Modeling preference evolution in discrete choice models: A Bayesian state-space approach

We develop discrete choice models that account for parameter driven preference dynamics. Choice model parameters may change over time because of shifting market conditions or due to changes in attribute levels over time or because of consumer learning. In this paper we show how such preference evolution can be modeled using hierarchial Bayesian state space models of discrete choice. The main feature of our approach is that it allows for the simultaneous incorporation of multiple sources of preference and choice dynamics. We show how the state space approach can include state dependence, unobserved heterogeneity, and more importantly, temporal variability in preferences using a correlated sequence of population distributions. The proposed model is very general and nests commonly used choice models in the literature as special cases. We use Markov chain monte carlo methods for estimating model parameters and apply our methodology to a scanner data set containing household brand choices over an eight-year period. Our analysis indicates that preferences exhibit significant variation over the time-span of the data and that incorporating time-variation in parameters is crucial for appropriate inferences regarding the magnitude and evolution of choice elasticities. We also find that models that ignore time variation in parameters can yield misleading inferences about the impact of causal variables. This paper is based on the first author's doctoral dissertation.     

Coalition Formation Procedures: The Impact of Issue Saliences and Consensus Estimation

Coalition formation procedures incorporate two properties that are not often found in other coalition formation models: the choice between different formation paths and constrained consensus positions. However, there are two aspects of coalition formation procedures that are often overlooked: issue saliences and consensus estimation. Issue saliences are a measure of the importance that parties can attribute to issue dimensions. Initially, we employ the classical application to implement issue saliences. The classical application combines the Euclidean distance with the center of gravity as a consensus estimate. Secondly, we introduce a consistent distance application where the coalition consensus position is determined by minimizing the sum of salience-weighted Euclidean distances. The impact of these aspects is examined with the help of both numerical and empirical applications. The results indicate that both the consensus estimation method and the inclusion of issue saliences do not only have an impact on the estimated consensus position. They also determine the individual parties’ preferences towards the potential coalition formation procedures.     

Linking Sources of Consumer Confusion to Decision Satisfaction: The Role of Choice Goals

Previous empirical research on the relationship between consumer confusion and customer satisfaction has largely neglected the role of choice goals. In a context of technologically complex products, the authors analyze the effect of selected choice goals on the consumer confusion‐decision satisfaction link. The empirical findings, which are based on a field study of smart phone users, show that different sources of confusion have distinctive effects on choice goals, which in turn influence decision satisfaction. For example, while confusion caused by ambiguous information and choice overload is found to reduce choice confidence, perceived attribute similarity between products or brands increases choice confidence. Among the choice goals, evaluation costs and negative affect are found to increase decision satisfaction. The findings have important implications for marketers and consumer policymakers in terms of marketing communication and customer satisfaction.     

Different strokes for different folks: A method to accommodate decision -making heterogeneity

Consumer decision making is complex and no single perspective offers a complete theory of consumer decision making. While the research community acknowledges that there is heterogeneity, homogenous choice models dominate consumer decision research. This paper provides insights from one method that was designed to accommodate decision -making heterogeneity. Computer process tracing methods can be used to observe different consumer decisions in one product category to understand what and how people choose. More than two-hundred and fifty decisions were observed in this research. Consumers were asked to select one of nine air conditioner alternatives described with six salient attributes. The research findings clearly reveal consumer differences. Specifically, the attributes and decision types used differed resulting in different product choices. This paper reveals how methods that accommodate decision -making heterogeneity can be used by retailers to inform product ranging decisions for categories.     

Integrated mixed logit and latent variable models

Extending the traditional discrete choice model by incorporating latent psychological factors can help to better understand the individual’s decision-making process and therefore to yield more reliable part-worth estimates and market share predictions. Several integrated choice and latent variable (ICLV) models which merge the conditional logit model with a structural equation model exist in the literature. They assume homogeneity in the part-worths and use latent variables to model the heterogeneity among the respondents. This paper starts from the mixed logit model that describes the heterogeneity in the part-worths and uses the latent variables to decrease the unexplained part of the heterogeneity. The empirical study presented here shows these ICLV models perform very well with respect to model fit and prediction.     

Using a Community of Knowledge to Build Intelligent Agents

The modeling of individual consumer preference can be aided by incorporating others' opinions which contain information above and beyond identified product attributes. The value of others' opinions is tested using two empirical data sets. The results indicate that incorporating others' opinions into an attribute-based model can reduce systematic error and increase predictive accuracy by serving as a proxy for missing information (e.g., undiscovered attributes or attribute interactions, sensory or experiential aspects of the product, as well as advertising or word of mouth effects). Additionally, modeling individual preference based on others' opinions alone is shown to predict as well or better than traditional multiattribute models thus bypassing the need for defining a product attribute space.     

Enhancing marketing with engineering: Optimal product line design for heterogeneous markets

Successful product line design and development often require a balance of technical and market tradeoffs. Quantitative methods for optimizing product attribute levels using preference elicitation (e.g., conjoint) data are useful for many product types. However, products with substantial engineering content involve critical tradeoffs in the ability to achieve those desired attribute levels. Technical tradeoffs in product design must be made with an eye toward market consequences, particularly when heterogeneous market preferences make differentiation and strategic positioning critical to capturing a range of market segments and avoiding cannibalization.We present a unified methodology for product line optimization that coordinates positioning and design models to achieve realizable firm-level optima. The approach overcomes several shortcomings of prior product line optimization models by incorporating a general Bayesian account of consumer preference heterogeneity, managing attributes over a continuous domain to alleviate issues of combinatorial complexity, and avoiding solutions that are impossible to realize. The method is demonstrated for a line of dial-readout scales, using physical models and conjoint-based consumer choice data. The results show that the optimal number of products in the line is not necessarily equal to the number of market segments, that an optimal single product for a heterogeneous market differs from that for a homogeneous one, and that the representational form for consumer heterogeneity has a substantial impact on the design and profitability of the resulting optimal product line — even for the design of a single product. The method is managerially valuable because it yields product line solutions efficiently, accounting for marketing-based preference heterogeneity as well as engineering-based constraints with which product attributes can be realized.     

Consumers' SKU choices in an online supermarket: a latent class approach

Abstract

This study examines the effects of loyalty and e-marketing mix variables on the choices of online consumers at the stock-keeping-unit (SKU) level. Using a panel dataset from an online supermarket, we estimate a discrete choice model of a frequently purchased product; generate the refined smoothing constants of the loyalty variables for brand, size, and SKU; and adopt the latent class approach to address consumer heterogeneity. The findings suggest that SKU loyalty is a better predictor of consumer choices than brand and size loyalty. Although online consumers are not sensitive to the net prices of SKU alternatives, they are attracted to price promotions. While webpage display has little effect on SKU choices, speedy delivery has a positive impact. The latent class approach significantly improves model fitness and classification accuracy. Analysing consumer choices at the SKU level can help online supermarkets with promotion planning and inventory and distribution management to improve customer satisfaction and profitability.     

Consumer strategic decision making and choice process: prefabricated house purchase

This study examines the consumer choice process in case of strategic purchases, such as house buying. In view of the existing literature exploring consumer decision making and choice for strategically important products, the purpose of this research is twofold: (a) to develop a conceptual model of strategic decision making; and (b) to empirically explore this model with regard to prefabricated house purchases. The results of our qualitative research suggest that in addition to the idiosyncratic characteristics of the customer, his or her personal situation, environmental factors, the role of feelings, experience, subconscious factors, needs, and goals should to be taken into account to better understand strategic consumer decision making and their choice process when buying a house.     

Market strategy for promoting green consumption: Consumer preference and policy implications for laundry detergent

Green consumption is a very common phrase in our daily lives, yet product characteristics that mainly contribute to the diffusion of green products are largely unknown. Based on microeconomic theory, we conduct a conjoint survey of consumer preferences for a ubiquitous green product—laundry detergent. We analyze the correlation between consumers' demographic variables and attributes of laundry detergents through a hierarchical Bayesian mixed logit model. We find that consumer preferences for attributes display significant heterogeneity. Age and income significantly influence the marginal preferences for attributes. An examination of consumer willingness to pay and of the relative importance of each attribute reveals that price and base material are the most important attributes. Green attributes, such as skin irritation potential and biodegradability, tend to be less important. This study also examines preference heterogeneity based on previous purchase experience. To promote green consumption, we emphasize the need for policies that reduce the value‐action gap.     

Hybrid Choice Models: Progress and Challenges

We discuss the development of predictive choice models that go beyond the random utility model in its narrowest formulation. Such approaches incorporate several elements of cognitive process that have been identified as important to the choice process, including strong dependence on history and context, perception formation, and latent constraints. A flexible and practical hybrid choice model is presented that integrates many types of discrete choice modeling methods, draws on different types of data, and allows for flexible disturbances and explicit modeling of latent psychological explanatory variables, heterogeneity, and latent segmentation. Both progress and challenges related to the development of the hybrid choice model are presented.     

Estimating time-varying parameters in brand choice models: A semiparametric approach

Nowadays, brand choice models are standard tools in quantitative marketing. In most applications, parameters representing brand intercepts and covariate effects are assumed to be constant over time. However, marketing theories, as well as the experience of marketing practitioners, suggest the existence of trends or short-term variations in particular parameters. Hence, having constant parameters over time is a highly restrictive assumption, which is not necessarily justified in a marketing context and may lead to biased inferences and misleading managerial insights.In this paper, we develop flexible, heterogeneous multinomial logit models based on penalized splines to estimate time-varying parameters. The estimation procedure is fully data-driven, determining the flexible function estimates and the corresponding degree of smoothness in a unified approach. The model flexibly accounts for parameter dynamics without any prior knowledge needed by the analyst or decision maker. Thus, we position our approach as an exploratory tool that can uncover interesting and managerially relevant parameter paths from the data without imposing assumptions on their shape and smoothness.Our approach further allows for heterogeneity in all parameters by additively decomposing parameter variation into time variation (at the population level) and cross-sectional heterogeneity (at the individual household level). It comprises models without time-varying parameters or heterogeneity, as well as random walk parameter evolutions used in recent state space models, as special cases. The results of our extensive model comparison suggest that models considering parameter dynamics and household heterogeneity outperform less complex models regarding fit and predictive validity. Although models with random walk dynamics for brand intercepts and covariate effects perform well, the proposed semiparametric approach still provides a higher predictive validity for two of the three data sets analyzed.For joint estimation of all regression coefficients and hyperparameters, we employ the publicly available software BayesX, making the proposed approach directly applicable.     

Random regret minimization for consumer choice modeling: Assessment of empirical evidence

This paper introduces to the field of marketing a regret-based discrete choice model for the analysis of multi-attribute consumer choices from multinomial choice sets. This random regret minimization (RRM) model, which has recently been introduced in the field of transport, forms a regret-based counterpart of the canonical random utility maximization (RUM) paradigm. This paper assesses empirical results based on 43 comparisons reported in peer-reviewed journal articles and book chapters, with the aim of finding out to what extent, when, and how RRM can form a viable addition to the consumer choice modeler's toolkit. The paper shows that RRM and hybrid RRM–RUM models outperform RUM counterparts in a majority of cases, in terms of model fit and predictive ability. Although differences in performance are quite small, the two paradigms often result in markedly different managerial implications due to considerable differences in, for example, market share forecasts.     

Need for cognitive closure and information search strategy

The authors investigated whether need for cognitive closure (NFCC) affected one's style of information search (attribute‐based search vs. alternative‐based search) in consumer choice. There has been growing interest in NFCC in marketing and its relationship to consumer information processing. However, no study to date has examined the different information search strategies that consumers employ when they (1) possess different levels of NFCC or (2) are exposed to situations that evoke more or less NFCC. Across two studies where Study 1 measured NFCC while Study 2 manipulated NFCC, the authors commonly found that higher NFCC compared to lower NFCC resulted in (1) a greater preference for the attribute‐based search over the alternative‐based search and (2) a consideration of a smaller amount of information to make a final choice. Implications for consumer information processing and sales strategies are discussed along with future research directions. © 2008 Wiley Periodicals, Inc.     

Profiling the reference price consumer

Researchers in marketing have devoted considerable attention to understanding how price impacts the purchase decision. Some individuals, termed memory-based reference price (MBR) consumers, take into account price expectations developed from past purchase behavior when making a current choice. Other individuals, termed stimulus-based reference price (SBR) consumers, make choices by constructing a reference point from the currently observed distribution of prices. Using a latent class model of structural heterogeneity applied to purchase histories from the toilet tissue category, we classify households in terms of the pricing mechanism used in buying decisions. We find strong evidence that memory-based (internal) reference price consumers are more price sensitive than other consumers. Moreover, we find that variables associated with the accessibility of price information are predictive of consumer use of memory-based reference prices. Managerial implications of these results are discussed.     

Modeling the effects of including/excluding attributes in choice experiments on systematic and random components

This paper examines the impact of attribute presence/absence in choice experiments using covariance heterogeneity models and random coefficient models. Results show that attribute presence/absence impacts both mean utility (systematic components) and choice variability (random components). Biased mean effects can occur by not accounting for choice variability. Further, even if one accounts for choice variability, attribute effects can differ because of attribute presence/absence. Managers who use choice experiments to study product changes or new variants should be cautious about excluding potentially essential attributes. Although including more relevant attributes increases choice variability, it also reduces bias.     

Overcoming data limitations for store choice modelling.: Exploiting retail chain choice data by means of aggregate logit models

This paper reports on research aimed at exploiting certain data sources for store choice modelling purposes. Many databases, such as some consumer panels, only record the firm chosen by consumers and not the specific store at which they shop. Four alternative approaches are proposed in order to use this raw information for studying patronage determinants at store level: (a) an ordinary logit model in which chain utility is averaged across stores within; (b) an ordinary logit model in which the choice set is assumed to be composed of the nearest store for each chain; (c) a straightforward application of an aggregate logit model; and (d) the application of an aggregate logit model with choice sets spatially bounded by a distance threshold representing the maximum distance that consumers are willing to travel for shopping. The models are empirically tested in the context of spatial choice behaviour. Goodness of fit indicators reveal that only models (b), (c) and (d) acceptably represent competitive interaction dynamics. As performance of (b) is slightly better than that of (c), it seems that a priori the ‘nearest store assumption’ is a better approach than the modelling of aggregate choice structures. However, when the latter approach is applied with more reliable choice sets, as suggested in model (d), the best performance is achieved. The results thus lead us to think that the aggregate logit model is a promising methodology for solving the problem at issue, but subject to an appropriate definition of the consumers’ choice sets. In fact, such an approach provides a more suitable modelling solution to the extent that the saturation and the intra-firm store heterogeneity become more intense, because these situations presumably imply that consideration sets include several stores from the same chain.     

Decisional Power in Group Decision Making: A Note on the Allocation of Group Members' Weights in the Multiplicative AHP and SMART

Many complex decisions are made in a group environment, where the decision is made jointly by a committee or group structure. The individual group members are often not equally qualified to contribute equitably to the decision process, or may have different saliences (desires) to influence the decision. A quantitative knowledge of the players' decisional power is useful for better understanding of the group decision process, and could even be used in weighted voting within the group structure. We adapt the REMBRANDT suite of decision models (multiplicative AHP and SMART) to measure decisional power in groups, and we generalise this to cater for the case where power itself is deemed to be multidimensional in nature, and the case of uncertain subjective judgements of power amongst group members.     

Bayesian Analysis of Simultaneous Demand and Supply

In models of demand and supply, consumer price sensitivity affects both the sales of a good through price, and the price that is set by producers and retailers. The relationship between the dependent variables (e.g., demand and price) and the common parameters (e.g., price sensitivity) is typically non-linear, especially when heterogeneity is present. In this paper, we develop a Bayesian method to address the computational challenge of estimating simultaneous demand and supply models that can be applied to both the analysis of household panel data and aggregated demand data. The method is developed within the context of a heterogeneous discrete choice model coupled with pricing equations derived from either specific competitive structures, or linear equations of the kind used in instrumental variable estimation, and applied to a scanner panel dataset of light beer purchases. Our analysis indicates that incorporating heterogeneity into the demand model all but eliminates the bias in the price parameter due to the endogeneity of price. The analysis also supports the use of a full information analysis.     

Decision neuroscience and consumer decision making

This article proposes that neuroscience can shape future theory and models in consumer decision making and suggests ways that neuroscience methods can be used in decision-making research. The article argues that neuroscience facilitates better theory development and empirical testing by considering the physiological context and the role of constructs such as hunger, stress, and social influence on consumer choice and preferences. Neuroscience can also provide new explanations for different sources of heterogeneity within and across populations, suggest novel hypotheses with respect to choices and underlying mechanisms that accord with an understanding of biology, and allow for the use of neural data to make better predictions about consumer behavior. The article suggests that despite some challenges associated with incorporating neuroscience into research on consumer decision processes, the use of neuroscience paradigms will produce a deeper understanding of decision making that can lead to the development of more effective decision aids and interventions.