Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

SD19

3 - Strategic Inventory under Supply Chain Competition Yanzhi David Li, City University of Hong Kong, Hong Kong, Li Xi, Ying-Ju Chen We examine the effect of strategic inventory in the presence of chain-to-chain competition. We show that as the competition between two supply chains becomes fiercer, retailers will carry more inventory, which intensifies the supply chain competition. Consequently, this competition intensification effect can overshadow the effect of double marginalization alleviation. 4 - Strategic Inventory in Non-exclusive Reselling Environments Abhishek Roy, Temple University, Fox School of Business, Philadelphia, PA, United States, Stephen M. Gilbert, Guoming Lai Although the effects of strategic inventory in dynamic contracts on supply chain agents in bilateral monopolies are well known, we find that those effects are altered when competing manufacturers sell partially substitutable products through a common retailer. We examine the manufacturers’ choices between a dynamic contract and a commitment contract that includes both price and quantity commitments. In contrast to what occurs in a bilateral monopoly, we find that manufacturers may prefer to use commitment contracts and that such contracts may arise in equilibrium. n SD18 North Bldg 128A Social Learning in Operations General Session Chair: Gad Allon, The Wharton School of University of Pennsylvania, Philadelphia, Pennsylvania 1 - Pricing and Prioritizing Differentiated Services when Customers Learn Socially Koushiki Sarkar, Kellogg School of Management, Northwestern University, Evanston, Saint Barth lemy, Gad Allon, Achal Bassamboo Dissemination of information via social networks play a critical role in modern decision-making. We model the effects of social information in service differentiation under a queuing setting. We consider a profit-maximizing firm serving two groups of customers in an M/M/1 queue with two priority classes, where the customers observe service reviews at random from each period and use that information to decide their action for the coming period. We study the steady-state behavior of this system and obtain the firm’s profit function and the optimal prices to be set. 2 - Learning, Welfare, and Profits in Two-sided Service Platforms Kostas Bimpikis, Stanford University, Stanford, CA, United States, Yiangos Papanastasiou, Wenchang Zhang Platforms such as Airbnb and Upwork have reduced search and information frictions in the service industry. The efficient gathering of information about the quality of providers in such a setting is of first-order importance. Our goal is to provide design guidelines taking into account that the quality of new providers is unknown and information about them can only be generated via transactions. The main design levers we explore are commission rates and information disclosure policy. We show that the optimal design features disclosing quality information with a delay and can result in significantly higher revenues than in the case that the platform does not optimize over information provision. 3 - Which Locations are the Most Valuable in a Network of Service Outlets? Using a comprehensive dataset from a major restaurant network, we evaluate the impact of a restaurant’s quality as measured by customer satisfaction and the impact of supply chain quality measured by quality complaints between supply chain members on the total sales at all locations within the network. As a result we identify the most valuable locations in terms of their impact on the total sales in the network, providing guidance on which service outlets should be prioritized for investment and development. 4 - Value of Traceability in Supply Chains Yao Cui, Cornell University, Ithaca, NY, United States, Ming Hu, JingChen Liu We consider supply chains where the buyer cannot identify which supplier is at fault when quality defect occurs (e.g., agricultural supply chains). We study how new technologies that enable traceability (e.g., blockchain) can create value under different supply chain structures. Masha Shunko, University of Washington, Foster School of Business, PACCAR Hall, Seattle, WA, 98195, United States, Qiuping Yu, Shawn Mankad

n SD19 North Bldg 128B Innovations in Pricing Sponsored: Revenue Management & Pricing Sponsored Session Chair: Adam Elmachtoub, Columbia University, New York, NY, 10027, United States Co-Chair: Michael Hamilton, Columbia University, New York, NY 1 - Network Revenue Management under Strategic Behavior and Customer Choice Nikolaos Trichakis, MIT, 100 Main Street, E62-576, Cambridge, MA, 02143, United States, Yiwei Chen We consider a canonical network revenue management model in which a seller offers multiple products, which consume capacitated resources, for sale to utility- maximizing customers who choose both (1) when to buy, and (2) which product to buy. We use a novel decomposition approach for a suitable dynamic mechanism design problem to derive for all non-anticipating dynamic pricing policies an upper bound to expected revenues. We employ the bound to obtain a constant-factor guarantee for the performance of static pricing in a fluid-type regime. 2 - The Impact of IPO on Peer-to-Peer Lending Platforms Kevin Jiao, NYU Stern, New York, NY, 10012, United States, Maxime Cohen In this talk, we investigate the event of initial public offering and its impact on the peer-to-peer lending U.S. market. Focusing on the two leading U.S. platforms, we exploit a research setting which compares the difference between the private platform (Prosper) and the public platform (Lending Club) from both operational and investment perspectives. Using econometrics tools, we carefully examine if the event of IPO causes the firm to manipulate some of its operational decisions. 3 - Price Optimization for Rotable Spare Parts Yunjie Sun, Columbia University, 500 West 120th Street, 535 S. W. Mudd Building, New York, NY, 10027, United States, Omar Besbes, Adam Elmachtoub Working in collaboration with Dassault Falcon Jet (DFJ), we address the problem of pricing for rotable parts, one special type of spare part that can be repaired and reused. We develop a model to maximize the expected profit rate through selling rotable parts. This model captures the special dynamics of rotable parts yet still easy to implement. We also discuss the on-going large scale implementation of the pricing model at DFJ. We leverage data in estimating the inputs for the optimization model. We build a user interface which takes the estimated inputs and returns a suggested price that is robust to uncertainties of inputs. We provide initial results of implementation for over 500 parts since May 2018. 4 - The Power of Opaque Products in Pricing Michael Hamilton, Columbia University, New York, NY, 10027, United States, Adam Elmachtoub We study the power of selling opaque products, i.e., products where a feature is hidden from the customer until after purchase. We consider models where traditional items are sold at a single price alongside opaque products corresponding to subsets of items and bench mark our opaque selling strategies against two common selling strategies: one which charges different prices for the items (discriminatory pricing), and one which charges a single price (single pricing). When customers are unit-demand and draw valuations exchangeably, we give a sharp characterization for when opaque selling outperforms discriminatory pricing, and characterize the maximal revenue lift opaque products can provide.

97