Knowledge that Transforms

Efforts on developing, implementing, and evaluating a marketing analytics framework at a real-world company are described. The framework uses individual-level transaction data to fit empirical models of consumer response to marketing efforts and uses these estimates to optimize segmentation and targeting. The models feature themes emphasized in the academic marketing science literature, including incorporation of consumer heterogeneity and state dependence into choice, and controls for the endogeneity of the firm’s historical targeting rule in estimation. To control for the endogeneity, we present an approach that involves conducting estimation separately across fixed partitions of the score variable that targeting is based on, which may be useful in other behavioral targeting settings. The models are customized to facilitate casino operations and are implemented at the MGM Resorts International’s group of companies. The framework is evaluated using a randomized trial implemented at MGM involving about 1.5 million consumers. Using the new model produces about $1 million to $5 million in incremental profits per campaign, translating to about 20¢ in incremental profit per dollar spent relative to the status quo. At current levels of marketing spending, this implies between $10 million and $15 million in incremental annual profit for the firm. The case study underscores the value of using empirically relevant marketing analytics solutions for improving outcomes for firms in real-world settings. Data are available at https://doi.org/10.1287/mksc.2017.1039 .

For many years, marketing managers have used dynamic sales response models to compute expected sales conditional on the available information. These models fail to recognize that the volatility (conditional variance) of sales can vary over time. Moreover, the covolatilities (conditional covariances) between sales and marketing-mix variables can be time varying. Both concepts introduce a new range of strategic and tactical considerations for product and brand managers. Using a multivariate volatility model, we investigate the covolatility of sales and the marketing mix of a focal brand and competing brands in the market. We also examine carryover effects from a volatility perspective. The methodology is applied to six product categories sold by Dominick’s Finer Foods. The results reveal valuable implications for marketing managers. Data and the online appendix are available at https://doi.org/10.1287/mksc.2016.1013 .

In this paper, I introduce a framework of price promotions by firms that preschedule their sale dates. I set up a dynamic model of demand accumulation in which high- and low-valuation consumers enter the market every period. The high-valuation consumers buy immediately and leave the market; the low-valuation consumers accumulate while waiting for the sale price. The firms schedule the dates of their promotions in advance. I find that often they use mixed strategies, choosing the future promotion period according to a probability distribution function. If the firms can cancel their prescheduled sales, typically they can wait longer until holding sales by shifting the probability distribution towards later periods. Scheduling promotions in advance increases the firms’ profits in comparison to the case when the promotion decision is made in the period when the promotion is offered. Data and the online appendix are available at https://doi.org/10.1287/mksc.2017.1052 .

Existing research on the compromise effect has focused exclusively on the individual. This paper investigates compromise effects in a setting that involves multiple individuals making a choice. We study whether the dyadic compromise effect (DCE) exists, the association between dyadic and individual compromise effects, and strategies to mitigate the DCE. We build a statistical model of dyadic choice that formally incorporates DCE. We conduct two studies to test our proposed models empirically. In Study 1, we begin with an investigation of the DCE with student subjects. In Study 2, we test for the presence of DCE among married couples when making retirement investment choices. In both studies, model-free and model-based evidence provides strong support for the presence of DCE. A model that incorporates DCE provides a better fit than models that do not. Evidence in support of DCE is shown to be robust to alternative compromise effect model specifications and utility aggregation methods. We find that the individual compromise effect tendency of a group member with a greater stake in the decision is likely to persist as a DCE in the joint choice setting. Our findings suggest that education of segments vulnerable to compromise effects reduces the DCE. Data and the web appendix are available at https://doi.org/10.1287/mksc.2016.1019 .

Most new-product frameworks in marketing and economics, as well as lay beliefs and practices, hold that the larger the stock of adoption of a new product, the greater the likelihood of additional adoption. Less is known about the underlying mechanisms as well as the conditions under which this central assumption holds. We use a series of field and consequential choice experiments to demonstrate the existence of nonpositive and even negative effects of large adoption stock information on the likelihood of subsequent adoption. The results highlight the degree of homophily with the adopting stock as well as the level of customer uncertainty as key characteristics determining the nature of the effect of stock information. In particular, information about a large existing adoption stock generates a positive effect on adoption only under moderate customer uncertainty combined with sufficient homophily; in other levels of uncertainty and/or homophily we find effects ranging from null to negative. This is the first direct test and demonstration of the intricate role of information about a large stock of adoption in the new product diffusion process, and it carries direct implications for marketers. Data and the online appendix are available at https://doi.org/10.1287/mksc.2016.1011