Analytics of Discrete Choice in Sequential Search and Price Promotion Settings
Author | : Natalia Kosilova |
Publisher | : |
Total Pages | : 0 |
Release | : 2022 |
ISBN-10 | : OCLC:1367872241 |
ISBN-13 | : |
Rating | : 4/5 ( Downloads) |
Download or read book Analytics of Discrete Choice in Sequential Search and Price Promotion Settings written by Natalia Kosilova and published by . This book was released on 2022 with total page 0 pages. Available in PDF, EPUB and Kindle. Book excerpt: In this dissertation we study structural models of discrete choice in sequential search and price promotions settings. We give the high-level overview of the dissertation in the first Chapter. In the second Chapter, we consider a problem of optimal upselling. Upselling is a sales technique used to motivate a customer to purchase a more expensive option than the one that the customer initially chooses. When carefully implemented, it can considerably improve revenues in the hospitality industry and possibly other revenue management contexts where it has not been used yet. We propose a simple and tractable framework based on a general Random Utility Model (RUM) that employs the information about the customer's initial choice to better understand their idiosyncratic reaction to an upsell offer. For a particular RUM -- Multinomial Logit Model -- we analyze the optimal upselling strategy of a firm providing hospitality services. We first analyze the case where the firm upsells a single product to a customer. We consider both the static (single-period) problem, and the dynamic problem, where the upsell price and choice of which product to upsell depend on inventory levels. In the static setting, we characterize the optimal upsell price for any product the firm may choose to offer as an upsell item. For one important special case we also provide a recommendation on which product should be offered. We formulate the dynamic problem and show some comparative statics of the optimal upsell price. We also demonstrate how easily the problem can be solved numerically. Finally, we generalize our framework to allow the firm to offer a portfolio of upsell offers simultaneously and develop the customer choice probabilities for this case. We estimate the generalized model on a real dataset provided by our industry partner and demonstrate that it adequately explains consumer behavior. Since our framework is based on MNL, the model parameters required for computing the terms of the optimal upsell offer can be estimated by well-known techniques that practitioners already use, and the application of our method appears straightforward and scalable. In the third Chapter, we marry an analytically tractable discrete choice model with a classic model of sequential search with perfect recall. Although significant progress has been made in the literature in analyzing customer choice behavior using random utility discrete choice models, discrete choice through a sequential search process has not received enough attention due to analytical intractability issues. We build on the seminal Pandora's Problem introduced by Weitzman (1979) as a model of sequential search, and the Exponomial Choice model (Alptekinoglu and Semple, 2016; Daganzo, 1979), which specifies random utility shocks to follow a (negative) exponential distribution. We derive closed-form choice probabilities and develop all the analytical tools to optimize prices for a given assortment of products. We also show some desirable analytical properties of the log-likelihood function, and discuss how our model can be estimated from search path and final choice data. In the fourth Chapter, we generalize the model of discrete choice via sequential search to accommodate heterogeneous consumers. We develop the estimation procedure that can be used to estimate the parameters of the model from the search path and final choice data. We perform empirical analysis to estimate search costs from a real dataset of consumers' clicks and subsequent purchases. We demonstrate that the assumption that researcher places on the distribution of the unobserved utility components in the models where search considerations are present significantly influences the estimates of the search costs.