Knowledge that Transforms

We introduce and test a behavioral model of consumer product search that extends a baseline normative model of sequential search by incorporating nonnormative influences that are local in the sense that they reflect consumers' undue sensitivity to recently encountered alternatives. We propose two types of such local behavioral influences that, at each stage of a search process, can manifest themselves both in which of the products inspected up to that point is deemed to be the most preferred one (the product comparison decision) and whether to terminate the search at that stage (the stopping decision). The first of these influences is that consumers respond excessively to the attractiveness of the currently inspected product, at the expense of all others (“focalism”). The second proposed behavioral influence is that consumers overreact to the difference in attractiveness between the current product and the one encountered just prior to it (“local contrast”). Converging evidence from two experiments, which combine to guarantee both high internal and high external validity, provides support for the proposed behavioral influences. Our findings demonstrate that consumers' product comparison and stopping decisions in sequential product search are jointly governed by normative principles and by the proposed local behavioral influences.

Critics often decry an earnings-focused short-term orientation of management that eschews spending on risky, long-term projects such as innovation to boost a firm's stock price. Such critics assume that stock markets react positively to announcements of immediate earnings but negatively to announcements of investments in innovation that have an uncertain long-term pay off. Contrary to this position, we argue that the market's true appreciation of innovation can be estimated by assessing the total market returns to the entire innovation project. We demonstrate this approach via the Fama-French 3-factor model (including Carhart's momentum factor) on 5,481 announcements from 69 firms in five markets and 19 technologies between 1977 and 2006. The total market returns to an innovation project are $643 million, more than 13 times the $49 million from an average innovation event. Returns to negative events are higher in absolute value than those to positive events. Returns to initiation occur 4.7 years ahead of launch. Returns to development activities are the highest and those to commercialization the lowest of all activities. Returns to new product launch are the lowest among all eight events tracked. Returns are higher for smaller firms than larger firms. Returns to the announcing firm are substantially greater than those to competitors across all stages. We discuss the implications of these results.

Virtual advisors often increase sales for those customers who find such online advice to be convenient and helpful. However, other customers take a more active role in their purchase decisions and prefer more detailed data. In general, we expect that websites are more preferred and increase sales if their characteristics (e.g., more detailed data) match customers' cognitive styles (e.g., more analytic). “Morphing” involves automatically matching the basic “look and feel” of a website, not just the content, to cognitive styles. We infer cognitive styles from clickstream data with Bayesian updating. We then balance exploration (learning how morphing affects purchase probabilities) with exploitation (maximizing short-term sales) by solving a dynamic program (partially observable Markov decision process). The solution is made feasible in real time with expected Gittins indices. We apply the Bayesian updating and dynamic programming to an experimental BT Group (formerly British Telecom) website using data from 835 priming respondents. If we had perfect information on cognitive styles, the optimal “morph” assignments would increase purchase intentions by 21%. When cognitive styles are partially observable, dynamic programming does almost as well—purchase intentions can increase by almost 20%. If implemented system-wide, such increases represent approximately $80 million in additional revenue.

Critics have long faulted the wide-spread practice of trade promotions as wasteful. It has been estimated that this practice adds up to $100 billion worth of inventory to the distribution system. Yet, the practice continues. In this paper, we propose a price discrimination model of trade promotions. We show that in a distribution channel characterized by a dominant retailer, a manufacturer has incentives to price discriminate between the dominant retailer and smaller independents. While offering all retailers the same pricing policy, price discrimination can be implemented through trade promotions because they induce different inventory-ordering behaviors on the part of retailers. Differences in inventory holding costs have been shown to be an important determinant of consumer promotions. Our analysis suggests that differences in holding costs are also potentially an important driver for the use of trade promotions. The implications from our model explain a number of anecdotal and/or empirically observed puzzles about how trade promotions are practiced. For example, our analysis explains why chain stores welcome trade promotions but independents do not. Our analysis outlines implications for managing trade promotions.

In this paper, we propose and operationalize a new method for optimizing shelf arrangements. We show that there are important dependencies between the layout of the shelf and stock-keeping unit (SKU) sales and marketing effectiveness. The importance of these dependencies is further shown by the substantive profit gains we obtain with our proposed shelf optimization approach. The basis of our model is a standard sales equation that explains sales using item-specific marketing effect parameters and intercepts. In a Hierarchical Bayes (HB) fashion, we augment this model with a second layer that relates the effect parameters to shelf and SKU descriptors. We also take into account potential endogeneity of facings. After estimating the parameters of the two-level model using Bayesian methodology, we carefully investigate the dependencies of SKU sales and SKU marketing effectiveness on the shelf layout. Next, we search for the shelf arrangement that maximizes the expected total profit using simulated annealing (SA). We appear to be able to increase profits for all the stores analyzed, and our approach appears to outperform well-known rules of thumb.

We investigate the sensitivity of cross-brand pass-through estimates to two types of pooling: across stores, and across regular price and promotional price weeks. Using the category data from Besanko, Dubé, and Gupta (2005), hereafter BDG, we find consistent support across all 11 categories for the predictive power of the wholesale prices of substitute products for retail shelf prices. A Bayesian procedure is used to address the small sample issues that arise in the absence of pooling. Even though the unpooled results render our inferences for specific cross-brand pass-through magnitudes reported in BDG as imprecise, consistent with McAlister (2007), we do find significant empirical support for cross-brand pass-through. We next assess the sensitivity of cross-brand pass-through estimates to pooling. This requires us to construct a much longer time series of 224 weeks for the refrigerated orange juice category, in contrast with the 52-week samples used in BDG and McAlister (2007). We find strong empirical support for the predictive power of wholesale prices of substitute products for retail shelf prices. In addition, we find evidence of nonzero own- and cross-brand pass-through elasticities for which our inferences are much more precise. These findings are robust to the separation of regular and promotional price weeks. However, the magnitudes of own-brand and cross-brand pass-through are quite different during promotional and regular price weeks. Our results clearly show that with longer data series and more robust models that can handle small sample sizes, there is evidence of cross-brand pass-through, substantiating the findings in BDG. Finally, we comment on why our results are entirely consistent with both the theoretical and empirical literatures on category pricing and retailer behavior.

This paper seeks to quantify the long-term financial impact of negative word of mouth (NWOM), an issue that has long challenged extant research. We do so with real-world data on firm security prices. The developed time-series models innovatively uncover (1) short- and long-term effects of NWOM on cash flows, stock returns, and stock volatilities, and (2) NWOM's “wear-in” effects (i.e., it takes a number of months before the stock price impact of NWOM reaches the peak point) and “wear-out” effects (i.e., it takes several months after the peak before the stock price impact of NWOM dies out completely). In addition, the results related to endogeneity and feedback effects from the stock market are also interesting, supporting the idea that historical underperformance in stock prices may breed more harmful future buzz in a “vicious” cycle of NWOM. After controlling for competition, NWOM's long-term financial harm becomes more destructive in magnitude, kicks in more quickly, and haunts investors longer. Overall, these findings offer some unique implications for buzz management, time-series models quantifying the financial impact of word of mouth, and the marketing-finance interface.