Fan Zhang, Marketing, PhD
Research Interests: Education:
My doctoral training is in the field of quantitative marketing. I use econometric techniques and methods to address important managerial issues arisen from today’s new technology evolution. My research approach is to identify key marketing problems and then construct econometric models to gain insights about firms and customers’ behavior using big data. My thesis studies the influence of technology-driven changes in delivery channels on customers and firms. In particular, I examine how the introduction of new channels delivering the same products or services affect customers’ purchasing behavior, firms’ strategies, and short-term and long-term profitability.
Job Market Paper Title: “Customer Migration from Online Retail Platforms”
Abstract: My job market paper examines a channel structure where a company sells the same products through both an online retail platform and its own online store. We focus on the long-term benefit of the partnership to the company in terms of helping migrate customers from the online retail platform to the online store. For this purpose, we use customer purchase data to estimate a simultaneous model of customers’ channel choice and expenditure decisions. We model customers’ consideration sets as being determined by their awareness of the company’s own store. We find that the company has greatly benefited from the partnership with the retail platform. It gains many new customers from the partner retail platform and many of these newly obtained customers migrate to the company’s own online store in later periods. We use counterfactuals to show how the company can manipulate product assortments to improve its long-term profit and create a “win-win” situation for both the retail platform and itself. We also demonstrate how the value of the partnership with retail platforms depends on the customer awareness and the intrinsic attractiveness of the company’s store.
Other Completed Papers
Title: “Technology Innovation and Consumer Choice in the Cable TV Industry: from Pay-per-View to Video-on-Demand” with Tat Y. Chan and P.B. Seethu Seetharaman
Abstract: This paper empirically investigates how technology innovation in the cable TV industry affects households’ movie consumption and firms’ profit. During the study period, the cable company starts to offer video-on-demand (VOD) programming in addition to older pay-per-view (PPV) technology to its subscribers in selected geographical regions. The new movie delivery technology allows households to select programs from a long list of movies at their preferred time. We study how such technology improvement influences households’ viewing preferences for regular movies and for adult movies. The latter genre is the main source of profit for the company. We model households’ decisions of movie consumption and service subscription, assuming that they are forward-looking in decision-making and heterogeneous in preferences for the two types of movies. Based on the model estimation results, we show that VOD helps the cable company to attract new subscribers and increase household movie consumption. The primary source of consumption increase, however, comes from regular movies that are lower priced and have lower profit margin for the cable company. New subscribers are also attracted by the improved experience of watching regular movies. For both existing and new subscribers, the benefit for watching adult movies from VOD is not different from PPV. Combining the estimation results and the data on movie prices and margins, we compute the new equilibrium division of revenue between the cable company and upstream movie studios after VOD has been widely adopted. We also compute equilibrium prices for the two types of movies.
Title: “The Proportional Hazard Model for Regular versus Irregular Purchase Times: A Comparison of Alternative Specifications” with P.B. Seethu Seetharaman
Abstract: Customers differ from each other in terms of purchase times in frequently purchased packaged goods categories. For example, some customers prefer to shop regularly (Type 1) while others have highly irregular purchase times (Type 2). In this paper, we employ the proportional hazard model (PHM), a commonly used model of purchase-timing behavior, and compare five different parametric specifications of the baseline hazard in terms of their relative abilities to explain purchase timing outcomes of Type 1 versus Type 2 customers. We study whether a given parametric baseline hazard specification leads to systematic bias in the estimated PHM for a given customer type. We find that the expo-power specification is most appropriate for Type 1 customers, while Weibull is most appropriate for Type 2 customers. We document the nature of parameter bias in employing one baseline hazard when the other is more appropriate. This study will be a useful reference to researchers hoping to use PHM for packaged goods data.
- M.A. in Economics, Stony Brook University, 2009
- B.A. in Finance, Fudan University, 2007
Curriculum Vitae | Website | Email