Knowledge that Transforms

In a viral marketing campaign, an organization develops a marketing message and encourages customers to forward this message to their contacts. Despite its increasing popularity, there are no models yet that help marketers to predict how many customers a viral marketing campaign will reach and how marketers can influence this process through marketing activities. This paper develops such a model using the theory of branching processes. The proposed viral branching model allows customers to participate in a viral marketing campaign by (1) opening a seeding e-mail from the organization, (2) opening a viral e-mail from a friend, and (3) responding to other marketing activities such as banners and offline advertising. The model parameters are estimated using individual-level data that become available in large quantities in the early stages of viral marketing campaigns. The viral branching model is applied to an actual viral marketing campaign in which over 200,000 customers participated during a six-week period. The results show that the model quickly predicts the actual reach of the campaign. In addition, the model proves to be a valuable tool to evaluate alternative what-if scenarios.

Many consumer durable retailers often do not advertise their prices and instead ask consumers to call them for prices. It is easy to see that this practice increases the consumers' cost of learning the prices of products they are considering, yet firms commonly use such practices. Not advertising prices may reduce the firm's advertising costs, but the strategic effects of doing so are not clear. Our objective is to examine the strategic effects of this practice. In particular, how does making price discovery more difficult for consumers affect competing retailers' price, service decisions, and profits? We develop a model in which a manufacturer sells its product through a high-service retailer and a low-service retailer. Consumers can purchase the retail service at the high-end retailer and purchase the product at the competing low-end retailer. Therefore, the high-end retailer faces a free-riding problem. A retailer first chooses its optimal service levels. Then, it chooses its optimal price levels. Finally, a retailer decides whether to advertise its prices. The model results in four structures: (1) both retailers advertise prices, (2) only the low-service retailer advertises price, (3) only the high-service retailer advertises price, and (4) neither retailer advertises price. We find that when a retailer does not advertise its price and makes price discovery more difficult for consumers, the competition between the retailers is less intense. However, the retailer is forced to charge a lower price. In addition, if the competing retailer does advertise its prices, then the competing retailer enjoys higher profit margins. We identify conditions under which each of the above four structures is an equilibrium and show that a low-service retailer not advertising its price is a more likely outcome than a high-service retailer doing so. We then solve the manufacturer's problem and find that there are several instances when a retailer's advertising decisions are different from what the manufacturer would want. We describe the nature of this channel coordination problem and identify some solutions.

This article explores some aspects of wine journalism in Norwegian newspapers. Two issues are discussed: First, are wine sales influenced by wine journalists' reviews? Second, do experts agree on what makes a good wine buy? The results show that wine sales are indeed significantly influenced by the judgements of wine critics; a 10% rise in newspapers' scores in Norway was accompanied by an average increase of 16%–18% in sales figures for table wines. The effect of wine reviews varied somewhat from newspaper to newspaper. It proved difficult to establish criteria for a good wine buy that are objective and independent of the person making the judgement. The journalists gave no unanimous recommendation of good wine buys to the consumers; the same wine could get good reviews in some papers and might well receive run-of-the-mill reviews in others. However, a majority of the reviewers seemed to agree in the ranking of most of the wines, even if the absolute value of the scores differed.

We study how market structure within a product category varies across retail formats. Building on the literature on internal market structure, we estimate a joint store and brand choice model where the loading matrix of brand attributes are allowed to be retail format specific. The approach allows us to recover brand maps for different retail formats while controlling for the short-term marketing mix activities at these stores and the self-selection of households that frequent a particular format. The model is applied to consumer panel data from two product categories, where households are observed to make purchases across three store types: high-end grocery store, traditional supermarket, and large everyday low pricing (EDLP) formats. Our results show strong correlations between the marketing mix sensitivities, store format preference, and unobserved brand attributes. These correlations translate into significant differences in market structure across retail formats and in the direction and size of preference vectors for unobservable brand attributes. We find a tight clustering of brands at the EDLP format, whereas brands are found to compete in distinct subgroups at other stores. Results show that failure to account for retail format effects can substantially bias the understanding of underlying market structure and could lead to incorrect implications in applications such as new product entry.

In this paper, we study the practice of overselling in a competitive environment where late-arriving consumers value the good higher than early-arriving ones but the former's arrival is uncertain. We show that overselling is a dominant strategy for the firms. However, it can lead to a prisoners' dilemma situation in which all firms are worse off overselling. We further show that only when demand from the late consumers far exceeds the supply and there is a sufficiently high profit margin from reselling does overselling result in a Pareto-dominant outcome for the firms.

Procter & Gamble (P&G) Asia-Pacific is interested in managing value growth. Only after fully understanding the true effects of the marketing-mix variables can P&G managers make strategic decisions answering questions such as the following: (1) Are the P&G brands in the detergent market inelastic or elastic with respect to price? How has the price elasticity changed over time? Can P&G increase the price of its brands to gain value growth? (2) What are the price, distribution, and sizing combinations needed to achieve the desirable value growth? (3) How can P&G gain market share from its competitors without cannibalizing its own brands? P&G Asia-Pacific approached us to develop a value growth framework to answer these questions. To generate the answers for the above questions, we develop a three-step weighted random coefficient estimator that captures the heterogeneity across cross sections (different stock-keeping units and states) and the endogeneity of distribution. Based on the parameter estimates, we provide strategic recommendations to P&G for a field test to validate our suggestions. We developed a simulator for P&G managers so that they can generate appropriate marketing-mix strategies for achieving the desired value growth. As a result, P&G gained over $39 million in value growth over a one-year period by implementing the recommendations from our modeling approach.

Remember James Boswell, ninth Laird of Auchinleck, author of the famous maxim that the road to hell is paved with good intentions? Trying to build realistic theories differs dramatically from having correct explanatory theories tested on objective criteria, e.g., verifiable prediction. Evaluating theories on whether assumptions are realistic is potentially subjective, biased, and arbitrary. A theory's domain determines whether its assumptions are sufficiently realistic and when assumptions must hold and to what degree, so testing assumptions in isolation puts an unnecessary burden on the assumptions (i.e., they must hold everywhere). For theories explaining cooperation and information exchange, predictions reveal that the prisoner's dilemma assumptions (only two prisoners, four possible outcomes, two possible actions, etc.) are sufficiently realistic. For theories explaining prisoner sentencing guidelines and probation policy, predictions might suggest otherwise. Scientific methods allow the evaluation of theories on criteria such predictive accuracy, reliability, validity, and robustness—not based on realism. When multiple explanatory theories survive initial testing, one derives conflicting predictions. For example, a theory that people are broccoli produces correct predictions (people are mortal) and incorrect predictions (people are biennial). Tragic consequences can occur when theory adoption depends on whether assumptions are disliked, unpopular, or exclude a favorite variable. Denounce journals that reject models with insightful new implications because the assumptions are too simple or merely disliked. The term “unrealistic” sometimes means personally disliked.

In his July–August 2007 editorial of Marketing Science, Steven Shugan argues that the realism of assumptions does not matter as long as a theory or model produces satisfactory predictions and claims further that unrealistic assumptions breed good theories. This commentary discusses the problems of his argument and presents a very different view about the realism of assumptions. Assumptions need not be realistic if the only goal of science is prediction. However, a major function of theory is also to explain and not just to predict. The role of explanation is more important in the social sciences because it is far more difficult to produce accurate predictions in the social than the natural sciences. Assumptions, especially core assumptions, often constitute the foundation of the mechanismic explanations provided by a theory. Unrealistic assumptions may lead to faulty explanations and false predictions. Contrary to Shugan's view, the realism of an assumption cannot be assessed just based on the output of a theory. It has to be tested independently of or in conjunction with the hypotheses of the theory. Also, contrary to Shugan's claim, more realistic assumptions result in better theories. As theory development advances, efforts should be directed toward making assumptions more realistic.

Panel survey data have been gaining importance in marketing. However, one challenge of estimating econometric models based on panel survey data is how to account for underreporting; that is, respondents do not report behavioral incidences that actually occur. Underreporting is especially likely to occur in a panel survey because the data-recording mechanism is often tedious, complex, and effortful. The probability of underreporting is likely to vary across respondents and also over the duration of the survey period. In this paper, we propose a model to simultaneously study reported behavioral incidences and partially observed actual behavioral incidences. We propose a Bayesian approach for estimating the proposed model. We treat those unobserved actual behavioral incidences as latent variables, and the Gibbs sampler makes it convenient to impute the nonreported consumption incidences along with making inferences on other model parameters. Our proposed method has two advantages. First, it offers a model-based approach to remove the underreporting bias in panel survey data and therefore allows marketing researchers to make accurate inferences about consumers' actual behavior. Second, the method also offers a natural way to study factors that influence respondents' propensity to underreport. Because we treat those underreported behavioral incidences as nonmissing at random, this underreporting propensity varies across respondents and over time. This understanding can help marketing researchers design the right strategy to intervene and incentivize respondents to authentically report and hence improve the quality of survey data. The proposed model and estimation approach are tested on both synthetic data and actual panel survey data on consumer-reported beverage-drinking behavior. Our analysis suggests that underreporting can significantly mask respondents' true behavior.