Choice Models in Marketing examines recent developments in the modeling of choice for marketing and reviews a large stream of research currently being developed by both quantitative and qualitative researches in marketing. Choice in marketing differs from other domains in that the choice context is typically very complex, and researchers' desire knowledge of the variables that ultimately lead to demand in marketplace. The marketing choice context is characterized by many choice alternatives. The aim of Choice Models in Marketing is to lay out the foundations of choice models and discuss recent advances. The authors focus on aspects of choice that can be quantitatively modeled and consider models related to a process of constrained utility maximization. By reviewing the basics of choice modeling and pointing to new developments, Choice Models in Marketing provides a platform for future research.

We develop a demand model for technology products that captures the effect of changes in the portfolio of models offered by a brand as well as the influence of the dynamics in its intrinsic preference on that brand's performance. In order to account for the potential correlation in the preferences of models offered by a particular brand, we use a nested logit model with the brand (e.g., Sony) at the upper level and its various models (e.g., Mavica, FD, DSC, etc.) at the lower level of the nest. Relative model preferences are captured via their attributes and prices. We allow for heterogeneity across consumers in their preferences for these attributes and in their price sensitivities in addition to heterogeneity in consumers' intrinsic brand preferences. Together with the nested logit assumption, this allows for a flexible substitution pattern across models at the aggregate level. The attractiveness of a brand's product line changes over time with entry and exit of new models and with changes in attribute and price levels. To allow for time-varying intrinsic brand preferences, we use a state-space model based on the Kalman filter, which captures the influence of marketing actions such as brand-level advertising on the dynamics of intrinsic brand preferences. Hence, the proposed model accounts for the effects of brand preferences, model attributes and marketing mix variables on consumer choice. First, we carry out a simulation study to ensure that our estimation procedure is able to recover the true parameters generating the data. Then, we estimate our model parameters on data for the U.S. digital camera market. Overall, we find that the effect of dynamics in the intrinsic brand preference is greater than the corresponding effect of the dynamics in the brand's product line attractiveness. Assuming plausible profit margins, we evaluate the effect of increasing the advertising expenditures for the largest and the smallest brands in this category and find that these brands can increase their profitability by increasing their advertising expenditures. We also analyze the impact of modifying a camera model's attributes on its profits. Such an analysis could potentially be used to evaluate if product development efforts would be profitable.

The IBSS is the essential tool for librarians, university departments, research institutions and any public or private institutions whose work requires access to up-to-date and comprehensive knowledge of the social sciences.

In this dissertation, I utilize and develop empirical tools to help academics and practitioners model the consumer's choice process. This collection of three essays strives to answer three main research questions in this theme. In the first paper, I ask: how is the consumer's purchase decision impacted by the search for general product-category information prior to search for their match with a retailer or manufacturer ("sellers")? This paper studies the impact of informational organic keyword search results on the performance of sponsored search advertising. We show that, even though advertisers can target consumers who have specific needs and preferences, for many consumers this is not a sufficient condition for search advertising to work. By allowing consumers to access content that satisfies their information requirements, informational organic results can allow consumers to learn about the product category prior to making their purchase decision. We develop a model characterize the situation in which consumers can search for general information about the product category as well as for information about the individual sellers' offerings. We estimate this model using a unique dataset of search advertising in which commercial websites are restricted in the organic listing, allowing us to identify consumer clicks as informational (from organic links) or purchase oriented (from sponsored links). With the estimation results, we show that consumer welfare is improved by 29%, while advertisers generate 19% more sales, and search engines obtain 18% more paid clicks, as compared to the scenario without informational links. We conduct counterfactuals and find that consumers, advertisers, and the search engine are significantly better off when the search engine provides "free" general information about the product. When the search engine provides information about the advertisers' specific offerings, however, there are fewer paid clicks and advertisers at high ad positions will obtain lower sales. We further investigate the implications on the equilibrium advertiser bidding strategy. Results show that advertiser bids will remain constant in the former scenario. When the search engine provides advertiser information, advertisers will increase their bids because of the increased conversion rate; however, the search engine still loses revenue due to the decreased paid clicks. The findings shed important managerial insights on how to improve the effectiveness of search advertising. In the second paper, I ask: how is the consumer's search for information, during their choice process and in an advertising context, influenced by the signaling theory of advertising? Using a dataset of travel-related keywords obtained from a search engine, we test to what extent consumers are searching and advertisers are bidding in accordance with the signaling theory of advertising in literature. We find significant evidence that consumers are more likely to click on advertisers at higher positions because they infer that such advertisers are more likely to match with their needs. Further, consumers are more likely to find a match with advertisers who have paid more for higher positions. We also find strong evidence that advertisers increase their bids when there is an improvement in the likelihood that their offerings match with consumers' needs, and the improvement cannot be readily observed by consumers prior to searching advertisers' websites. These results are consistent with the predictions from the signaling theory. We test several alternative explanations and show that they cannot fully explain the results. Furthermore, through an extension we find that advertisers can generate more clicks when competing against advertisers with higher match value. We offer an explanation for this finding based on the signaling theory. In the third paper, I ask: can we model the consumer's choice of brand as a sequential elimination of alternatives based on shared or unique aspects while incorporating continuous variables, such as price? With aggregate scanner data, marketing researchers typically estimate the mixed logit model, which accounts for non-IIA substitution patterns among brands, which arise due to similarity and dominance effects in demand. Using numerical examples and analytical illustrations, this research shows that the mixed logit model, which is widely believed to be a highly flexible characterization of brand switching behavior, is not well designed to handle non-IIA substitution patterns. The probit allows only for pair-wise inter-brand similarities, and ignores third-order or higher dependencies. In the presence of similarity and dominance effects, the mixed logit model and the probit model yield systematically distorted marketing mix elasticities. This limits the usefulness of mixed logit and probit for marketing decision-making. We propose a more flexible demand model that is an extension of the elimination-by-aspects (EBA) model (Tversky 1972a, 1972b) to handle marketing variables. The model vastly expands the domain of applicability of the EBA model to aggregate scanner data. Using an analytical closed-form that nests the traditional logit model as a special case, the EBA demand model is estimated with marketing variables from aggregate scanner data in 9 different product categories. It is compared to the mixed logit and probit models on the same datasets. In terms of multiple fit and predictive metrics (LL, BIC, MSE, MAD), the EBA model outperforms the mixed logit and the probit in a majority of categories in terms of both in-sample fit and holdout predictions. The results show significant differences in the estimated price elasticity matrices between the EBA model and the comparison models. In addition, a simulation shows that the retailer can improve gross profits up to 34.4% from pricing based on the EBA model rather than the mixed logit model. Finally, the results suggest that empirical IO researchers, who routinely use mixed logit models as inputs to oligopolistic pricing models, should consider the EBA demand model as the appropriate model of demand for differentiated products.