Search

This paper shows that audience segments that are sold by data brokers for the purpose of ad targeting are often surprisingly inaccurate. Data brokers often use online browsing records to create digital consumer profiles that they sell to marketers as predefined audiences for ad targeting. However, this process is a "black box"—little is known about the reliability of the digital profiles that are created or of the audience identification provided by buying platforms. In this paper, we investigate using three field tests the accuracy of a variety of demographic and audience-interest segments. We examine the accuracy of more than 90 third-party audiences across 19 data brokers. Audience segments vary greatly in quality and are often inaccurate across leading data brokers. In comparison with random audience selection, the use of black box data profiles, on average, increased identification of a user with a desired single attribute by 0%–77%. Audience identification can be improved, on average, by 123% when combined with optimization software. However, given the high extra costs of targeting solutions and the relative inaccuracy, we find that third-party audiences are often economically unattractive except for higher-priced media placements. [ABSTRACT FROM AUTHOR]

Following a successful idea generation exercise, a company might easily be left with hundreds of ideas generated by experts, employees, or consumers. The next step is to screen these ideas and identify those with the highest potential. In this paper we propose a practical approach to involving consumers in idea screening. Although the number of ideas may potentially be very, large, it would be unreasonable to ask each consumer to evaluate more than a few ideas. This raises the challenge of efficiently selecting the ideas to be evaluated by each consumer. We describe several idea-screening algorithms that perform this selection adaptively based on the evaluations made by previous consumers. We use simulations to compare and analyze the performance of the algorithms as well as to understand their behavior. The best-performing algorithm focuses on the ideas that are the most likely to have been misclassified (as "top" or "bottom" ideas) based on the previous evaluations, and avoids discarding ideas too fast by adding random perturbations to the misclassification probabilities. We demonstrate the convergent validity of this algorithm using a field experiment, which also confirms the convergence pattern predicted by simulations. [ABSTRACT FROM AUTHOR]

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 tile 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. [ABSTRACT FROM AUTHOR]

In this paper, we provide theoretical arguments and empirical evidence for how Genetic Algorithms (GA) can be used for efficient estimation of macro-level diffusion models. Using simulations we find that GA and Sequential Search-Based-Nonlinear Least Squares (SSB-NLS) provide comparable parameter estimates when the data including peak sales are being used, for a range of error variances, and true parameter values commonly encountered in the literature. From empirical analyses we find that the forecasting performance of the GA estimates is better than that of SSB-NLS, Augmented Filter, Hierarchical Bayes, and Kalman Filter when only pre-peak sales data is available for estimation. When sales data until the peak time period are available for estimation, SSB-NLS is able to obtain parameter estimates when the starting values provided are the estimates from using GA. The estimates from GA are not biased and do not change in a systematic fashion when postpeak sales data are used,whereas the estimates from SSB-NLS are biased and change in a systematic fashion. Summarizing, we find that GA may be better suited for diffusion model estimation under the three conditions where SSB-NLS has been found to have problems. [ABSTRACT FROM AUTHOR]

We introduce the idea of a massively categorical variable, a variable such as zip code that takes on too many values to treat in the standard manner. We show how to use a massively categorical variable directly as an explanatory variable. As an application of this concept, we explore several of the issues that analysts confront when trying to develop a direct marketing campaign. We begin by pointing out that the data contained in many of the common sources are masked through aggregation in order to protect consumer privacy. This creates some difficulty when trying to construct models of individual level behavior. We show how to take full advantage of such data through a hierarchical Bayesian variance components (HBVC) model. The flexibility of our approach allows us to combine several sources of information, some of which may not be aggregated, in a coherent manner. We show that the conventional modeling practice understates the uncertainty with regard to its parameter values. We explore an array of financial considerations, including ones in which the marginal benefit is non-linear, to make robust model comparisons. To implement the decision rules that determine the optimal number of prospects to contact, we develop an algorithm based on the Monte Carlo Markov chain output from parameter estimation. We conclude the analysis by demonstrating how to determine an organization's willingness to pay for additional data. [ABSTRACT FROM AUTHOR]

Beauty contests are auction mechanisms used to buy or sell differentiated products where the auctioneer does not specify a decision rule to pick the winning bidder. Beauty contests are widely used in procuring welfare-to-work projects, freelance services, selling online ads, real estate transactions, and hiring, dating/ marriage decisions. Unlike price-only auctions, beauty contests have no closed-form bidding strategies and suffer from nonmultiplicatively separable unobserved auction heterogeneity, which makes their estimation challenging. To address these challenges, we formulate beauty contests as incomplete information games and present a two-step estimator. A key contribution of our method is its ability to account for common-knowledge auction-specific unobservables using finite unobserved types. We show that unobserved auction types and distributions of bids are nonparametrically identified and recoverable in the first step using a nonparametric Expectation-Maximization (EM)-like algorithm, and that these can then be used in the second step to recover cost distributions. We present an application of our method in the online freelancing context. We find that seller margins in this marketplace are around 15% of the bid, and that not accounting for unobserved heterogeneity can significantly bias cost estimates in this setting. Based on our estimates, we run counterfactual simulations and provide guidelines to managers of freelance firms. Data, as supplemental material, are available at http://dx.doi.org/10.1287/mksc.2015.0929. [ABSTRACT FROM AUTHOR]

This paper presents a new multidimensional scaling (MDS) methodology which operationalizes the Krumhansl (1978) distance-density model for the analysis of asymmetric proximity data. In Krumhansl's conceptualization, the symmetric Euclidean interbrand distances typically associated with the traditional MDS model are augmented by the spatial density around the brands in the derived space. This modification allows the distance-density model to accommodate many of the empirically observed violations of the metric axioms (such as asymmetry). An operationalization of the distance-density model is particularly attractive to marketers who often work with asymmetric brand switching data to investigate market structure and competition. We describe this new MDS procedure which is sufficiently flexible to fit a number of competing hypotheses of proximity. The algorithm employed for estimation is technically presented together with various program options. An analysis of an asymmetric brand switching matrix for automobiles is presented to illustrate the methodology. The results of this analysis are also compared with the solutions obtained from several of the currently available procedures for handling asymmetric proximity data. Finally, technical extensions to three-way analyses, hybrid models, etc., are discussed. [ABSTRACT FROM AUTHOR]

This paper indicates how an reformation theoretic approach via the MDI (minimum discrimination information) statistic can be used to help provide a uniform approach to both statistical testing and estimation in various kinds of marketing analyses. Extensions to constrained versions of the MDI statistic also make it possible to test the consistency of market information with management plans or policies that can be represented m "external" constraints, i.e., constraints formulated without reference to the data base, and to estimate their Impact on the market. Composite hypotheses, which are difficult to deal with by the more customary methods used m market research, can be dealt with naturally and easily via these MDI approaches. Basically MDI is more efficient than classical approaches because distribution estimation and hyopthesis testing are done simultaneously and the resulting estimates obtain regardless of the conclusion of the test. Numerical illustrations are supplied and discussed. Recent developments in mathematical programming (duality) theory and methods, which are also pertinent, are briefly examined for their bearing on still further possibilities for constrained MDI modeling. (Market Segmentation; Market Equilibrium; Switching Equilibrium; MDI Statistic; Composite Hypotheses; Mathematical Programming; External Constraints) [ABSTRACT FROM AUTHOR]

A choice model based on direct utility maximization subject to an arbitrary number of constraints is developed and applied to conjoint data. The model can accommodate both corner and interior solutions, and it provides insights into the proportion of respondents bound by each constraint. Application to volumetric choice data reveals that the majority of respondents make choices consistent with price and quantity restrictions. Estimates based on a single monetary-constraint choice model are shown to lead to biased estimates of the monetary value of attribute levels. [ABSTRACT FROM AUTHOR]

We model the diffusion of innovations in markets with two segments: influentials who are more in touch with new developments and who affect another segment of imitators whose own adoptions do not affect the influentials. This two-segment structure with asymmetric influence is consistent with several theories in sociology and diffusion research, as well as many "viral" or "network" marketing strategies. We have four main results. (1) Diffusion in a mixture of influentials and imitators can exhibit a dip or "chasm" between the early and later parts of the diffusion curve. (2) The proportion of adoptions stemming from influentials need not decrease monotonically, but may first decrease and then increase. (3) Erroneously specifying a mixed-influence model to a mixture process where influentials act independently from each other can generate systematic changes in the parameter values reported in earlier research. (4) Empirical analysis of 33 different data series indicates that the two-segment model fits better than the standard mixed-influence, the Gamma/Shifted Gompertz, and the Weibull-Garnma models, especially in cases where a two-segment structure is likely to exist. Also, the two-segment model fits about as well as the Karmeshu-Goswami mixed-influence model, in which the coefficients of innovation and imitation vary across potential adopters in a continuous fashion. [ABSTRACT FROM AUTHOR]

In this study, we develop a multivariate generalization of the negative binomial distribution (NBD). This new model has potential application to situations where separate NBDs are correlated, such as for page views across multiple websites. In turn, our page view model is used to predict the audience for internet advertising campaigns. For very large Internet advertising schedules, a simple approximation to the multivariate model is also derived. In a test of nearly 3,000 lnternet advertising schedules, the two new models are compared with some proprietary and nonproprietary models previously used for Internet advertising and are shown to be significantly more accurate. [ABSTRACT FROM AUTHOR]

This paper is concerned with the intriguing idea of "automatic" methods for determining those characteristics a product should have for greatest market share. The authors compare four interesting algorithms that approach the problem quite differently. Their undertaking was ambitious and well carried out; the computational effort was large and the results are presented clearly and are believable. The basic idea is compelling: each product is assumed to occupy a position in space, as is each respondent's "ideal point." His "liking" for a product is presumed to be related to the nearness of his ideal point to that product. But beyond this generality are important details about which little is known. Two such questions concern the estimation of ideal points and the choice of the function relating distance to preference. If an ideal point and dimensional weights are estimated from preference data. it is common to find negative weights, unless they are constrained to be nonnegative. To accept these is to refute the underlying model. Procedures for estimating ideal points and dimensional weights simultaneously appear vulnerable to colinearity. This suggcsts there may be problems in quality of estimation even in those cases where negatives are not observed.

This article focuses on a simulation comparison of methods for new product location. The value of analytic approaches to product idea generation lies in the fact that they are market based. Although a mere recitation of possible shortcomings in the market model cannot prove that analytic search will not be as productive as other means for idea generation, the authors believe that their value to managers would be improved by efforts to characterize markets more "realistically." In their paper the authors speculate that PRODSRCH and GRID SEARCH could ultimately prove the more useful of the methods compared because they are more flexible and can accommodate many of the suggestions for improving the realism of market models. Research is necessary not only to investigate which improvements to realism offer the greater promise, but also how to go about incorporating such changes. Modeling competitive action and reaction, finding ways to predict changes in market size and structure as the result of product additions, developing means for predicting the behavior of costs as a consequence of product location, predicting the attractiveness of attributes or features which do not at present characterize existing products, determining when the addition or deletion of a feature Will change the definition of a market or its structure, and the like represent major challenges for future research.