The authors thank Digital Equipment Corporation for generously providing the data.ASIM ANSARI, SKANDER ESSEGAIER, and RAJEEV KOHLI* Several online firms, including Yahoo!, Amazon.com, and Movie Critic, recommend documents and products to consumers. Typically, the recommendations are based on content and/or collaborative filtering methods. The authors examine the merits of these methods, suggest that preference models used in marketing offer good alternatives, and describe a Bayesian preference model that allows statistical integration of five types of information useful for making recommendations: a person's expressed preferences, preferences of other consumers, expert evaluations, item characteristics, and individual characteristics. The proposed method accounts for not only preference heterogeneity across users but also unobserved product heterogeneity by introducing the interaction of unobserved product attributes with customer characteristics. The authors describe estimation by means of Markov chain Monte Carlo methods and use the model with a large data set to recommend movies either when collaborative filtering methods are viable alternatives or when no recommendations can be made by these methods.

Recently proposed methods for product-line selection use the total utilities of candidate items to construct product lines maximizing seller's return or buyers' welfare. For conjoint (hybrid conjoint) data, enumerating the utilities of candidate items can be computationally infeasible if the number of attributes and attribute levels is large and most multi-attribute alternatives are feasible. For such problems, constructing product lines directly from part-worths data is preferable. We propose such methods, extending Kohli and Krishnamurti's (1987) dynamic-programming heuristic for selecting a single item maximizing share to structure product lines maximizing share, seller's return, or buyers' (utilitarian) welfare. The computational performance of the heuristics and their approximation of product-line solutions is evaluated using simulated data. Across problem instances, the dynamic-programming heuristics identify solutions that are no worse, in terms of approximating optimal solutions, to the solutions of heuristics for the current two-step approaches to product-line design. An application using hybrid-conjoint data for a consumer-durable product is described.marketing: product policy, conjoint analysis, programming: heuristics

The authors propose two variants of lexicographic preference rules. They obtain the necessary and sufficient conditions under which a linear utility function represents a standard lexicographic rule, and each of the proposed variants, over a set of discrete attributes. They then: (i) characterize the measurement properties of the parameters in the representations; (ii) propose a nonmetric procedure for inferring each lexicographic rule from pairwise comparisons of multiattribute alternatives; (iii) describe a method for distinguishing among different lexicographic rules, and between lexicographic and linear preference models; and (iv) suggest how individual lexicographic rules can be combined to describe hierarchical market structures. The authors illustrate each of these aspects using data on personal-computer preferences. They find that two-thirds of the subjects in the sample use some kind of lexicographic rule. In contrast, only one in five subjects use a standard lexicographic rule. This suggests that lexicographic rules are more widely used by consumers than one might have thought in the absence of the lexicographic variants described in the paper. The authors report a simulation assessing the ability of the proposed inference procedure to distinguish among alternative lexicographic models, and between linear-compensatory and lexicographic models.lexicographic preferences, noncompensatory preference models, linear models, optimization techniques, greedy algorithm, approximation algorithms, utility theory, conjoint analysis, hierarchical clustering, market segmentation, hierarchical market structures

Multipart pricing is commonly used by service providers such as car rentals, prescription drug plans, health maintenance organizations, and wireless telephony. The general structure of these pricing schemes is a fixed access fee, which sometimes entitles users to a certain level of product use; a variable fee for additional use; and still another fee for add-on features that are priced individually and/or as bundles. The authors propose a method using conjoint analysis for multipart pricing. The method reflects the two-way dependence between prices and consumption and incorporates consumers' uncertainty about their use of a service. The proposed method estimates both choice probabilities and usage levels for each consumer as functions of the product features and the different price components. The authors then use these estimates to evaluate the expected revenues and profits of alternative plans and pricing schemes. They illustrate this method using data from a conjoint study of cell phone services. They compare the results with those obtained from using several competing models. Finally, they use the proposed procedure to identify the optimal set of features in a base plan and the pricing of optional features for a provider of cell phone services.

The authors propose two generalizations of conjunctive and disjunctive screening rules. First, they relax the requirement that an acceptable alternative must be satisfactory on one criterion (disjunctive) or on all criteria (conjunctive). Second, they relax the assumption that consumers make deterministic judgments when evaluating alternatives. They combine the two generalizations into a probabilistic subset-conjunctive rule, which allows consumers to use any number or subset of decision criteria when screening alternatives and permits them to be uncertain about the acceptability of attribute levels. These two features allow for a screening process that is uncertain and more flexible than the deterministic conjunctive and disjunctive rules currently described in the literature. The authors describe a latent-class method for the estimation of the subset-conjunctive rules and the attribute-level consideration probabilities using either consideration or choice data. Applications using both types of data suggest that the proposed models predict as well as linear models do; can make different predictions of consideration, choice, and market shares; and provide insights into consumer decision processes that are different from those obtained with linear models.

This article examines the time between product development and market launch, and its relation to the subsequent diffusion of consumer durables. We find that this “incubation time” is long. Further, it is a useful predictor of the shape of the subsequent sales diffusion curve. Using the Bass model as a base, we find that the longer the incubation time, the lower the coefficient of innovation (p) and the longer the time to peak sales. Further, using the incubation time in a Bayesian forecasting model significantly improves forecasts early in the life cycle. © 1999 Elsevier Science Inc.

A significant application of conjoint analysis is in pricing decisions for new products. Conceptually, profit maximization is an important criterion for selecting the price of a product. However, the maximization of profit necessitates estimation of fixed and variable costs, which are difficult to estimate reliably for the large number of products available for evaluation in conjoint analysis. Consequently, users of conjoint models have begun to use a share simulation to screen a small set of attractive products. For each screened product, fixed and variable costs are estimated separately and used to simulate its profits at different price levels. The limitation of this approach is that the profit simulations are based on the assumption that the conjoint data, and hence the predicted profits, are error free. Also, though the purpose of examining alternative prices is to determine the best price at which to offer a new product, current conjoint simulators do not focus explicitly on optimal pricing decisions. The authors describe and illustrate a model for optimal pricing of screened products in conjoint analysis, incorporating the effect of measurement and estimation error on predicted profits.

A dynamic-programming heuristic is described to find approximate solutions to the problem of identifying a new, multi-attribute product profile associated with the highest share-of-choices in a competitive market. The input data consist of idiosyncratic multi-attribute preference functions estimated using conjoint or hybrid-conjoint analysis. An individual is assumed to choose a new product profile if he/she associates a higher utility with it than with a status-quo alternative. Importance weights are assigned to individuals to account for differences in their purchase and/or usage rates and the performance of a new product profile is evaluated after taking into account its cannibalization of a seller's existing brands. In a simulation with real-sized problems, the proposed heuristic strictly dominates an alternative lagrangian-relaxation heuristic in terms of both computational time and approximation of the optimal solution. Across 192 simulated problems, the dynamic-programming heuristic identifies product profiles whose share-of-choices, on average, are 98.2% of the share-of-choices of the optimal product profile, suggesting that it closely approximates the optimal solution.marketing, product design, conjoint analysis, heuristics