هوش وفاداری و تبعیض قیمت در یک انحصار دوگانه فروش

Business intelligence tools have enabled novel and relatively low-cost capabilities to collect and analyze vast amount of customer information. Accumulation of customer specific information along with transactional data empowers firms to categorize customers into segments and offer customized prices. We study the impact of price discrimination and market segmentation on competition and consumer purchase behavior in a game-theoretic model with two asymmetric firms. At equilibrium, both firms price discriminate and segment the market. Contrary to previous price discrimination and market segmentation findings, the game is not necessarily a prisoner’s dilemma. The firm dominating the industry is likely to improve its profits at the expense of the rival firm, and consumer welfare will increase with segmentation. We define two fundamental parameters, market dominance and the technology cost to industry dominance ratio, to drive segmentation technology adoption decisions, as a basis for our analytical approach.

Market segmentation is the practice of splitting customers into different groups within which customers have similar characteristics, needs and wants. Interest in market segmentation has been growing lately in both the real and the virtual economy, spurred by the availability of sophisticated business intelligence and data mining tools, which provide relatively low-cost capabilities to collect and analyze vast amount of customer information.1 Firms can also infer customers’ price-sensitivity from large data warehouses of customer information such as age, income, and purchasing history. For example, Principal Financial Group (PFG)2 is a firm that manages retirement savings, investments and insurance for more than fifteen million employees. By collecting high-quality data about demographics, life milestones, and benefits-enrollment habits of its customers, PFG delivers personalized investment advice and customer service to a fragmented customer base and sells retirement plans, supplementary mutual funds and insurance. Market segmentation today can be carried out to an extent never envisioned before. The granularity of segments is only limited by the amount of data available, the sophistication of the technological tools, and cost. In their seminal paper, Dickson and Ginter (1987) state that segments form homogeneous sub-markets, where customers within the same segment respond similarly to firms’ product offerings. Wedel and Kamakura (2000) extensively reviewed the marketing segmentation literature, compiling a list of methods and identifying the variables used to assign customers to segments. Frank et al. (1972) categorized segmentation research into two main streams: microeconomic theory and behavioral sciences. Our focus in this paper will be on the economical implications of market segmentation rather than on the segmentation methods and variables. In general, when firms decide to charge different prices to different segments, three important questions arise: What should be the price for a particular segment? What will be the impact of segmentation and price discrimination on market share? How would consumer welfare change with segmentation and price discrimination? In this study, we address these issues and the impact of segmentation technology cost in a duopolistic price discrimination model, and specifically investigate the impact of segment granularity on price, market share and social welfare. A duopoly is a market situation in which two firms are dominant. Unlike monopolistic price discrimination and segmentation models (Board, 2008, Jing, 2007 and Saak, 2008) which have been extensively studied, duopolies, especially in the case of asymmetric firms, have received less attention, as will become clear in the literature review section. Yet duopolies occur in many sectors of the economy. Examples include Pepsi and Coca-Cola in the soft drink industry, Intel and AMD in microprocessors, FedEx and UPS in package delivery, and Airbus and Boeing in aircraft manufacturing. Our model will consist of two asymmetric firms competing over a population of customers with distinct levels of loyalty. We will use a game-theoretic model where firms may or may not segment the market and offer different prices to each segment when they do. We assume that firms make their pricing decisions simultaneously. Our work is an extension of Shaffer and Zhang’s (2000) model to address the impact of multiple segments on price, market share and social welfare. We show that firms have a strong incentive to segment the market, as price discrimination coupled with market segmentation does not necessarily lead to prisoner’s dilemma when the dominance of a firm is high, where dominance is a decision-making measure we defined as the product of its market share and customer loyalty. The dominant firm is very likely to increase its profits at the expense of its rival. Segmentation leads to lower prices for price-sensitive customers, but higher prices for loyal customers. Consumer welfare improves due to intensified competition. We found also that a firm’s decision to implement segmentation technology compels its rival to do the same to remain competitive. However, this decision is impacted by the technology cost. A very dominant firm and a high segmentation technology cost may inhibit its rival from acquiring the technology. As a result, a manager in practice may be interested in tools to support technology acquisition decisions. We define a simple measure called the technology cost to industry dominance ratio to drive technology acquisition decisions. The rest of the paper is organized as follows. We survey the relevant literature and provide a motivating example is provided in Section 2. The model is described in Section 3. An extensive analysis is given in Section 4, and the various segmentation strategies are then summarized. The implications of the model and the impact of segmentation technology costs are discussed in Section 5. We conclude in Section 6 with a summary of the issues investigated and the results contributed. The proofs of our main results are mostly relegated to appendices.

We have examined price discrimination and the effect of market segmentation on competition and consumer purchase behavior in a duopoly. While these issues have been addressed extensively in the literature in monopolistic models, the case of duopolies involving two asymmetric firms remains largely unexplored. Yet duopolies can be found in many economic sectors where firms compete for customers with varying loyalties. Unlike many models, we have also incorporated the cost of technology into the pricing strategies, as it has a direct impact on the profitability of the firms. Based on the results, it appears clearly that firms have a strong incentive to segment the market in order to increase their profits and gain market share. These profits are proportional to customer loyalty, in terms of the premium tolerated by customers at a firm. This result explains in part the proliferation of loyalty enhancing initiatives recently. Importantly, we have identified situations when the competitive game is not a prisoner’s dilemma. This is dependent on the dominance factor of the firms, which we defined as the product of the size of each firm’s loyal customer base and their customer loyalty. The dominant firm is likely to improve its profit at the expense of its competitor. The firms face a prisoner’s dilemma when both of the two firms have equivalent customer loyalty. Price discrimination and market segmentation lead to lower prices for price-sensitive customers, but higher prices for loyal customers. Segmenting the market for not very loyal customers is detrimental for firms due to intense price competition, as can be seen in many real situations. For example, long distance phone service turned out to be a low profit business due to specialized discounts that some firms offered to other rival firms’ customers. AT&T, MCI and Sprint, in particular, engaged in price wars in the late 1990s. While very loyal customers pay higher prices, the majority of the customers enjoy lower prices. The actual savings achieved by less loyal customers offset losses incurred by very loyal customers. Thus, overall consumer welfare is improved by segmentation. Importantly, the premiums incurred by loyal customers never exceed the tolerance threshold that will push them to switch to the competitor. We have also shown that the high technology cost coupled with a high dominance indicator leads to a situation in which only the very dominant firm segments the market. Even though the less dominant firm’s profits decrease, it will choose not to counteract the dominant firm’s segmentation decision. An important contribution of our research is the definition of a practical measure, the technology cost to industry dominance ratio, to support managers in their assessment of the competitive landscape and to drive their technology acquisition decisions. Several assumptions have been made in the model, and the relaxation of each of them is appropriate for further research. First, we assume that segmentation technology is a one time investment, which starts to pay off as soon as a firm adopts the technology. In fact, a firm would need to first collect customer specific data to accurately before forecasting an individual customer’s loyalty tolerance. We plan to incorporate a time lag and the investment effect in future research. Second, we assumed that a customer’s loyalty declines linearly to simplify our mathematical derivations. A U-shaped loyalty curve can capture what happens in the market better, as most of the population is price sensitive and their loyalty is moderately low. A linear loyalty curve favors the firms with high loyalty customers. Segments are also assumed to be of the same size. Actually, a firm’s profit improves if the granularity of segments among the very loyal customers is more refined. Third, a customer’s brand loyalty is fixed in our model. However, we know that the experience of purchasing from a rival firm may change a customer’s behavior over time, as loyalty enticing programs demonstrate. A model incorporating change in brand loyalty over time is likely to capture the behavior of customers in the market. Finally, and more importantly, the model assumes perfect information in that the customers are assigned to the right segment with certainty. Obviously, business intelligence tools are far from being able to achieve this level certainty. We plan to introduce uncertainty in our future studies.