Knowledge that Transforms

Each year in the postsecondary education industry, schools offer admission to nearly 3 million new students and scholarships totaling nearly $100 billion. This is a large, understudied targeted marketing and price discrimination problem. This problem falls into a broader class of configuration utility problems (CUPs), which typically require an approach tailored to exploit the particular setting. This paper provides such an approach for the admission and scholarship decisions problem. The approach accounts for the key distinguishing feature of this industry—schools value the average features of the matriculating students such as percent female, percent from different regions of the world, average test scores, and average grade point average. Thus, as in any CUP, the value of one object (i.e., student) cannot be separated from the composition of all of the objects (other students in the enrolling class). This goal of achieving a class with a desirable set of average characteristics greatly complicates the optimization problem and does not allow the application of standard approaches. We develop a new approach that solves this more complex optimization problem using an empirical system to estimate each student's choice and the focal school's utility function. We test the approach in a field study of an MBA scholarship process and implement adjusted scholarship decisions. Using a holdout sample, we provide evidence that the methodology can lead to improvements over current management decisions. Finally, by comparing our solution to what management would do on its own, we provide insight into how to improve management decisions in this setting.

This paper analyses firms' decisions to provide connectivity to their customers. We distinguish between intraconnectivity—the ability of one firm's customers to connect to each other—and interconnectivity—the ability of one firm's customers to connect with another firm's customers. The profitability implications of allowing connectivity are not a straightforward consequence of the consumer value of connectivity, because connectivity affects not only the customer value but also the intensity of competition by creating or changing network externality. We find that if sales are driven by brand switching rather than by category expansion, a firm may find it optimal not to provide intraconnectivity even if providing it is not costly and may find it optimal to provide interconnectivity even at a cost exceeding the consumer value of connectivity. On the other hand, if category expansion is possible, providing intraconnectivity may be profitable. In this case, either the equilibrium intraconnectivity provision may be asymmetric or both firms may find it (individually) optimal to provide intraconnectivity. Under certain conditions in the latter case, the firms' choice of intraconnectivity is a prisoner's dilemma game.

We reanalyze endogenous sample selection in the context of customer scoring, targeting, and influencing decisions. Scoring relies on ordered lists of probabilities that customers act in a way that contributes revenues, e.g., purchase something from the firm. Targeting identifies constrained sets of covariate patterns associated with high probabilities of these acts. Influencing aims at changing the probabilities that individual customers act accordingly through marketing activities. We show that successful targeting and influencing decisions require inference that controls for endogenous selection, whereas scoring can proceed relatively successfully based on simpler models that provide (local) approximations, capitalizing on spurious effects of observed covariates. To facilitate the type of inference required for targeting and influencing, we develop a prior that frees the analyst from having to specify (often arbitrary) exclusion restrictions for model identification a priori or to explicitly compare all possible models. We cover exclusions of observed as well as unobserved covariates that may cause the successive selections to be dependent. We automatically infer the dependence structure among selection stages using Markov chain Monte Carlo-based variable selection, before identifying the scale of latent variables. The adaptive parsimony achieved through our prior is particularly helpful in applications where the number of successive selections exceeds two, a relevant but underresearched situation.

A widely debated question in recent years by both strategy theorists and antitrust practitioners is what role product differentiation between firms plays in their ability to sustain a collusive agreement in order to reduce the strength of competition and gain higher profits. This paper addresses the following question: What happens to the “product differentiation–collusion sustainability” relationship when setting up and maintaining an agreement is costly? We show that introducing collusion costs into the discussion has relevant implications. Indeed, sufficiently high collusion costs modify the underlying market structure, thus altering the product differentiation–collusion sustainability relationship with respect to the case where collusion costs are absent or low. In particular, if the gains from collusion are increasing (decreasing) with the degree of product differentiation, the relationship between product differentiation and collusion sustainability is always positive (negative), whereas if the gains from collusion are inverted U-shaped, the relationship is inverted U-shaped too. These results stress the importance of considering those markets where the coordination between firms is sufficiently costly as structurally different from those markets where coordination has no costs for firms.

The purpose of this paper is to describe the implementation of a category management tool known as Category Optimizer™ at Foster's Wine Estates Americas for one of its brands, the Beringer California Collection. Foster's was facing a common management problem: harnessing its portfolio of Beringer California Collection wines to increase profitability, improve its competitive position, and defend against a disruptive new entrant in the U.S. wine market called Yellow Tail. Category Optimizer combines the parsimony of an internal market structure with the advances that have been made in assortment planning in operations research, assortment and stock-keeping-unit–level modeling, mixed logits, and the marketing literature on the perceptions of variety of assortment to develop and estimate a model on readily available store scanner data. The model subsequently uses these results to inform strategic and tactical decision making. This approach led to recommendations that initially seemed counterintuitive; the normal response would be for Foster's to consider lowering prices to maintain share and volume, a strategy not inconsistent with many of the recommendations of past models. However, considering the additional degrees of freedom that a product range offered for defense, we demonstrated that a combination of price increases together with the introduction of a volume-flanker product in a new channel would improve profits, increase revenue, and protect and enhance market share. These were successfully implemented in early 2008, earning rich dividends for the company; increasing profitability by 70%, revenue by 3%, and earnings before interest and taxes by 8.5%; and having a positive impact on its brand ranking. In fact, in 2008, it debuted as sixth among the international wine brands. It also managed to play an important role in deposing Yellow Tail, the market share leader, from its dominant position. We conclude the paper by providing examples of other companies where this approach has also been successfully implemented and by discussing some avenues for future research.

This paper describes two new data sets available to academic researchers (at http://www.informs.org/Community/ISMS ). The first is a panel data set containing the transactions of 19,936 households made over the period from December 1998 to November 2004 at a major U.S. consumer electronics retailer. There are a total of 173,262 transactions, including purchases and returns of products as well as extended warranties. There are 16 product categories and 292 subcategories, ranging from big-ticket items such as televisions to small-ticket items such as CDs and batteries. The second data set features a field experiment for a Christmas promotion that took place in December 2003 in the form of a direct mailing sent to a randomly selected group of households at the end of November 2003. We describe the data and the potential research issues that can be studied using these two durable goods data sets.

One of the characteristics of the fashion marketplace is the unpredictability and apparent randomness of fashion hits. Another one is the information asymmetry among consumers. In this paper, we consider fashion as a means consumers use to signal belonging to a higher social rank and propose an analytical model of fashion hits in the presence of competition and consumers who can coordinate on which product to use. We show that, consistent with the observed market phenomenon, in equilibrium, consumer coordination involves randomization between products chosen, i.e., in randomness of fashion hits. Analyzing optimal consumer choice, we find that whenever low-type consumer demand for a product is positive, a price increase results in a higher probability of high-type consumers choosing this product but lower low-type consumer demand. We also show that although high-type consumers may prefer (higher) prices that would lead to complete separation of the high- and the low-type consumers through product use, in equilibrium, firms always price as to attract positive demand from low-type consumers. The equilibrium price and profits turn out to be nonmonotonic in the low-type consumer valuation of being recognized as belonging to a higher social rank. Equilibrium profits first increase and then decrease in this valuation.

This research examines how individual differences and institutional practices influence consumer bidding in auctions. Bidders may be motivated by different goals, e.g., thrill (of winning the item, with minimal attention to what they pay for it) versus prudence (winning the item at a price at or below its perceived value). Also, innate or auctioneer-induced differences may exist in the precision and salience of bidder cognitions about the item's value. We report two studies on how these motivational and cognitive factors influence bids in descending and ascending auctions, respectively. Each study also manipulated a situational variable (wait time at each price step). The two auctions realized different average prices for the same item set. Average bids were higher in the descending (versus ascending) auction in several study conditions. In both auction formats, bidders primed with thrill (versus prudence) bid higher, but more precise and/or salient values attenuated this goal effect. Among other results, in the descending auction, longer wait times elicited higher bids from bidders primed with thrill (but not prudence). In the ascending auction, longer wait times produced lower bids for bidders primed with prudence (but not thrill). These findings on consumer bidding behavior have practical implications for auction design.

We investigate a business-to-business context and ask when and why a firm should announce a “reference program” that commits the firm to facilitating the flow of information about the efficacy of its products from early adopters to potential late adopters. We model a monopolist manufacturer with a new innovation that can be sold to two potential customers. We demonstrate here two benefits of a reference program that relate not to an increase in later adopters' willingness to pay but to an increase in the willingness to pay of the early adopters themselves. The impact on the early adopters' willingness to pay arises in two ways as a result of their observation of the firm's commitment to information transmission. First, in a model of symmetric uncertainty, we show that the announcement of a reference program facilitates dynamic pricing by the manufacturer in the sense that it allows the firm to provide temporary exclusive use of the technology to one of the customers. This creates more value, which the manufacturer can extract via a higher price. In this way, a reference program can serve as a partial substitute for an exclusive-use contract. In a model with asymmetric information, we demonstrate that under certain conditions, the firm is able to use the reference program as a signal—again, to the early adopting customer—that its technology is of high quality. However, such a signal requires significant discounts to early adopters to ensure separation. As a result, a pooling equilibrium dominates in which the manufacturer fosters references regardless of its quality. Finally, by allowing the firms' private information to be stochastic, we show that separation may be a dominant outcome.