Knowledge that Transforms

Peer-to-peer (P2P) platforms facilitate the direct exchange of goods, services, or financial transactions between individuals without the involvement of intermediaries. To maintain trust among their user bases, these platforms must implement stringent access controls to determine which users are eligible to participate in their digital marketplace. We argue that when a platform transitions from private to public ownership, it may be incentivized to strategically lower its access standards and admit users who might otherwise have been deemed unqualified. Lowering standards before an initial public offering (IPO) can enable platforms to rapidly increase their user base—and, consequently, enhance their perceived valuation—which could appeal to stock investors and positively influence the IPO price. While this strategy may bolster short-term growth, it could be costly to the platform’s user base. We support this hypothesis using data from two major P2P lending platforms—one that went public and one that remained private. Using a difference-in-differences analysis, we find that the platform preparing for an IPO admitted borrowers who exhibited higher risk levels. Additionally, lenders, in some cases, did not effectively screen out these subpar borrowers and ended up issuing loans to them, leading to higher default rates and lower returns for the lenders. This effect was particularly evident for a specific segment of lenders—namely, those who brokered small-valued loans and loans for necessary purchases (e.g., health emergencies). By contrast, lenders who screened for large-valued loans or loans for discretionary expenditures (e.g., vacations or weddings) were more successful in screening out these subpar borrowers and not issuing loans to them. These results fill a gap in the operations management literature on platform governance by integrating capital-raising objectives into the discussion of access control and operational screening. Our study highlights the nuanced trade-off between quality control and user expansion during capital-raising events, emphasizing both the opportunities and potential inefficiencies that arise in this context.

Dynamic treatment regimes improve health outcomes by tailoring each treatment to a patient’s evolving condition, but they also allow providers to learn and game the system over time. How should insurers pay? We study this new class of reimbursement problems, where the provider can privately learn and manipulate the progression of the patient’s condition. (i) We characterize the optimal payment policy: it internalizes two intertemporal effects of each treatment, and rewards provider honesty with incentive pay; moreover, it admits a simple implementation of risk-adjusted cost-sharing policy. (ii) We show that, ignoring dynamic learning and gaming, the existing payment models may have overestimated the harm of information asymmetry. Using the optimal policy, insurers only need to pay for initial private information; they can exploit provider uncertainty and elicit future private information at no cost. (iii) Our study informs U.S. healthcare payment reform with new insights; using two sets of real data, our study also quantifies when and why the optimal policy outperforms the existing ones. By highlighting the critical role of dynamic learning and gaming, this study advances our understanding of healthcare payment theory and practice.

Work Experience Samuel Curtis Johnson Graduate School of Management, Cornell University (06/2004 – now) Kelley School of Business, Indiana University (08/2002 – 05/2004) SmartOps Corporation (02/2002 – 08/2002) Maxager Technology Inc. (02/2000 – 11/2001) Schlumberger Ltd. (09/1997 – 02/2000) Intel (09/1995 – 11/1995) Journal Publications 1. “Understanding Online Hotel Reviews through Automated Text Analysis” with Hyun Jeong Hun, Shawn Mankad, and Joel Goh. Service Science, To Appear. 2. “Self-selecting Priority Queues with Burr Distributed Waiting Costs” with Vidyadhar Kulkarni. Production and Operations Management, To Appear. 3. “Impact of Information Sharing, Random Yield, Correlation, and Lead Times in Closed Loop Supply Chains” with Takamichi Hosoda and Stephen M. Disney. European Journal of Operational Research, Vol. 246, 2015, pp. 827-836. 4. “Impact of Information Errors on Supply Chain Performance” with Jin Kyung Kwak. Journal of the Operational Research Society, Vol. 66, No. 2, 2015, pp. 288-298. 5. “Transfer Pricing and Sourcing Strategies for Multinational Firms” with Masha Shunko and Laurens Debo. Production and Operations Management, Vol. 23, No. 12, 2014, pp. 2043-2057. 6. “Optimal Selection of Line Extensions: Incorporating Operational, Financial, and Marketing Constraints” with Liying Mu, Milind Dawande, and Chelliah Sriskandarajah. IEEE Transactions on Engineering Management, Vol. 61, No. 4, 2014, pp. 738-754. 7. “Concierge Medicine: Applying Rational Economics to Health Care Queuing” with Vidyadhar Kulkarni. Cornell Hospitality Quarterly, Vol. 55, No. 3, 2014, pp. 314-325.

Maintaining a high service level with customers involves a fundamental tradeoff between investing in inventory and investing in expediting the shipping of the additional units needed to achieve that service level. We use a multi-method approach to show how and when ordering decisions are influenced by different levels of expediting costs and target service levels. First, we derive a mathematical model that provides a closed-form solution to this tradeoff. Second, we run a behavioral study to show how increments in the expediting cost and the target service level impact buyers’ ordering behavior. Results show that ordering decisions are influenced mainly by the expediting cost. Our econometric estimations provide a generalization of the pull-to-center effect to a setting with expedited shipping. Moreover, we find that buyers adjust their orders only when facing a high target service level. To reduce the observed behavioral biases, we propose that managers can increase the salience of key performance metrics in the buyers’ decision-making process. We test for the role of salience with a second behavioral study and a behavioral model. Results show that the extent to which salient information helps improve buyers’ ordering decisions depends on the level of the expediting cost. Interestingly, our behavioral model highlights how ordering decisions improve not by eliminating people’s biases but by amplifying some of those biases. We contribute to the literature on expedited shipping and behavioral operations, and provide practical recommendations for how managers can improve ordering decisions.

Disturbances in production along with volatile demand have raised concerns over shortfalls in the global supply chain and prompted the need to build a more diversified supply chain with competitive suppliers. This research investigates the impact of disturbances on a two-tier supply chain network with asymmetric competing firms. We establish the equilibrium in a unique structure that represents the maximum set of profitable upstream supply paths achievable through competition and exhibits stability under specific conditions. We evaluate the efficiency of the supply chain configuration by a shortfall problem and solve it with an adapted pseudoflow algorithm that efficiently identifies the mismatches between shortfalls and capacity surpluses in the multitier network. The parametric analysis reveals that the disturbance loss can be significantly offset by supplier competition, although the marginal benefit of competition decreases rapidly with the number of suppliers. Furthermore, shortfalls could be magnified by network asymmetries that increase configuration inefficiency, and supply chain performance could be improved by pushing high-cost firms to cease production. Simulation results indicate that the supply chain with a moderate level of competition and a balanced configuration can be robust against disturbance and demand volatility.

We study a multi-server queueing system where a customer is satisfied (and generates a unit revenue) if their queueing time is at most a given constant. If the queueing time of the admitted customer exceeds this constant, the customer gets served, but is unsatisfied and generates no revenue. Such queueing systems arise in the context of modeling service systems where excessive delays are of concern. A key challenge is how to design an admission control policy to maximize the number of satisfied customers per unit time in the long run, assuming that we can observe the number of customers in the system at any time. We call this the binary reward structure system and show that a threshold-type admission policy is optimal. The optimal threshold policy has to be computed numerically. Hence we propose a square-root admission policy to approximate the optimal admission control policy, and compare the performance of these two policies. We derive an analytical upper bound on the performance of optimal admission control policy by deriving an optimal admission policy assuming we have full information over the queueing time of the admitted customers. This is equivalent to a queueing system where customers abandon the queue (i.e., leave without service) if their queueing time exceeds the given constant. We demonstrate that the optimal policy that includes customer abandonment, or alternatively, the optimal policy under full information, the optimal threshold policy, and the square-root admission policy, all exhibit identical performance in the asymptotic regions of the parameter space. Our numerical results indicate that the worst optimality gap of the square-root admission policy is within 3.9% of the optimal revenue, and implementing the square-root admission policy in the observable queueing system leads to a revenue loss that is at most 5.6% of the maximum possible revenue rate in the full information system. We also compare the binary reward structure with the more common linear reward structure where the system incurs holding cost per unit queueing time per customer. In addition, we also show that the analysis based on queueing time is applicable to the system time as well.

Considering the real-world situations where a customer’s purchase choices in previous stages can influence the prices she encounters in subsequent stages, this research examines the multiproduct price optimization problem under a multistage choice model. Particularly, the seller commits to a multistage pricing policy and determines product prices based on the customer’s purchase history, and the customer makes purchase decisions such that the total expected utility is maximized. We show that the pricing problem has a unique optimal solution under some mild conditions and the optimal solution satisfies a modified equal adjusted markup property. Based on the property, the problem can be solved efficiently by reducing it to a single-dimensional search problem. Moreover, the optimal pricing policy has an important property, namely, the product with a higher adjusted markup in earlier stages should always lead to lower prices in subsequent stages. We also show that compared to customers who are myopic, the seller should offer higher first-stage prices and lower second-stage prices to forward-looking customers, which will lead to a higher profit. Numerical analyses are also conducted to demonstrate the above results.

Online retail platforms are increasingly challenged by the proliferation of low-quality products, which may damage their reputation and sales. To address this problem, we propose a system architecture to proactively identify products that are likely to go “out of favor.” Our approach uses historical data to extract useful information from customer ratings and textual reviews. Available data are fed into a state-of-the-art deep learning sequence model to forecast future ratings. We then analyze rating trends, extracting hyperparameters that a binary classifier uses to label products as “out-of-favor” or not. We tested this system on an Amazon dataset comprising nearly 800,000 observations across 2826 electronics products. Our results show that the Long Short-Term Memory (LSTM) model excels in forecasting future product ratings compared to other benchmarks. Ablation analysis shows sentiment-related features significantly improve rating forecasts by up to 40%, with review topics adding 10% and other review characteristics, 4%. Counterintuitively, topic extraction from reviews does not provide substantial benefits, despite the heavy computational resources it requires. Finally, the two-stage classification process, which leverages time-series data and rating trends, offers a more stable and robust performance than conventional single-stage methods. We provide considerations for system architecture development through robustness checks ensuring its resilience to stressors. Our experiments indicate that rating trends can change in subtle ways over time, leading a promising “star” product to turn into a liability (“dog”). E-commerce platforms can use the proposed system architecture proactively to identify and remove potentially dubious products instead of waiting to take reactive action.

To encourage small business inclusion in public procurement, the U.S. federal government has established the set-aside program that mandates government agencies to award a portion of their contracts to small businesses. We study whether and to what extent the performance of R&D contracts awarded through this program differs from those awarded through open competition. Analyzing a large dataset of federal R&D contracts, we find that despite restricting competition to small businesses, set-aside R&D contracts experience lower schedule and cost overrun than R&D contracts awarded through open competition. Furthermore, although set-aside R&D contracts experience lower schedule and cost overrun when they are awarded to more experienced contractor firms, this benefit arises primarily from a contractor firm's experience in executing R&D contracts across different agencies compared to the firm's experience with the same agency. Finally, set-aside R&D contracts awarded early in a fiscal year experience lower schedule and cost overrun than those awarded later. Post-hoc analysis examining the underlying dimensions of different-agency experience highlights the asymmetric effects of related-agency experience and unrelated-agency experience of contractor firms on the performance of set-aside R&D contracts awarded by the Department of Defense. While related-agency experience improves contract performance, unrelated-agency experience has a detrimental effect on contract performance. These findings demonstrate that small business inclusion policies may not necessarily compromise contract performance. Importantly, they emphasize the need for federal agencies and contracting officers to consider the underlying dimensions of contractor firm experience and contract award timing to improve contract performance and taxpayer money utilization.

The flexibility of choosing the ad action as a function of the consumer state is critical for modern-day marketing campaigns. We study the problem of identifying the optimal sequential personalized interventions that maximize the adoption probability for a new product. We model consumer behavior by a conversion funnel that captures the state of each consumer (e.g., interaction history with the firm) and allows the consumer behavior to vary as a function of both her state and firm’s sequential interventions. We show our model captures consumer behavior with very high accuracy (out-of-sample area under the curve of over 0.95) in a real-world email marketing dataset. However, it results in a very large-scale learning problem, where the firm must learn the state-specific effects of various interventions from consumer interactions. We propose a novel attribution-based decision-making algorithm for this problem that we call model-free approximate Bayesian learning. Our algorithm inherits the interpretability and scalability of Thompson sampling for bandits and maintains an approximate belief over the value of each state-specific intervention. The belief is updated as the algorithm interacts with the consumers. Despite being an approximation to the Bayes update, we prove the asymptotic optimality of our algorithm and analyze its convergence rate. We show that our algorithm significantly outperforms traditional approaches on extensive simulations calibrated to a real-world email marketing dataset.