Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MC21

choices. By linking the design characteristics to specific mechanisms, we provide insight into how program managers can influence consumers’ decision. 2 - How do Price Promotions Affect Customer Behavior on Retailing Platforms? Evidence from a Large Randomized Experiment on Alibaba Dennis Zhang, Washington University in St. Louis, University City, MO, 63124, United States, Hengchen Dai, Lingxiu Dong We study how a promotion strategy—offering customers a discount for products in their shopping cart—affects customer behavior in the short and long term on a retailing platform. We conducted a randomized field experiment involving more than 100 million customers and 11,000 retailers with Alibaba Group, the world’s largest retailing platform. We randomly assigned eligible customers to either receive promotions for products in their shopping cart or not. In the short term, our promotion program doubled the sales of promoted products. In the long term, we causally document unintended consequences of this promotion program during the month following our treatment period. 3 - Underrepresented Minorities and LGBT in the Sharing Economy: Bias and Financial Incentives in Ridesharing Platforms Christopher Dalton Parker, Pennsylvania State University, 411 Business Building, University Park, PA, 16802, United States, Jorge Mejia Operational transparency can be good for business. However, it may also enable biased behavior if those with information about customers can choose not to provide a service for the customer. We explore this through a field experiment on a major ridesharing platform which recently changed the timing of information provided to drivers in order to reduce bias. We find significant bias still exists against URM and LGBT individuals. However, dynamic pricing moderates the effects. Policy implications will be discussed. 4 - Clearing Matching Markets Efficiently: Informative Signals and Match Recommendations Yash Kanoria, Columbia Business School, 404 Uris Hall, New York, NY, 10027, United States, Itai Ashlagi, Peng Shi, Mark Braverman We study how to reduce congestion in two-sided matching markets with private preferences. We measure congestion by the number of bits of information that agents must (i) learn about their own preferences, and (ii) communicate with others, before obtaining their final match. Previous results by Segal (2007) and Gonczarowski et al. (2015) suggest that a high level of congestion is inevitable under arbitrary preferences before the market can clear with a stable matching. We show that when the unobservable component of agent preferences satisfies certain natural assumptions, it is possible to recommend potential matches and encourage informative signals such that the market reaches a stable matching with a low level of congestion. The main idea is to only recommend partners with whom the agent has a non-negligible chance of both liking and being liked by. Chair: Francesca Parise, MIT, Cambridge, MA, 02141, United States Co-Chair: Asuman Ozdaglar, Massachusetts Institute of Technology, Cambridge, MA, 02139, United States 1 - Path to Stochastic Stability: Comparative Analysis of Stochastic Learning Dynamics in Games Hassan Jaleel, KAUST, Jeddah, Saudi Arabia, Jeff S. Shamma Stochastic stability (SS) analysis can accurately explain the long-term behavior of stochastic learning dynamics. However, this solution concept does not explain the transient behavior of these dynamics. Consequently, we cannot distinguish between different learning rules with the same steady state using SS analysis. We develop a framework for the comparative analysis of stochastic learning dynamics with different update rules that lead to a same steady-state behavior. We propose multiple criteria to quantify the differences in the short and medium-run behaviors of these dynamics. We apply these criteria to compare Log-Linear Learning and Metropolis Learning and gain valuable insights. 2 - A Variational Inequality Framework for Network Games: Existence, Uniqueness, Convergence and Sensitivity Analysis Francesca Parise, Massachusetts Institute of Technology, Cambridge, MA, United States, Asuman Ozdaglar We provide a unified variational inequality framework for the study of fundamental properties of the Nash equilibrium in network games. We identify several conditions on the underlying network (in terms of spectral norm, infinity norm and minimum eigenvalue of its adjacency matrix) that guarantee existence, uniqueness, convergence and continuity of equilibrium in general network games with multidimensional and possibly constrained strategy sets. We delineate the relations between these conditions and characterize classes of networks that satisfy each of these conditions. n MC21 North Bldg 129B Game Theory and Networks Sponsored: Revenue Management & Pricing Sponsored Session

n MC19 North Bldg 128B Revenue Management in Uncertain Environments Sponsored: Revenue Management & Pricing Sponsored Session Chair: Omar Besbes, Columbia University, New York, NY, 10027, United States 1 - Personalized Dynamic Pricing with Machine Learning N. Bora Keskin, Duke University, Fuqua School of Business, 100 Fuqua Drive, Durham, NC, 27708-0120, United States, Gah-Yi Ban Motivated by online retail applications, we consider a seller who offers personalized prices to individual customers. The seller initially does not know the impact of individual customer characteristics on demand, but can learn about this relationship via sales observations. We construct and analyze near-optimal policies that balance the learn-and-earn tradeoff in this setting. 2 - Stein Shrinkage for Stochastic Optimization Vishal Gupta, USC Marshall School of Business, 3670 Trousdale Parkway, Los Angeles, CA, 90026, United States, Nathan Kallus Inspired by Stein’s phenomenon in statistics, we propose a new shrinkage algorithm for solving many data-driven stochastic optimization problems simultaneously. Our procedure pools data across problems. Perhaps surprisingly, as the number of problems increases, our method outperforms methods that decouple the problems, even when the problems are unrelated and data are drawn independently. Unlike the Stein phenomenon in statistics, our method does not require strong distributional assumptions and applies to general constrained optimization problems. 3 - Less Can Be More in Price Experimentation; The Uncertain Demand Case Divya Singhvi, MIT, 516 University Avenue, Ithaca, NY, 14850, United States, Georgia Perakis We consider a dynamic pricing problem where the retailer has no knowledge of the demand curve and there is a cost on price experimentation. The retailer seeks to efficiently learn the demand curve and keep the cost of price experimentation low. We propose an optimistic-pessimistic approach for price experimenting and learning which is simple and mimics industry practice. We provide bounds on the number of price experimentations needed to achieve a threshold revenue level. We show that with few price experimentations (aka 4) we can be within 18% of the optimal unknown price. 4 - Prior-independent Optimal Auctions Amine Allouah, Columbia University, New York, NY, 10027, United States, Omar Besbes In this work, we study the design of optimal prior-independent selling mechanisms. In particular, the seller faces buyers whose values are drawn from an unknown distribution, and only knows that the distribution belongs to a particular class of distributions. We analyze a maximin setting in which the seller attempts to optimize the worst-case fraction of revenues compared to those of an oracle with knowledge of the distribution. We first characterize the structure of optimal mechanisms. Leveraging such structure, we then establish tight lower and upper bounds on performance, leading to a crisp characterization of optimal performance for a spectrum of families of distributions.

n MC20 North Bldg 129A Joint Session RMP/Practice Curated: Marketplace Analytics Sponsored: Revenue Management & Pricing Sponsored Session Chair: Wedad Elmaghraby Co-Chair: Chris Parker, Pennsylvania State University, University Park, PA

1 - Should I Pay for this Purchase or Redeem Points? Effects of Loyalty Program Design on Consumer Decisions to Redeem Points So Yeon Chun, Georgetown University, Arlington, VA, 22202- 7416, United States, Rebecca Hamilton In many loyalty programs, customers earn points for their purchases, which they can later exchange for additional products and services. In a sense, points function as a currency that consumers can spend instead of money. However, we uncover systematic differences in the way consumers spend points compared with money. We first propose a conceptual model of the consumer’s choice to redeem points or use money for a specific purchase, and then we conduct a series of studies to investigate the impact of program design characteristics on consumers’

191