Informs Annual Meeting 2017

MC20

INFORMS Houston – 2017

MC20

MC22

342B Managing New Product Development Uncertainties Sponsored: Technology, Innovation Management & Entrepreneurship Sponsored Session 1 - To Better Manage Risks in Multi-Period New Product Development — Select Riskier Projects Janne Kettunen, Assistant Professor, George Washington University, Department of Decision Sciences, 2201 G. Street NW, Washington, DC, 20052, United States, jkettune@gwu.edu, Shivraj Kanungo We consider multi-period new product development (NPD) where the decision maker chooses, periodically, which development projects to fund from the pool of available projects. We show that the availability of new development projects in future periods, which the initiated projects can be switched for, has major implications for what projects are optimal to select and thereby for the portfolio value and risk. Furthermore, we show that a risk-averse NPD portfolio selection approach implies higher risk than a risk-neutral selection approach. We name this phenomenon the risk aversion paradox and develop an approach to overcome it. 2 - A Theoretical Analysis of the Lean Start-up’s Agile Product Development Process Onesun Yoo, University College London, London, United Kingdom, o.yoo@ucl.ac.uk, Tingliang Huang, Kenan Arifoglu Lean startup paradigm is emerging as a best practice for early product development for entrepreneurs and is widely adopted by entrepreneurship curricula in business schools. Despite its influence, there is no theoretical underpinning, leading to lack of clarity in its generalizability and explanation of implementation challenges in practice. We present a stylized model of the lean startup implementation problem. We show how the (vertical) quality of minimum viable product impacts learning about consumer’s (horizontal) taste, and also how product-market characteristics impact its effectiveness. 3 - Learning Strategies to Deal with Market Disruptions and Turbulences in a Finite Time Horizon Leonardo Santiago, Copenhagen Business School, Copenhagen Business School, Solbjerg Plads 3, Blok B.5. sal, Frederiksberg, 2000, Denmark, leonardo.p.santiago@gmail.com, Julia Couto, Nitin Joglekar The ability to continuously innovate is a key asset to maintain a competitive advantage. However, very often, market turbulences or disruptions force firms to shift gears in their new product development strategies. This work considers the question of how firms should react to such disruptions when pursuing a finite time initiative. In particular, we focus on the role of pivoting strategies and cognitive maps to manage efforts to explore and exploit. 4 - Firm Clockspeeds: Toward a Theory of Relativity Sina Moghadas Khorasani, Doctoral Student, David Eccles School Fast clockspeeds imply short product lifespans. However, we find that two firms running at the same internal clockspeed may, to an outside observer, have quite different external clockspeeds, depending on product “type,” market competitiveness, and whether product innovation or process innovation prevails. 5 - When is Necessity the Mother of Invention? Sezer Ulku, McDonough School of Business, Georgetown University, DC, United States, Sezer.Ulku@georgetown.edu We conduct several experiments to investigate how time, resource and performance constraints influence the search behavior and the performance achieved in problem solving tasks. We find that solution quality attained under a moderate constraint is superior to that attained when no such constraint is present. This is explained by differences in search effectiveness. We also find that in the presence of a challenging performance goal, participants exert more effort on average, they are more attentive to the tasks and as a result they achieve better results. As expected, when the resource constraint becomes very tight, performance suffers. of Business, Salt Lake City, UT, 84112, United States, sina.moghadas@eccles.utah.edu, Glen M.Schmidt

342D Data-driven Studies in Revenue Management Sponsored: Revenue Management & Pricing Sponsored Session Chair: Ken Moon, The Wharton School, Philadelphia, PA, 19104, United States, kenmoon@wharton.upenn.edu Co-Chair: Kostas Bimpikis, Stanford University, 655 Knight Way, Stanford, CA, 94305, United States, kostasb@stanford.edu 1 - Leveraging Comparables for New Product Sales Forecasting Georgia Perakis, Massachusetts Institute of Technology, Sloan We introduce a simple and intuitive approach for forecasting demand for new products. We identify comparable products that allow us to make accurate predictions. The approach we propose is both simple and intuitive and mimics the current operational practice of retailers. We show that our approach is fast and scalable. In collaboration with two big industry partners, we provide prediction improvements of the order of 40%-50% (WMAPE) that are robust across various product categories. 2 - Predicting with Proxies: Pitfalls and Solutions Hamsa Sridhar Bastani, IBM Thomas J. Watson Research Center, Yorktown Heights, NY, United States, hamsab@wharton.upenn.edu An increasingly common problem with machine learning is that true outcomes of interest are costly to measure. Instead, practitioners use proxy outcomes for prediction and subsequent decision-making. For example, a content recommender may make critical decisions based on predicted click-through rates (easily available) rather than user experience (costly to measure). However, the proxy outcome may be biased and result in poor decisions. We introduce a novel estimator that leverages many proxy outcomes and a few true outcomes to provably detect and reduce bias. Simulations on real datasets demonstrate significant improvements in prediction accuracy. 3 - Welfare Implications of Congestion Pricing: Evidence from SF Park Hsin-Tien Tsai, UC Berkeley, 1822 Francisco St., Apt 4, Berkeley, CA, 94703, United States, hsintien@berkeley.edu, Pnina Feldman, Jun Li Using data from San Francisco both before and after the implementation of a congestion pricing policy, we estimate the welfare implications of the policy. We find that congestion pricing increases consumer welfare in popular areas, but when implemented in less-congested areas, it may actually hurt consumers. In less congested areas, a simpler pricing policy may actually achieve higher welfare than a complex one. 4 - Optimal Retail Location: Empirical Methodology and Application to Practice Chloe Kim Glaeser, University of Pennsylvania, 3730 Walnut St. Suite 500, Philadelphia, PA, 19104, United States, chloekim@wharton.upenn.edu, Marshall L.Fisher, Xuanming Su We empirically study the spatio-temporal location problem motivated by an online grocery retailer. Customer demand is influenced by the convenience of pick-up locations and days. We combine demographic and economic data, business location data, and the retailer’s historical sales and operations data to predict demand at potential locations. We introduce a novel procedure that combines machine learning and econometric techniques. Based on the predicted demand, we solve the spatio-temporal integer program using quadratic program relaxation to find the optimal pick-up location configuration and schedule. School of Management, MIT, 100 Main Street Rm E62-565, Cambridge, MA, 02142-1347, United States, georgiap@mit.edu, Lennart Baardman, Divya Singhvi

210