In today's marketplace, consumers purchase a basket of goods of different categories from many different types of retail stores. While most marketing literature focuses on studying a single retail chain or a single category, my dissertation considers consumers' collective basket purchases across different retail formats. By doing so, my dissertation not only provides new insights into retail competition, but also opens new paths of leveraging consumers' collective behavior from multiple categories and multiple retail formats to facilitate practitioners' decision making. In the first chapter, "Competition Among Retail Formats'", I study how households allocate their budget for food and related items across different types of retailers. I construct a structural demand model that coherently characterizes both consumer expenditure and expenditure allocation among six retail formats. The model considers how price and assortments influence expenditure allocations. To illustrate how my demand model can be applied in practice, I consider a situation faced by discount stores like Target, during the Great Recession. In 2008, discount stores experienced a loss in share to other retail formats like warehouse clubs. I use my model to examine two different policies that discount stores could use to compete with other retail formats. First, I consider whether discount stores should compete by lowering prices. Second, I evaluate the effectiveness of introducing a new retail format that has smaller assortments. I find that introducing a small-size format is relatively more effective than competing with lower prices. In the second chapter, "Evaluation of the Tobacco Removal by CVS" (with Brett Gordon), we ask when a channel ceases distribution of a product category, whether consumers shift to other channels or reduce their purchases entirely. We study this question in the context of the tobacco removal by CVS. Recently, concerns have arisen that selling tobacco in pharmacies contradicts their mission of providing health care. Especially in this category, understanding the cross-channel substitution effects is necessary to inform the public health debate. To measure the impacts of the tobacco removal, we apply difference-in-difference and matching approaches to individual-level purchase data. We find that the tobacco removal did not reduce cigarette purchases, as smokers substituted to other drug stores and other channels. Since cigarettes are a key category that drives store visits, we also find that the tobacco removal led to a large reduction in shopping trips and spending at CVS from the affected smokers. Our findings cast doubt to the policies that aim reducing cigarette use by limiting tobacco access at a single pharmacy retailer. In the third chapter, "Harbingers of Failure ... and Success" (with Eric Anderson and Blakeley McShane), we show that there is tremendous value in looking beyond a single category and leveraging consumers' purchase behavior from many different categories for new product forecasting. In particular, we develop a cohesive, flexible semi-parametric model to predict new product lifetime, which not only improves predictive accuracy but also generates interpretable individual-level estimates that are modeled as a function of past behavior. we find evidence for harbingers of failure, customers who systematically purchase new products that go on to fail, as well as harbingers of success, customers who systematically purchase new products that survive for a long time. We explore mechanisms by testing behavioral hypotheses via a survey. The findings are consistent with the explanation that harbingers of failure are variety-seeking and make fewer repeat purchases compare to other customers.

Last modified

• 04/09/2018

Creator

• Chaoqun Chen

DOI

• https://doi.org/10.21985/N2Q69H

Subject

• Marketing

Keyword

• Marketing

Date created

• 2017-01-01

Resource type

• Dissertation

Rights statement

• In Copyright