Knowledge that Transforms

Market response models based on field-generated data need to address potential endogeneity in the regressors to obtain consistent parameter estimates. Another requirement is that market response models predict well in a holdout sample. With both requirements combined, it may seem reasonable to subject an endogeneity-corrected model to a holdout prediction task, and this is quite common in the academic marketing literature. One may be inclined to expect that the consistent parameter estimates obtained via instrumental variables (IV) estimation predict better than the biased ordinary least squares (OLS) estimates. This paper shows that this expectation is incorrect. That is, if the holdout sample is similar to the estimation sample so that the regressors are endogenous in both samples, holdout sample validation favors regression estimates that are not corrected for endogeneity (i.e., OLS) over estimates that are corrected for endogeneity (i.e., IV estimation). We also discuss ways in which holdout samples may be used sensibly in the presence of endogeneity. A key takeaway is that if consistent parameter estimates are the primary model objective, the model should be validated with an exogenous (rather than endogenous) holdout sample.

Previous research on marketing budget decisions has shown that profit improvement from better allocation across products or regions is much higher than from improving the overall budget. However, despite its high managerial relevance, contributions by marketing scholars are rare. In this paper, we introduce an innovative and feasible solution to the dynamic marketing budget allocation problem for multiproduct, multicountry firms. Specifically, our decision support model allows determining near-optimal marketing budgets at the country–product–marketing–activity level in an Excel-supported environment each year. The model accounts for marketing dynamics and a product's growth potential as well as for trade-offs with respect to marketing effectiveness and profit contribution. The model has been successfully implemented at Bayer, one the world's largest pharmaceutical and chemical firms. The profit improvement potential is more than 50% and worth nearly €500 million in incremental discounted cash flows.

Marketing researchers have used models of consumer demand to forecast future sales, to describe and test theories of behavior, and to measure the response to marketing interventions. The basic framework typically starts from microfoundations of expected utility theory to obtain an econometric system that describes consumers' choices over available options, and to thus characterize product demand. The basic framework has been augmented significantly to account for quantity choices, to accommodate purchases of several products on a single purchase occasion (multiple discreteness and multicategory purchases), and to allow for asymmetric switching between brands across different price tiers. These extensions have enabled researchers to bring the analysis to bear on several related marketing phenomena of interest. This paper has three main objectives. The first objective is to articulate the main goals of demand analysis—forecasting, measurement, and testing—and to highlight several considerations associated with these goals. Our second objective is to describe the main building blocks of individual-level demand models. We discuss approaches built on direct and indirect utility specifications of demand systems, and we review extensions that have appeared in the marketing literature. The third objective is to explore a few emerging directions in demand analysis, including considering demand-side dynamics, combining purchase data with primary information, and using semiparametric and nonparametric approaches. We hope researchers new to this literature will take away a broader perspective on these models and see the potential for new directions in future research.

We study a duopoly model where consumers are heterogeneous with respect to their willingness to pay for two product characteristics and marginal costs are increasing with the quality level chosen on each attribute. We show that although firms seek to manage competition through product positioning, their differentiation strategies critically depend on how costly it is to provide higher quality. When the cost of providing quality is not too high, firms use only one attribute to differentiate their products: they maximally differentiate on one dimension and minimally differentiate on the other (a Max-Min equilibrium). Furthermore, they always differentiate along the dimension with the greater attribute range. As for the dimension with the smaller range and along which they agglomerate, firms either choose the highest quality level or the lowest quality level possible, depending on whether the marginal costs of quality provision are low or intermediate, respectively. However, for larger quality provision costs, firms exploit both dimensions to differentiate their products. In particular, we characterize a maximal differentiation equilibrium in which one firm chooses the highest quality level on both attributes while its rival offers the lowest quality level on both attributes (a Max-Max equilibrium). We discuss the managerial implications of our findings and explain how they enrich and qualify previous results reported in the literature on two-dimensional differentiation models.

We develop a two-stage consumer-level model of paid search advertising response based on standard aggregated data provided to advertisers by major search engines such as Google or Bing. The proposed model uses behavioral primitives in accord with utility maximization and allows recovering parameters of the heterogeneity distribution in consumer preferences. The model is estimated on a novel paid search data set that includes information on the ad copy. To that end, we develop an original framework to analyze composition and design attributes of paid search ads. Our results allow us to correctly evaluate the effects of specific ad properties on ad performance, taking consumer heterogeneity into account. Another benefit of our approach is allowing recovery of preference correlation across the click-through and conversion stage. Based on the estimated correlation between price- and position-sensitivity, we propose a novel contextual targeting scheme in which a coupon is offered to a consumer depending on the position in which the paid search ad was displayed. Our analysis shows that total revenues from conversion can be increased using this targeting scheme while keeping cost constant.

Our main objective in this paper is to measure the value of customers acquired from Google search advertising accounting for two factors that have been overlooked in the conventional method widely adopted in the industry: (1) the spillover effect of search advertising on customer acquisition and sales in off-line channels and (2) the lifetime value of acquired customers. By merging Web traffic and sales data from a small-sized U.S. firm, we create an individual customer-level panel that tracks all repeated purchases, both online and off-line, and tracks whether or not these purchases were referred from Google search advertising. To estimate the customer lifetime value, we apply the methodology in the customer relationship management literature by developing an integrated model of customer lifetime, transaction rate, and gross profit margin, allowing for individual heterogeneity and a full correlation of the three processes. Results show that customers acquired through Google search advertising in our data have a higher transaction rate than customers acquired from other channels. After accounting for future purchases and spillover to off-line channels, the calculated value of new customers using our approach is much higher than the value obtained using conventional method. The approach used in our study provides a practical framework for firms to evaluate the long-term profit impact of their search advertising investment in a multichannel setting.

How should forward-looking managers plan advertising if they envision a product-harm crisis in the future? To address this question, we propose a dynamic model of brand advertising in which, at each instant, a nonzero probability exists for the occurrence of a crisis event that damages the brand's baseline sales and may enhance or erode marketing effectiveness when the crisis occurs. Because managers do not know when the crisis will occur, its random time of occurrence induces a stochastic control problem, which we solve analytically in closed form. More importantly, the envisioning of a possible crisis alters managers' rate of time preference: anticipation enhances impatience. That is, forward-looking managers discount the present—even when the crisis has not occurred—more than they would in the absence of crisis. Building on this insight, we then derive the optimal feedback advertising strategies and assess the effects of crisis likelihood and damage rate. We discover the crossover interaction: the optimal precrisis advertising decreases, but the postcrisis advertising increases as the crisis likelihood (or damage rate) increases. In addition, we develop a new continuous-time estimation method to simultaneously estimate sales dynamics and feedback strategies using discrete-time data. Applying the method to market data from the Ford Explorer's rollover recall, we furnish evidence to support the proposed model. We detect compensatory effects in parametric shift: ad effectiveness increases, but carryover effect decreases (or vice versa). We also characterize the crisis occurrence distribution that shows that Ford Explorer should anticipate a crisis in 2.1 years and within 6.3 years at the 95% confidence level. Finally, we find a remarkable correspondence between the observed and optimal advertising decisions.

Many online shoppers initially acquired through paid search advertising later return to the same website directly. These so-called “direct type-in” visits can be an important indirect effect of paid search. Because visitors come to sites via different keywords and can vary in their propensity to make return visits, traffic at the keyword level is likely to be heterogeneous with respect to how much direct type-in visitation is generated. Estimating this indirect effect, especially at the keyword level, is difficult. First, standard paid search data are aggregated across consumers. Second, there are typically far more keywords than available observations. Third, data across keywords may be highly correlated. To address these issues, the authors propose a hierarchical Bayesian elastic net model that allows the textual attributes of keywords to be incorporated. The authors apply the model to a keyword-level data set from a major commercial website in the automotive industry. The results show a significant indirect effect of paid search that clearly differs across keywords. The estimated indirect effect is large enough that it could recover a substantial part of the cost of the paid search advertising. Results from textual attribute analysis suggest that branded and broader search terms are associated with higher levels of subsequent direct type-in visitation.

We study the bidding strategies of vertically differentiated firms that bid for sponsored search advertisement positions for a keyword at a search engine. We explicitly model how consumers navigate and click on sponsored links based on their knowledge and beliefs about firm qualities. Our model yields several interesting insights; a main counterintuitive result we focus on is the “position paradox.” The paradox is that a superior firm may bid lower than an inferior firm and obtain a position below it, yet it still obtains more clicks than the inferior firm. Under a pay-per-impression mechanism, the inferior firm wants to be at the top where more consumers click on its link, whereas the superior firm is better off by placing its link at a lower position because it pays a smaller advertising fee, but some consumers will still reach it in search of the higher-quality firm. Under a pay-per-click mechanism, the inferior firm has an even stronger incentive to be at the top because now it only has to pay for the consumers who do not know the firms' reputations and, therefore, can bid more aggressively. Interestingly, as the quality premium for the superior firm increases, and/or if more consumers know the identity of the superior firm, the incentive for the inferior firm to be at the top may increase. Contrary to conventional belief, we find that the search engine may have the incentive to overweight the inferior firm's bid and strategically create the position paradox to increase overall clicks by consumers. To validate our model, we analyze a data set from a popular Korean search engine firm and find that (i) a large proportion of auction outcomes in the data show the position paradox, and (ii) sharp predictions from our model are validated in the data.

Interpersonal communications have long been recognized as an influential source of information for consumers. Internet-based media have facilitated information exchange among firms and consumers, as well as observability and measurement of such exchanges. However, much of the research addressing online communication focuses on ratings collected from online forums. In this paper, we look beyond ratings to a more comprehensive view of online communications. We consider the sales effect of the volume of positive, negative, and neutral online communications captured by Web crawler technology and classified by automated sentiment analysis. Our modeling approach captures two key features of our data, dynamics and endogeneity. In terms of dynamics, we model daily measures of online communications about a firm and its products as contributing to a latent demand-generating stock variable. To account for the endogeneity, we extend the latent instrumental variable technique to account for dynamic endogenous regressors. Our results demonstrate a significant effect of positive, negative, and neutral online communications on daily sales performance. Failure to account for endogeneity results in a severe attenuation of the estimated effects. From a managerial perspective, we demonstrate the importance of accounting for communication valence as well as the impact of shocks to positive, negative, and neutral online communications.