Abstract: In recent years, many U.S. and Japanese firms have adopted Quality Function Deployment QFD. QFD is a total-quality-management process in which the "voice of the customer" is deployed throughout the R&D, engineering, and manufacturing stages of product development. For example, in the first "house" of QFD, customer needs are linked to design attributes thus encouraging the joint consideration of marketing issues and engineering issues. This paper focuses on the "Voice-of-the-Customer" component of QFD, that is, the tasks of identifying customer needs, structuring customer needs, and providing priorities for customer needs. In the identification stage, we address the questions of 1 how many customers need be interviewed, 2 how many analysts need to read the transcripts, 3 how many customer needs do we miss, and 4 are focus groups or one-on-one interviews superior? In the structuring stage the customer needs are arrayed into a hierarchy of primary, secondary, and tertiary needs. We compare group consensus affinity charts, a technique which accounts for most industry applications, with a technique based on customer-sort data. In the stage which provides priorities we present new data in which product concepts were created by product-development experts such that each concept stressed the fulfillment of one primary customer need. Customer interest in and preference for these concepts are compared to measured and estimated importances. We also address the question of whether frequency of mention can be used as a surrogate for importance. Finally, we examine the stated goal of QFD, customer satisfaction. Our data demonstrate a self-selection bias in satisfaction measures that are used commonly for QFD and for corporate incentive programs. We close with a brief application to illustrate how a product-development team used the voice of the customer to create a successful new product.

Abstract: This paper presents a technique for estimating a firm's brand equity that is based on the financial market value of the firm. Brand equity is defined as the incremental cash flows which accrue to branded products over unbranded products. The estimation technique extracts the value of brand equity from the value of the firm's other assets. This technique is useful for two purposes. First, the macro approach assigns an objective value to a company's brands and relates this value to the determinants of brand equity. Second, the micro approach isolates changes in brand equity at the individual brand level by measuring the response of brand equity to major marketing decisions. Empirically, we estimate brand equity using the macro approach for a sample of industries and companies. Then we use the micro approach to trace the brand equity of Coca-Cola and Pepsi over three major events in the soft drink industry from 1982 to 1986.

Abstract: The PIMS Profit Impact of Marketing Strategies data entail sparse time-series observations for a large number of strategic business units SBUs, In order to estimate disaggregate marketing mix elasticities of demand, a natural solution is to pool different SBUs The traditional, a priori approach is to pool together those SBUs which one believes in advance to be very similar with respect to their marketing mix elasticities We propose an alternative maximum likelihood, latent-pooling method for simultaneously pooling, estimating, and testing linear regression models empirically This method enables the determination of a "fuzzy" pooling scheme, while directly estimating a set of marketing mix elasticities and intertemporal covariances for each pool of SBUs Our analyses reveal different magnitudes and patterns of marketing mix elasticities for the derived pools Pool membership is influenced by demand characteristics, business scope, and order of market entry

Abstract: Based upon a recently developed multiattribute generalization of prospect theory's value function Tversky and Kahneman 1991, we argue that consumer choice is influenced by the position of brands relative to multiattribute reference points, and that consumers weigh losses from a reference point more than equivalent sized gains loss aversion. We sketch implications of this model for understanding brand choice. We develop a multinomial logit formulation of a reference-dependent choice model, calibrating it using scanner data. In addition to providing better fit in both estimation and forecast periods than a standard multinomial logit model, the model's coefficients demonstrate significant loss aversion, as hypothesized. We also discuss the implications of a reference-dependent view of consumer choice for modeling brand choice, demonstrate that loss aversion can account for asymmetric responses to changes in product characteristics, and examine other implications for competitive strategy.

Abstract: We develop a comprehensive utility maximizing framework to study the impact of marketing variables on the category purchase, brand choice and purchase quantity decisions of households for frequently purchased packaged goods. The model allows for dependence among the three decisions while ensuring that these decisions provide, in combination, the greatest possible utility to the household. By accounting for variations in reservation prices and intrinsic brand preferences across households, the modeling framework explicitly captures the effects of unobserved heterogeneity on all three purchase decisions. The principal empirical finding from analyzing the A. C. Nielsen data for the yogurt product category is that the substantive implications for the effects of marketing variables are sensitive to whether these effects are determined conditional or unconditional on a product category purchase. Our results show that reservation prices and intrinsic brand preferences vary across households, and not accounting for these variations in the estimation could lead to biased estimates for the coefficients of the marketing variables. A comparison of our results to those obtained from a nested logit model of purchase incidence and brand choice reveals that our proposed model performs better using both a formal statistical test as well as the criterion of predictive validity in a holdout sample of panelists. Further, the purchase quantity model compares favorably with two alternative models of quantity choice in the validation sample.

Abstract: "Single Source" databases based on scanner data offer new opportunities for evaluating promotions and improving their effectiveness. Decision support needs vary depending on the decision maker's organizational vantage point. Some managers require the evaluation of promotion results in the short term. Others should take a medium-to long-term focus. An implemented model and automated system for measuring short-term incremental volume due to promotions by developing baselines of store-level "normal" sales is presented using store-level scanner data. Empirical validation results and real life applications are presented and discussed, including the use of the baselines as measures of "brand health."

Abstract: This paper characterizes the manufacturer warranty policy and its effect on consumer behavior under the following conditions: consumers are heterogeneous in risk-preferences, consumer actions affecting the probability of warranty redemption are unobservable to the manufacturer, and the product reliability is known. We obtain the "menu" of warranty contracts, and then make connections with its institutional counterpart: the extended service contract. The model's implications for consumer behavior are examined using data obtained from a sample of recent buyers of new cars. The role of risk in consumer behavior with respect to choice of extended service contracts, and the allocation of effort for maintenance are found to be consistent with the model's predictions. The empirical analysis permits quantifying the demand for extended service contracts as a function of the extent of manufacturer warranty. The estimates show that for our sample of buyers a manufacturer warranty of three years is optimal in the sense of overcoming the role of risk-aversion in the choice of extended service contracts.

Abstract: Using data on search and choice behavior from a local automobile market, we estimate monetary returns to search in terms of lower prices resulting from additional time invested in price search. For our analytical framework, we adapt a model developed in the job search literature to the problem of consumer search; this framework is especially useful for illuminating the relationship between time spent searching, the outcome of search, and demand and supply side variables. Our results indicate that, for this particular sample of buyers, marginal returns to search are broadly consistent with what one might expect if consumers balance costs and benefits of search, and that potential gains from additional search for lower car prices do not appear to be large for most consumers. Our study highlights many of the methodological difficulties involved in estimating returns to search, including isolating returns to different outcomes of search, and sensitivity of results to model specification and sampling error. We deal with these problems by trying to isolate time spent searching for price from other uses of search time, by deriving our model used in estimation from a specific conceptual framework, and by extensive specification testing.

Abstract: Recent research suggests that the signal e.g., sign or marker with a point of purchase promotion will stimulate a significant sales increase, regardless of whether or not that signal is accompanied by a price cut. This paper develops a model of retailer profitability that incorporates this "promotion signal sensitivity." In a field test, the profitability of the promotion policy prescribed by this model is compared to the profitability of two other promotion policy-setting paradigms: a model-based policy that does not consider promotion signal sensitivity and one prescribed by industry experts. The test results support the proposed model. Its policy generates 11% more category profit per unit than the model-based policy and 12% more than the industry experts. Implications for retailers and future research are discussed.

Abstract: Given that firms pulse in advertising, should firms pulse in or out of phase? It is shown that out of phase maximizes the oligopoly profits and is also the Markov perfect equilibrium of the infinite horizon game. The basic intuition for this result comes from the following fact: it is more profitable to increase consideration when the competitor's consideration is lower. Evidence from several product categories seems to support this theoretical result.

Abstract: In this paper we develop a model relating market share to average costs. We start with a theoretical model of the factors that affect the firm's average cost curve, partitioning these factors into a measurable firm and competitive environment characteristics, and b unobserved factors that are either fixed, random, or follow a first-order autoregressive process. We then link this theoretical model to an empirical model in which we specify three average cost equations for the organizational areas of purchasing, production, and marketing. Main effects for initial lagged market share position, as well as their interactions with factors characterizing the firm's competitive environment, represent the variables of key theoretical interest in our equations. We estimate these equations using PIMS data, and control for fixed, contemporaneous, and autoregressive unobservable factors. Our results suggest that market share can often lead to market power in the form of lower average costs. However, the firm's operating environment greatly moderates the effect of market share on average cost. In particular, we find that market share position only leads to lower average costs when the organizational unit operates in a competitive environment that gives it both motivation and ability to realize power from its market share position.

Abstract: A set of alternatives under consideration is often divided into subsets or local sets by some external e.g., product display format at the store or internal e.g., a decision rule factor. We propose that the manner in which a global set of alternatives varying in price and quality is divided into local sets can have a systematic effect on consumer choice between a lower price alternative and a higher quality alternative. Three such effects are investigated: 1 Consumers who first choose from pairs of products {A, B}, {B, C}, {A, C} and then choose from the set of all three products {A, B, C} are more likely to prefer the cheapest alternative than consumers who only choose from the complete set; 2 The choice share of a low-price, low-quality brand is greater when alternatives varying in price, quality, and features are displayed by brand i.e., each display presents different models of one brand as compared to a display by feature level i.e., each display presents comparable models of different brands; and 3 Consumers considering a pair of two-option local sets, each consisting of different brands and feature levels e.g., a feature enhanced model of a low quality brand vs. a basic model of a high quality brand, are more likely to select a feature enhanced model from the global set of four options than those who consider the same alternatives in other local set configurations or consider only the global set. These predictions were supported in seven studies, which also provided insights into the boundaries of the effects focusing on the paired-comparison effect. We discuss the theoretical and practical implications of the findings.

Abstract: Publisher Summary This chapter describes some tangible benefits that emerge from coexistence and outlines some techniques in management science that have been developed to address this issue. At this point, it seems worthwhile to consider some additional perspectives in order to identify existing gaps that may offer further research opportunities. Few of the models incorporate competition, undoubtedly because the fact that the existence of both production and marketing decisions creates models which are already considerably complex. Despite the daunting nature of competitive formulations, these would be well worth investigating. Another dimension which should be looked into is the case of multiple products. Such models may be quite difficult to analyze, however, with the existing tools.

Abstract: Publisher Summary The chapter describes quantitative modelers and management scientists unfamiliar with marketing an appreciation of the way in which models of consumer behavior are developed and used. The chapter is also designed to provide a reference and teaching resource for marketing specialists. The future of consumer behavior modeling is bright; newer models are richer, more flexible, and more closely attuned to modern data sources. Yet many phenomena are poorly modeled at the moment. The chapter highlights the areas of modeling consumer purchase heuristics (and information-processing biases), modeling consumers' mental processes, matching models to market segments, and modeling choice for truly new or non-comparable alternatives as fruitful areas that deserve concerted attention in the future.

Abstract: Publisher Summary Marketing has seen a rapid expansion in the widespread use of quantitative methods. Correlation and regression analysis were among the first techniques used as marketing research emerged as a discipline after World War II. In the early 1970s regression analysis became econometrics. Simultaneous equation systems could be estimated almost as easily as single regression equations. While econometrics as a whole continues to flourish as new and more sophisticated estimation techniques and associated computer software have become available, simultaneous-equation systems have not become widely prevalent. The dynamic structure of marketing variables themselves is addressed in the chapter, followed by discussions of leads and lags among marketing variables and the assessment of the direction of causality. Dynamic properties of sales-response functions have been discussed in more detail. Marketing generalizations that have been uncovered as well as empirical evidence on the shape of the sales response function is reported.

Abstract: In this paper, a formal test on prediction errors is developed for the cross-validation of regression models under the simple random splitting framework. Analytic as well as simulation results relate the statistical power of the test to the allocation of sample observations to estimation and validation subsets. The results indicate that splitting the data into halves is suboptimal. More observations should be used for estimation than validation. Furthermore, the proportion of the sample optimally devoted to validation is small for very limited samples N 60. However, although the 50/50 split is suboptimal, it is not tremendously so in a wide variety of circumstances.

Abstract: Publisher Summary This chapter provides the OR/MS researcher with an overview of conjoint's origins, foundations and progress, culminating in prescriptive models for optimal product-positioning. Its evolution has moved beyond initial preoccupation with utility measurement and buyer-choice simulations to interest in product design, market segmentation, and competitive strategy. OR/MS researchers are probably less knowledgeable with the more plebeian methodology of conjoint analysis, a multi attribute utility-measurement approach applied primarily by marketing researchers. Conjoint researchers are usually concerned with the more day-to-day decisions of consumers what brand of soap, automobile, phone service, photocopy machine to buy. While, in principle, conjoint methodology can be used to measure corporate administrators' multi attribute values, in most applications this is not the case.

Abstract: Economies of scale are evident when a firm's average costs decline while its output expands, as when an advertising agency raises its gross income by serving more accounts and/or larger accounts. Economies of scope appear when cost savings can be realized by a single agency producing several products jointly, as compared to many agencies each producing them separately. How important are economies of scale and scope in advertising agency operations? In this paper cost models are formulated which represent how the principal component of agency costs, employment level, varies according to the mix of media and services an agency provides and the total volume of advertising it produces. These models are estimated and tested cross-sectionally utilizing data pertaining to the domestic operations of 401 U.S. agencies for 1987. The empirical evidence reported here indicates that both scale and particularly scope economies are highly significant in the operations of U.S. advertising agencies. We find that of the 12,000 establishments comprising the industry in 1987, approximately 200-250 had domestic gross incomes of $3-4 million or more or equivalently, billings of $20-27 million and therefore had service mixes and operating levels sufficiently large to take full advantage of all available size-related efficiencies. Furthermore, the overall structure of the industry is one where these large, fully efficient firms created and produced more than half of all the national advertising utilized in the U.S. during 1987. At the same time, vast numbers of very small agencies appear to operate with substantial cost disadvantages compared to large firms as a consequence of these scale and scope economies. These findings carry important implications concerning possible future changes in the industry structure. It seems highly doubtful that scale economies could motivate further mergers among the largest 200-250 agencies. On the other hand, for small agencies, mergers and acquisitions might be attractive as means of mitigating their size-related cost disadvantages. Finally, our findings demonstrating the existence of scale and scope economies are consistent with the diminishing reliance on fixed rates of media commissions as the principal basis of agency compensation. They also cast strong doubts on size-related economies in operating costs as a viable explanation for the limited degree of vertical integration of agency services by large advertisers.

Abstract: Publisher Summary Modeling has a place in the understanding and planning of sales promotion, and much progress has been recently made. However, there are still many unanswered questions, even regarding some of the areas that have been thoroughly researched. In addition, one should begin to see a shift toward prescriptive models and hopefully an increase in managerial use of models. The potential need is there, and so is the technology. If it is successful, this chapter, by summarizing the field, helps to accelerate progress in the development and implementation of sales-promotion models.

Abstract: Publisher Summary This chapter presents the first attempt in marketing to assemble the vast literature dealing with a variety of non-spatial tree models for the assessment of competitive market structure The chapter also presents taxonomy of the non-spatial approaches using two major dimensions, namely data and model characteristics The data characteristics include: type of data (behavioral or judgmental), level of aggregation (individual, segment, or market), and measure of competition relation (perceptions, switching, or preference/choice) The model characteristics include representation (hierarchical/ultrametric, additive, networks, or latent class), type of analysis (confirmatory or exploratory), ability to accommodate marketing-mix effects, various modes of analysis, and error specifications The chapter analyzes the brand-switching data on eight brands of soft drinks collected by Bass, Pessemier & Lehmann using various methodologies for assessment of non-spatial tree-type competitive market structure and discussed insights provided by various methods and their limitations The chapter also discusses stochastic tree models

Abstract: Publisher Summary This chapter describes that three basic principles motivate the specification of the market-share models. Market-share models should be competitive, descriptive as well as predictive, and profit-oriented. Being fundamentally competitive implies that one cannot know the effect or effectiveness of a marketing action without accounting for the actions of competitors. Market-share models are models for understanding how the marketing efforts of every brand impact the results in a competitive marketplace. Only by describing the influence of each marketing instrument can one gain a basis for marketing planning. Prediction alone is not enough. Time-series models that forecast the future from the past sales provide no insight into how sales are generated. The emphasis on being descriptive also embraces the need to understand the areas in which consumer choice probabilities are synonymous with market shares as well. Part of the goal of description transcends what can be done by market-share models alone. Managers need to understand that their efforts have (potentially) competitive effects and (potentially) market-expansive effects. In sales-response models these effects are comingled, but by combining descriptive market-share models for the competitive effects with descriptive category-volume models for the market-expansive effects, managers obtain a much richer understanding of the market. The profit-oriented goal of market-share analysis urges us to ask how the firm's allocations of resources to aspects of the marketing mix produce bottom-line results.

Abstract: Probabilistic discrete-choice models, such as multinomial logit models, are widely used to predict changes in market shares or total demand resulting from changes in policy variables under management control. These models often are evaluated in terms of their ability to predict choices in a holdout sample. This paper presents a new test for comparing predicted and observed choices. The results of a Monte Carlo experiment indicate that the new test has good finite-sample properties and high power in several circumstances likely to arise frequently in applications.

Abstract: The paper argues that: (i) A successful market segmentation study requires the application of structure and a proper integration of qualitative and quantitative research. Also, that the prime emphasis should be on actionability. (ii) Lifestyle and value-based segmentation are too general to be of great use in category specific studies. Also, their international application is limited as lifestyles vary, internationally. (iii) The growing phenomenon of 'consumer' schizophrenia' makes lifestyle increasingly less useful in market segmentation. (iv) Well-developed psychological segmentations, that relate to a specific market or product category, are more diagnostic, more predictive and more actionable, on a global basis.

Abstract: Publisher Summary This chapter illustrates the importance and prevalence of group choice in marketing Few choices are made by individuals truly independent of others and many are explicitly joint The examples in the chapter are designed to illustrate the hazards of assuming that most choices are independent Purchase histories can be misinterpreted if they are assumed to belong to a single individual and actually reflect the preferences of multiple family members, buyers, and sellers often negotiate issues that have been assumed to be set by one party and either accepted or rejected by the other, and strategy is more often formulated by formal or informal management teams than by individuals that may result in inaccurate choice predictions for them

Abstract: Publisher Summary This chapter brings together a set of diverse efforts in the recent literature on the modeling of price decisions and related questions in marketing. There has been an impressive growth in the array of topics investigated in the literature. Various trends in the development of pricing models are evident from the foregoing review. First, one trend has been to develop theoretical models to describe observed pricing strategies in the marketplace and to derive conditions under which certain strategies are optimal. This trend is clearly evident when one considers the area of dynamic pricing models. Another trend is an attempt to develop pricing models in which certain aspects of consumer behavior (e.g asymmetric response to price increases versus price decreases) are incorporated. This development is quite recent and does offer a large potential. A third dominant direction is the use of game-theoretic models for developing equilibrium pricing strategies.

Abstract: Publisher Summary This chapter provides a critical perspective on one of the new developments, namely, artificial intelligence (AI) modeling, as it compares to decision modeling using conventional OR/MS techniques. Decision models are playing an increasingly important role in supporting management decision-making. Little has noted that a problem-solving technology is emerging that consists of people, knowledge, software, and hardware successfully wired into the management process'. Even since then, developments in modeling and computer technologies have created new opportunities for marketing scientists to develop decision models that can significantly influence marketing decision-making. The chapter describes how decision models and characterizing differ from theoretical models. It also identifies and summarizes some key concepts underlying conventional decision models in marketing.

Abstract: Publisher Summary When the term ‘marketing’ comes to mind, many people think of ‘pet rocks', cans of ‘New York City air’, and the cyclical movement of hemlines in women's fashions; the analysis of the demand for such items seems well removed from the reliance on mathematical models that characterizes much of the work in operations research and management science (OR/MS). Indeed, many company executives despair of putting marketing on a more scientific basis. Many see marketing processes as lacking the neat quantitative properties found in production and finance. In marketing, human factors play a large role, marketing expenditures affect demand and cost simultaneously and information to support truly systematic decisions is rarely available, Further, the effects of most marketing actions are typically delayed, nonlinear, stochastic, and difficult to measure.

Abstract: Farris, Parry and Ailawadi 1992; hereafter denoted FPA demonstrate that bias can arise in a regression involving a composite dependent variable where a subset of components of the dependent variable are used as explanatory factors. They correctly observe that the Jacobson and Aaker 1985; hereafter denoted JA model has explanatory factors that are also components of the ROI dependent variable and, as such, is subject to "composite variable bias." FPA note that another way of viewing composite variable bias is that the coefficients in the model reflect not their impact on the dependent variable but rather their impact on the dependent variable less the elements of the components included as explanatory factors. As such, additional effects analogous to indirect effects may be present to the extent strategic factors influence the included components. FPA conclude that such bias explains the low estimate of the market share effect reported in JA. However, FPA's attempt to replicate our analysis and assess composite variable bias is flawed by a mistake in their analysis, i.e., their disaggregate models do not follow from the JA aggregate specification. The purpose of this note is to correctly assess the extent to which the JA estimate of the market share effect is affected by composite variable bias and to suggest approaches for modeling a composite dependent variable in the presence of unobservable factors. In particular, we i show that the disaggregate specifications of FPA do not follow from JA, ii look at specifications not subject to composite variable bias to investigate the magnitude of the composite variable bias in JA, and iii provide a disaggregate modeling framework that controls for unobservable effects.

Abstract: Publisher Summary Despite the advances in the sophistication and proliferation of marketing-science models, relatively few true marketing strategy models have been developed, and those that have been developed generally had limited impact on management. The gap between management needs (for models that help them generate, evaluate and implement better marketing strategies) and the available models (especially non-traditional) is narrowing. As the environment and the characteristics of successful businesses are changing, the nature of the marketing paradigm, and the nature of desired marketing strategy models is changing as well.