Article Preview
TopIntroduction
Marketing is the discipline that deals with the marketing mix, which is classically the combination of (1) product, (2) place, (3) promotion, and (4) price, according to the original definition of the “four P’s” (McCarthy, 1960, p. 45). The discipline has developed since 1917, the year in which the first book about marketing was published (Butler, 1917). He already considers the costs of marketing and states “Ineffective retail advertising is of value to no one. The store that gets no satisfactory returns from its advertising expenditure soon ceases to advertise or ceases to exist (p. 325).” In other words, advertising is a financial risk with respect to all four P’s. Responsible marketers in shops – today it might be also Internet shops – have to spend money in advance for a campaign without knowing the returns, only having an uncertain expectation about them in mind. Usually, they make estimations about the expected success, if possible considering experience. Thus, marketing was early seen as a management task (Mauser, 1961).
Without giving a complete historical overview of classical work, the authors want to mention the early work about optimization and risk in marketing (Little, 1966) “In risk analysis a set of alternative plans are laid out, the problem being to choose the best. Cost and revenue forecasts are made, with uncertain elements being described by probability distributions (p. 17).” In this work, the authors will start just by taking this thought into account and model the risk by distributions, involving optimization. Little (1966) concludes: “We have reviewed a number of developments in operations research in marketing. Some of them have a brave new world aura (p. 25).” Fortunately, today an optimal cost modeling can be handled convenient for the marketers without any soupçon of science fiction, as the authors will demonstrate.
From those earlier days on, many statistical techniques were used in the 1970s or 1980s, often named as marketing or market research (Lehmann, 1979), leading to database marketing (Shaw & Stone, 1989; McCorkell, 1997) due to ever increasing amounts of stored customer data. Data mining and machine learning from data warehouses become popular in the following years (Linoff, 2011), including marketing applications of decision trees (Kim, Jung, Suh, & Hwang, 2006), neural networks (Yao, Teng, Poh, & Tan, 1998), genetic optimization (Bhattacharyya, 2003), fuzzy logic (Kaufmann & Meier, 2009), association rules (Wang, Zhou, Yang, & Yeung, 2005), and support vector machines (Shin & Cho, 2006) for example. Two kinds of IT systems can be build around the learning algorithms: (1) recommender systems (Adomavicius & Tuzhilin, 2005) and (2) customer relationship management (CRM) systems (Ngai, Xiu, & Chau, 2009). Recommender systems learn from data specific offers for customers, for example a recommendation for lending the next movie based on previous purchases. CRM systems are a kind of framework for handling all the customer, marketing, and sales data in the sense of a customer life cycle. Even larger amounts of data are stored in file systems with efficient techniques for saving and accessing the data. A popular name for this field of activity is “big data” (Arthur, 2013; Hu, Wen, Chua, & Li, 2014).