over contract terms—and have the ability to make multiple back-and-forth offers—while also providing feedback on the offers they receive. One key result from our study is that an option contract and a service-level agreement are best at increasing first-best investment levels and overall supply chain profits. However, these same contracts also generate the largest inequity in expected profits between the two parties. We find that both of these results are driven by the bargaining tendencies of retailers and suppliers, which we refer to as “superficial fairness.” In particular, retailers and suppliers place more emphasis on negotiating the wholesale price, while partially overlooking any secondary parameter, such that final wholesale prices end up roughly halfway between the retailer’s selling price and the supplier’s production cost. We show that this bargaining behavior contributes to higher investment levels observed in the option contract and service-level agreement, along with the inequitable payoffs.

Abstract: The relationship between a buyer and its suppliers often relies on factors beyond the terms of a contractual agreement. Buyers can therefore benefit from identifying trustworthy suppliers. We argue that precontractual actions by a supplier, for example making costly buyer-specific investments without a long-term contract, can signal a supplier’s trustworthiness. We develop a theoretical model to reflect supplier trustworthiness, and determine when a buyer can benefit from identifying trustworthy suppliers. We show that costly relationship-specific investments can serve as a signal of trustworthiness, and that supply chain profits increase when trustworthy suppliers are able to identify themselves in this fashion. We demonstrate the importance of the signaling mechanism using laboratory experiments. The experimental results show that relationship-specific investments lead to more collaborative transactions, with buyers offering higher prices and suppliers returning higher-quality products. This results in increased profits for both buyers and suppliers. Additionally, we design a treatment which shuts down the signaling mechanism and show that the benefits of the buyer-specific investment are no longer present in this case. Finally, we show that the benefits of buyer-specific investments for both suppliers and buyers are strengthened when firms interact repeatedly.

Abstract: Bringing a new product to market involves both a creative ideation stage and an execution stage. When time-to-market constraints are binding, important questions are how to divide limited time between the two stages and who should make this decision. We introduce a laboratory experiment that closely resembles this setting: it features a product development task with an open design space, a downstream cost increase, and two development stages. We show that performance is significantly worse when designers choose for themselves when to transition from ideation to execution and that decision control explains a large share of performance variation even after controlling for individual differences. How the time is allocated between ideation and execution does not affect mean performance, but later transition increases risk. One driver of poor design outcomes in the designer-initiated transition regime are delays in physical construction and testing of designs. We show that such delays can be prevented by “nudging” designers toward early prototyping. However, the most important performance driver is the lack of task structure in endogenous regimes, which can be remedied by demanding a concrete, performance-oriented deliverable prior to a transition.

Abstract: Using coordination games, we elicit social norms directly for two different games where either an agreement to take the first best action has been reached or where no such agreement exists. We combine the norms data with separately measured choice data to predict changes in behavior. We demonstrate that including social norms as a utility component significantly improves predictive performance. Then we compare social norms to guilt aversion and lying aversion. We estimate that honoring an agreement in the double dictator game is worth giving up approximately 10% of total earnings and more than 120% in the Bertrand game. We show that informal agreements affect behavior through their direct effect on social norms as well as through an indirect effect on beliefs.

Abstract: We study experimentally bargaining in a multiple-tier supply chain with horizontal competition and sequential bargaining between tiers. Our treatments vary the cost differences between firms in tiers 1 and 2. We measure how these underlying costs in uence the efficiency, negotiated prices and profit distribution across the supply chain, and the consistency of these outcomes with existing theory. We find that the structural issue of cost differentials dominates personal characteristics in explaining outcomes, with profits in a tier generally increasing with decreased competition in the tier and increasing with decreased competition in alternate tiers. The Balanced Principal model of supply chain bargaining does a good job explaining our data, and outperforms the common assumption of leader-follower negotiations. We find a significant anchoring effect from a firm's first bid but no effect of the sequence of those bids, no evidence of failure to close via escalation of commitment, and mixed results for a deadline effect. We also find an interesting asymmetry between the buy and sell sides in employed bidding strategy. All firms make predominantly concessionary offers after the initial anchor, however sell side firms that engage in aggressive anti-concessionary bidding successfully increase prices while not compromising closure rates. Buy side firms achieve much smaller price changes from anti-concessionary tactics, and risk reduced closure, yielding no net benefit.

Long Version Working Paper

Abstract: A large and growing literature has demonstrated that imposing control on agents has the potential to backfire, leading agents to withhold effort. Consistent with principles of procedural fairness, we find that the way in which control is imposed—in particular whether control is imposed symmetrically on both principals and agents and whether both parties have a say in whether control is imposed—affects how agents respond to control. In our setting, control leads agents to withhold effort only when procedural fairness concerns are ignored and control is imposed unilaterally with an asymmetric effect on the agent.

Abstract: Recent laboratory evidence suggests that personality traits, in particular social preferences, may affect contractual outcomes under moral hazard. Using the British Workplace Employment Relations Survey 2004 we find that behaviour of employers and employees is consistent with the presence of gift-exchange motives: firms that screen applicants for personality are less likely to pay low wages and more likely to provide (non-pecuniary) benefits. Firms likewise benefit from employee screening, as they can implement more team-working and are generally more successful. Other human resource management practices only poorly predict these patterns. Moreover, there is no association between dismissals and personality tests, indicating that personality tests do not merely improve the fit between applicant and employer. Hence, we conclude that motivation based on gift-exchange motives is a plausible explanation for our results.

Abstract: Using data from an anti-vote-buying field experiment we conducted in the Philippines, we report and validate a proxy measure for vote-selling. We demonstrate that our proxy measure, vote-switching, changes as expected with voter preferences and monetary offers from candidates. Voters are less likely to vote for someone different than their initial preference the larger the favorability rating difference between the preferred and alternative candidates. Similarly, vote-switching increases the more money the alternative candidate offers compared to the preferred candidates. We also describe the effects of the promise-based interventions on vote-switching, reported in full in a companion paper.

Abstract: We study theoretically and empirically the consumption of access services. We demonstrate that consumption is affected by contract structure (pay-per-use versus three-part tariffs) even if the optimal consumption plans are identical. We find that, although there is extensive individual heterogeneity, on average, consumers' choices follow a structure that is similar to a nearly optimal heuristic and correctly react to imbalances between the number of free calls and call opportunities remaining. However, consumers use the free units too quickly, leading to overconsumption and lost surplus. These errors are partially driven by mistaken beliefs about the value distribution. We also measure subjects' willingness to pay for a contract with free access units, and we find that nearly half of subjects are willing to pay at least the full per-unit price, with a substantial fraction willing to overpay. In response, the optimal firm strategy offers a three-part tariff at a very small discount, which increases revenue by 8%–14% compared to only offering a pay-per-use contract.

Abstract: We investigate how the convexity of a firm's incentives interacts with worker overconfidence to affect sorting decisions and performance. We demonstrate, experimentally, that overconfident employees are more likely to sort into a nonlinear incentive scheme over a linear one, even though this reduces pay for many subjects and despite the presence of clear feedback. Additionally, the linear scheme attracts demotivated, underconfident workers who perform below their ability. Our findings suggest that firms may design incentive schemes that adapt to the behavioral biases of employees to "sort in" ("sort away") attractive (unattractive) employees; such schemes may also reduce a firm's wage bill. 

Abstract: We solve for the optimal contract when agents are reciprocal, demonstrating that generous compensation can substitute for performance-based pay. Our results suggest several factors that make firms more likely to use reciprocal incentives. Reciprocity is most powerful when output is a poor signal of effort and when the agent is highly reciprocal and/or productive. Similarly, reciprocal incentives are attractive when firm managers have strong incentive pay and discretion over employee compensation. While reciprocal incentives can be optimal even when identical firms compete, a reciprocity contract is most likely when one firm has a match-specific productivity advantage with the agent.

Abstract: The shortage of transplant kidneys has spurred debate about legalizing monetary payments to donors to increase the number of available kidneys. However, buying and selling organs faces widespread disapproval. We survey a representative sample of Americans to assess disapproval for several forms of kidney market, and to understand why individuals disapprove by identifying factors that predict disapproval, including disapproval of markets for other body parts, dislike of increased scope for markets and distrust of markets generally. Our results suggest that while the public is potentially receptive to compensating kidney donors, among those who oppose it, general disapproval toward certain kinds of transactions is at least as important as concern about specific policy details. Between 51% and 63% of respondents approve of the various potential kidney markets we investigate, and between 42% and 58% want such markets to be legal. A total of 38% of respondents disapprove of at least one market. Respondents who distrust markets generally are not more disapproving of kidney markets; however we find significant correlations between kidney market disapproval and attitudes reflecting disapproval toward certain transactions—including both other body markets and market encroachment into traditionally nonmarket exchanges, such as food preparation.

Abstract: Recent papers have demonstrated that the way people acquire information about a decision problem, by experience or by abstract description, can affect their behavior. We examined the role of experience over time in the emergence of the Gambler's Fallacy in binary prediction tasks. Theories of the Gambler's Fallacy and models of binary prediction suggest that recency bias, elicited by experience over time, may play a significant role. An experiment compared a condition where participants sequentially predicted the colored outcomes of a virtual roulette wheel spin with a condition where the wheel's past outcomes were presented all at once. In a third condition outcomes were presented sequentially in an automatic fashion without intervening predictions. Subjects were yoked so that the same history of outcomes was observed in all conditions. The results revealed the Gambler's Fallacy when outcomes were experienced (with or without predictions). However, the Gambler's Fallacy was attenuated when the same outcomes were presented all at once. Observing the Gambler's Fallacy in the third condition suggests that the presentation of information over time is a significant antecedent of the bias. A second experiment demonstrated that, while the bias can emerge with an all-at-once presentation that makes recent outcomes salient (Burns & Corpus, 2004), the bias did not emerge when the presentation did not draw attention to recent outcomes.

Abstract: We conduct a field experiment in a large real-world social network to examine how subjects expect to be treated by their friends and by strangers who make allocation decisions in modified dictator games. Although recipients' beliefs accurately account for the extent to which friends will choose more generous allocations than strangers (i.e., directed altruism), recipients are not able to anticipate individual differences in the baseline altruism of allocators (measured by giving to an unnamed recipient, which is predictive of generosity toward named recipients). Recipients who are direct friends with the allocator, or even recipients with many common friends, are no more accurate in recognizing intrinsically altruistic allocators. Recipient beliefs are significantly less accurate than the predictions of an econometrician who knows the allocator's demographic characteristics and social distance, suggesting recipients do not have information on unobservable characteristics of the allocator.

Abstract: We conducted online field experiments in large real-world social networks in order to decompose prosocial giving into three components: (1) baseline altruism toward randomly selected strangers, (2) directed altruism that favors friends over random strangers, and (3) giving motivated by the prospect of future interaction. Directed altruism increases giving to friends by 52% relative to random strangers, whereas future interaction effects increase giving by an additional 24% when giving is socially efficient. This finding suggests that future interaction affects giving through a repeated game mechanism where agents can be rewarded for granting efficiency-enhancing favors. We also find that subjects with higher baseline altruism have friends with higher baseline altruism.

Abstract: In many contexts we are warned against engaging in risky behavior only after having past safe experience. We examine the effect of safe experience on a warning’s impact by comparing warnings received after having safe personal experience with those received before people start making choices. A series of five experiments studies this question with a paradigm that combines both descriptive information (i.e. the warning) and experiential information (safe outcomes). The results demonstrate two separate advantages to an early warning that go beyond the warning’s mere informational content. When an early warning coincides with the beginning of a decision-making process, the warning is both weighted more heavily in future decisions (the Primacy Effect) and induces safer behavior that becomes the status quo for future choices (the Initial History Effect). While both effects operate indirectly through choice inertia, the primacy effect also operates directly on choices. This pattern of behavior is inconsistent with the “ideal” Bayesian for whom the order of information revelation does not influence subsequent behavior. The effect was robust across settings with and without forgone payoffs and when the consequences for risk taking are delayed until the end of the experiment. The results imply that, even after being adequately warned, some people may continue to take risks simply because they incurred good outcomes from the same choice in the past. Implications for policy and theory are discussed.

Description: Behavioral Operations Management (BOM) incorporates insights from psychology and behavioral economics to study how individuals make decisions in an operational context. Examples of important behavioral factors include bounded rationality and decision heuristics, folk intuitions about random processes, preference regularities such as loss aversion and reference dependent preferences, and interpersonal factors such as trust and fairness. Behavioral research frequently contrasts observed behavior with the predictions of `standard' analytical models. Two major goals of BOM are to provide a better understanding of (and make better predictions about) behavioral regularities, and to provide guidance to firms on how to design mechanisms that will lead to better decisions and improved performance. Much of the existing BOM research has utilized laboratory experiments; however there are a growing number of behaviorally-influenced theoretical models, empirical research and field experiments. The field has grown tremendously in the last fifteen years - leading to several special issues in top journals (Management Science, Manufacturing & Service Operations Management, Production and Operations Management and Journal of Operations Management), an annual Behavioral Operations Research Conference, and INFORMS and POMS sections. This book aims to be a comprehensive resource on BOM research for both those active in the field and those new to it. 

Excerpt: Strategic interactions – circumstances where multiple parties, potentially with imperfectly aligned preferences, make decisions in a decentralized manner that affect some or all of the parties – are important for many different aspects of operations management. Firms in a supply chain need to set prices, decide on capacity, provide quality goods, etc. Customers joining a queueing system may need to anticipate the abandonment decisions of others. Managers in a production system can choose various incentive and monitoring policies to encourage workers to maintain productivity and quality. 

This chapter will discuss three different sets of analytical tools commonly used in operations management to understand strategic interactions: game theory, principal–agent theory, and bargaining theory. Each tool is useful to understand different aspects of a strategic interaction. Game theory typically takes as given the nature of the interaction (e.g. the set of possible actions for each party and the payoff consequences of those actions) and asks what strategy we should expect each party to employ. Principal–agent theory considers cases where one party can shape the structure of the interaction (e.g. by establishing financial incentives for various outcomes or offering a menu of options that the other party can choose from) and asks what is the optimal structure to achieve that party’s goal. Bargaining theory asks how the surplus from a transaction will likely be divided between the parties, which in turn is an important determinant of the choices those parties might take to increase the surplus.

Abstract: We experimentally study dynamic resource allocation decisions using product development as the context. A product manager must accept or reject a series of design improvement opportunities, given a limited budget. Human subjects perform well when the cost-to-implement is fixed throughout the project. However, in a more complex setting where the cost increases for the latter half of the project, subjects' performance worsens substantially. We use the Strategy Frequency Estimation Method to analyze subjects' decision mechanisms, and find that many subjects are (a) mis-weighting future periods (underweighting in the simple case, overweighting in the complex) and (b) focusing on only the highest value opportunities. These heuristics perform poorly in the complex setting, leading to excess savings and are a counterproductive reaction to the cost increase. Top performers in the complex setting do well by decomposing the problem into two sub-problems resembling the simpler setting, which they can handle nearly optimally. In a second study, we test managerial interventions based on prompting this decomposition approach to improve performance in the complex setting. Merely prompting subjects to consider problem decomposition is largely ineffective. However, additionally sharing a "best practice" budget plan that gives information about how and why top performers decompose the problem significantly improves performance. Our results highlight when decisionmakers will perform well or poorly in a dynamic resource allocation problem, and shows effective ways to re-frame the problem and improve their performance.