2016 INFORMS Annual Meeting Program

SB49

INFORMS Nashville – 2016

SB48

4 - Tractable Equilibria For Sponsored Search With Budget Optimizing Bidders Dragos Florin Ciocan, INSEAD, florin.ciocan@insead.edu, Krishnamurthy Iyer We examine a model of sponsored search markets where bidders strategically choose their budgets and bids, while the ad network can throttle bidders to optimize its own revenues. We show the equilibria in this market take a simple form and that for these equilibria the network’s optimal throttling policy is greedy. SB47 209C-MCC New Topics in Revenue Management and Pricing Sponsored: Revenue Management & Pricing Sponsored Session Chair: So Yeon Chun, McDonough School of Business, Georgetown University, Washington, DC, United States, sc1286@georgetown.edu 1 - Worker Poaching In A Supply Chain: Enemy From Within? Gad Allon, Northwestern University, g-allon@kellogg.northwestern.edu, Achal Bassamboo, Evan Barlow Poaching workers has become a universal practice. We explore worker poaching between firms linked in a supply chain. We show that the classical intuition from labor economics is insufficient in explaining poaching between supply chain partners. We also show how and under what conditions worker poaching can actually improve supply chain performance. Finally, we show how the equilibrium identity of the supply chain bottleneck depends on the interaction between hiring, poaching, and productivity. 2 - The Revenue Impact Of Dynamic Pricing Policies In Major League Baseball Ticket Sales Joseph Xu, University of Pennsylvania, Philadelphia, PA, United States, jiaqixu@wharton.upenn.edu, Senthil Veeraraghavan, Peter Fader We study RM implementation of multiproduct dynamic pricing by a Major League Baseball franchise for their single game tickets. We develop a comprehensive customer choice model to calibrate and design a dynamic pricing policy for the franchise. Our model also incorporates external factors that drive customer valuation of sports tickets, such as the effect of the home team’s on-field performance and the effect of overall attendance level. Our counterfactuals show potential revenue improvement of up to 15% through the effective use of dynamic pricing. We also find that a properly calibrated fixed pricing policy can achieve similar levels of performance as the optimal dynamic pricing policy. 3 - Designing Rewards-based Crowdfunding Campaigns For Strategic Contributors Soudipta Chakraborty, Duke University, Durham, NC, 27708, United States, sc390@duke.edu, Robert Swinney We study a model of rewards crowdfunding with the all or nothing funding mechanism. The creator of a crowdfunding campaign sets a target funding level and the campaign is successful only if the funding it receives meets this target. A creator can have two possibly competing objectives: maximize the likelihood of success and maximize the expected total funding. The contributors incur a transaction cost while pledging to a campaign. As a result, they behave strategically and decide whether to pledge at the beginning or to wait till the target is met. We analyze how a creator, who encounters such strategic behavior, can achieve her objectives by optimally using the operational parameters of her campaign. 4 - Setting The Optimal Value Of Loyalty Points So Yeon Chun, McDonough School of Business, Georgetown University, sc1286@georgetown.edu, Dan Andrei Iancu, Nikolaos Trichakis A loyalty program introduces a new currency—the points—through which customers transact with a firm. We study the problem of optimally setting the monetary value of points, i.e., pricing in this new currency, in a multi-period setting. We first show that point pricing is different from cash pricing primarily due to the way points are accounted for, as liabilities on the firm’s balance sheet, and then we characterize the optimal cash and point pricing policies.

210-MCC Social Media Analytics for Businesses Invited: Social Media Analytics Invited Session

Chair: Panagiotis Adamopoulos, New York University, School of Business, New York, NY, 11111, United States, padamopo@stern.nyu.edu 1 - Monetizing Sharing Traffic Through Incentive Design: Evidence From A Randomized Field Experiment Tianshu Sun, University of Southern California, 3330 Van Munching Hall, Los Angeles, CA, 20742, United States, tianshu.sun@gmail.com, Siva Viswanathan, Elena Zheleva Using a large-scale randomized field experiment, we examine whether and how firms can engage customers involved in online social sharing, through the design of novel incentive mechanisms. We find evidence that incentive design has a significant impact on both sender’s purchase and referrals, but in a different ways. Specifically, compared to the senders who receive non-shareable promotional code, senders who receives shareable code are less likely to make purchases themselves, but much more likely to make further referrals. We further leverage variation in incentive design to untangle three motives underlying the sender’s sharing:self-regarding, other-regarding, group-regarding motive. 2 - Realizing The Activation Potential Of Online Communities Marios Kokkodis, Boston College, kokkodis@bc.edu In this work we present a data-driven stochastic framework that identifies which users and when are more likely to become heavy contributors in an online community. 3 - Word Of Mouth Vs. Word Of Health Inspectors: Evidence From Restaurant Reviews Chenhui Guo, University of Arizona, 1130 E Helen St, McClelland Hall 430, Tucson, AZ, 85721, United States, chguo@email.arizona.edu, Paulo B Goes, Mingfeng Lin Prior to purchase, consumers are naturally exposed to multiple sources of quality information. We study whether and how consumer word of mouth of restaurants—both volume and valence—is influenced by co-presence of information from health inspectors. We build a simple analytical model and conduct an empirical study using data from a leading consumer review site, showing that the availability of official information has a significant dampening effect on the volume of reviews generated by consumers. Moreover, the effect on valence is significantly positive, with a very small magnitude. 4 - The Role Of Dimensionality Reduction In Binary Classification For Social Network Data Jessica Clark, New York University, jclark@stern.nyu.edu, Foster Provost Dimensionality reduction is regarded as a key part of the predictive analytics process. We take a design-science approach to analyzing the role of dimensionality reduction via matrix factorization for binary classification using large, sparse social network data. The experiments in this work (which span a variety of data sets, modeling techniques, and DR methods) find that DR at best provides little advantage in terms of classification performance, and at worst can significantly negatively impact performance. The results emphasize the need for caution when utilizing DR in predictive modeling, which should serve as a guideline for applied data science researchers and industry practitioners. 211-MCC Case Competition II Sponsored: Education (INFORMED) Sponsored Session Chair: Palaniappa Krishnan, University of Delaware, Newark, DE, United States, baba@udel.edu 1 - Dynamic In-Game advertising: Managing Complex High-Stakes Operations Alan Scheller-Wolf, Carnegie Mellon University, Pittsburgh, PA, United States, awolf@andrew.cmu.edu, John Turner Dynamic in-game advertising is an advanced form of advertising in which ads are displayed on electronic billboards, stadium walls, or in other visually appealing spots within the 3D worlds of video games. This case teaches students not only about the economics of online advertising and how to solve complex multi- objective ad planning problems using goal programming, but also covers broader modeling concepts, practical modeling considerations, and discusses relevant strategic issues from the fast-growing and fast-changing online advertising industry. SB49

59