Informs Annual Meeting 2017

MA26

INFORMS Houston – 2017

MA24

have a purchase rights structure: agent is paid a retainer plus a bonus each time the principal makes an acquisition based on the agent’s lead. The size of this bonus is decreasing in time and depends on probability of finding a suitable opportunity but not on its value. The outsourcing decision is complex and depends on several factors, with information and cost asymmetries being most important. Delaying outsourcing for later parts of search is shown to be optimal. 2 - Breaking the Stalemate: Signaling, Experience, and Value Creation in Technology Licensing Pascale Crama, Singapore Management University, 50 Stamford Road, Singapore, 178899, Singapore, pcrama@smu.edu.sg, Reddi Kotha, Phillip Kim We address the question of how signals can be employed to broker technology licensing agreements when information asymmetry prevents risk-averse parties from working together. In our setting, inventors and licensee firms are brought together by licensing managers (LM) to commercialize new technologies. We argue that experienced LM intentionally alter contractual payment structures to signal value creation to the licensee based on ability and need. With experience, LM gain the ability to create value based on selecting promising inventions and matching them to capable licensing partners. In addition, experienced negotiators weaken and amplify signals as needs to address uncertainties vary. 3 - Truth or Funds? Human Behavior under Competition for Resources in R&D Özge Tüncel, Singapore University of Technology and Design, 8 Somapah Rd, Singapore, tuncel_ozge@mymail.sutd.edu.sg, Niyazi Taneri, Nektarios Oraiopoulos, Jochen Schlapp Self-interested project managers working on R&D projects compete for their firm’s limited resources. This may lead to misreporting of project prospects (adverse selection) and distort both the efforts made by project managers (moral hazard) and the firm’s resource allocation decision in the absence of carefully designed incentive mechanisms. We study how human behaviour further complicates this process and suggest remedies to alleviate the problem. 4 - Behavioral Responses in the Multi-armed Bandit Model Fabian J. Sting, University of Cologne, Albertus-Magnus-Platz, c/o WiSo-Sekretariat Universitaetsstrasse 91, Cologne, 50923, Germany, sting@wiso.uni-koeln.de, Mirko Kremer, Andolhossein (Navid) Ghayazi How do human decision makers respond to changing business landscapes, e.g., in project portfolio management? Toward answering that question we investigate the well-studied multi-armed bandit model—-yet enriched with disruptions, i.e., changes in the stochastic processes that yield the arms’ payoffs. We compare normative results from comprehensive simulation studies with behavioral results from lab experiments. 350B DEA Methodological Extensions and Applications Invited: Data Envelopment Analysis Invited Session Chair: Mehdi Toloo, Technical University of Ostrava, Czech Republic, VŠB-TU Ostrava, Faculty of Economics, Sokolská tr. 33., Ostrava, 701 21, Czech Republic, m_toloo@yahoo.com 1 - Extended DEA Efficiency Model of Rice Growing Farmers with Undesirable Output and Interval Data Azizul Baten, Shahjalal University of Science and Technology, Sylhet, Bangladesh, baten_math@yahoo.com An extended Data Envelopment Analysis model based on Directional Distance Function approach considering undesirable outputs, climatic factors and interval data is developed. The inclusion of these factors enables to represent the actual production environment. The interval data approach is used to model the uncertainties in the production environment and obtain the interval efficiency score of rice farmers who produces undesirable outputs and operates under uncertainty. The optimistic and pessimistic cases are considered to estimate the upper and lower bound of the efficiency score for each farmers as this model cannot be solved in its original form. 2 - Measuring Nursing Home Quality Performance using DEA Christine Pitocco, Stony Brook University, 6 Essex Drive, Ronkonkoma, NY, 11779, United States, christine.pitocco@stonybrook.edu A primary concern of consumers as they select a nursing home is the quality of care provided by the home. No single quality measure exists to evaluate nursing home performance; therefore a multidimensional approach is required. We illustrate the use of Data Envelopment Analysis using nine quality indicators to evaluate 589 nursing homes in New York State in 2015. MA26

342F Revenue Management Sponsored: Revenue Management & Pricing Sponsored Session Chair: Srikanth Jagabathula, NYU Stern School of Business, New York, NY, 10012, United States, sjagabat@stern.nyu.edu 1 - Can Managers Plan Assortments? An Experimental Study Yulia Vorotyntseva, University of Texas at Dallas, Richardson, TX, United States, Yulia.Vorotyntseva@utdallas.edu, Elena Katok, Dorothee Honhon Assortment planning can be a very complex problem due to product interaction effects, such as cannibalization and inventory pooling benefits. We want to find out if human decision makers have an intuitive understanding of such effects, and whether they can gain it with practice. To do this, we develop and conduct a behavioral experiment, where the subjects are repeatedly picking assortments in a simplified environment. We find that the subjects perform better when the profit maximizing assortment consists of fewer products and that subjects improve their decisions over time. 2 - Group Conversion: A Better KPI to Compare Stores Anup Hanamant Walvekar, INSEAD, 1 Ayer Rajah Avenue, PhD Office, Singapore, 138676, Singapore, anup.walvekar@insead.edu, Nils Rudi, Serguei Netessine Conversion rate is an important KPI for a retail store that captures the fraction of sales opportunity that the store was able to capture. As a group of people shopping together seldom have multiple transactions, the number of groups visiting the store is a more appropriate measure of sales opportunities than the number of individuals. Identifying groups in traffic using video footage is a laborious and expensive task. In this paper, we mathematically model the customer arrival process and derive methods to identify of groups in traffic data obtained from simple counters. 3 - A Model-based Projection Technique for Segmenting Customers Ashwin Venkataraman, New York University, New York, NY, United States, ashwin@cs.nyu.edu, Srikanth Jagabathula, Lakshminarayanan Subramanian We consider the problem of segmenting a population of customers into non- overlapping groups with similar preferences, using diverse signals such as purchases, ratings, clicks, etc., over a large universe of items. We propose a model-based projection technique, which transforms the diverse observations into a comparable scale and deals with missing data by projecting the transformed data onto a low-dimensional space. We then cluster the projected data to obtain the customer segments. Our method achieves upto 84% improvement in the accuracy of new movie recommendations on the MovieLens data set and 6% improvement for similar item recommendations algorithm on an offline dataset at eBay. 4 - Building Optimized and Hyperlocal Product Assortments: A Nonparametric Choice Approach Srikanth Jagabathula, NYU. Stern School of Business, 44 W. 4th St, Kmc Rm 8-74, New York, NY, 10012, United States, sjagabat@stern.nyu.edu, Vivek Farias, Devavrat Shah Fashion retailers want to “hyper-localize” product assortments, i.e., customize offerings to the preferences of customers visiting each store, to increase sales. The challenge is demand prediction because of the large variety. To address this challenge, we propose: (a) a nonparametric choice modeling technique that uses purchase transactions tagged by customer IDs and (b) an optimization framework that uses our demand predictions to optimally allocate merchandise to different stores. In a controlled experiment at a large US fashion retailer (~$3B annual revenue and ~300 stores), our methods resulted in additional 7% revenue growth (~$200M profit impact) over the current method. 350A Topics in Entrepreneurship and Innovation Invited: New Product Development Invited Session Chair: Pascale Crama, Singapore Management University, Singapore, 178899, Singapore, pcrama@smu.edu.sg 1 - The When and How of Outsourced Technology Scouting Sasa Zorc, INSEAD, 1 Ayer Rajah Avenue, PhD Office, Singapore, 138676, Singapore, sasa.zorc@insead.edu, Sameer Hasija, Ilia Tsetlin A pharmaceutical firm seeks to conduct a search for acquisition opportunities. We consider the decision of whether to outsource this process or do it in-house, and model it in dynamic principal-agent framework. The optimal contract is shown to MA25

141