Chandler, Raymond an American author, once said, "Such is the brutalization of commercial ethics in this country that no one can feel anything more delicate than the velvet touch of a soft buck". Some may say that "business ethics" is an oxymoron. In the world of business where competition is fierce and a "survival of the fittest" mentality seems to rule the jungle, ethics is a controversy that will forever be debatable. And as technology continues to change at an astounding rate, survival is not only dependant on the wisdom and prudence of our business leaders, but also the code of ethics in which they choose to adopt. All areas of business seem to have ethical issues that are debated over, not excluding data mining. According to Laudon and Traver, authors of E-commerce, data mining is "a set of different analytical techniques that look for patterns in the data of a database or data warehouse, or seek to model the behavior of customers" (370). The gathering, utilization, and analysis of data to better understand consumers has been common practice for many years. However, with the progression of technology, methods such as data mining have evolved in the attempt to make the analysis of consumer and consumer behavior more precise. Just as data mining takes data warehousing one step further, web mining takes data mining one step further. Web mining can use the same definition as data mining, except that web mining is conducted within the realm of the internet and world wide web. This paper will explore the purpose, methods, and ethical issues of web mining. Although, data mining and web mining are so similar in nature that the two terms could be treated synonymously, for the purpose of this paper, web mining and data mining will be used to differentiate between internet data mining and traditional data mining, respectively. Purpose
Why do companies use data mining techniques? One of the buzz words that has existed in the business world for years and still does today is competitive advantage. A competitive advantage is what a firm has that they do better than any one else. An example would be producing a high quality product or providing a high quality service at price lower than any of their competitors. Some consider the ability to recognize strengths, weaknesses, opportunities, and threats, which is termed the SWOT analysis, better than competitors a competitive advantage. SWOT analysis is also referred to as visibility. One of the determinants of good visibility is an organization's measurement system and practices; it is the information that gets noticed, captured, analyzed, and acted upon (Hendrix 41). Visibility and the competitive advantage that it offers an organization can be applied not only to the firm's overall strategic plan, but also to areas within the strategic plan, such as their marketing strategy. Marketing strategies have in past been focused on gaining market share, which is a measure of the number of customers. This philosophy that the higher number of customers that a firm can acquire, the higher the market share, which results in higher profits has shifted to a philosophy based on customer share and not market share. The customer-based philosophy demonstrates that a firm becomes more profitable when they focus their attention on gaining share of a customer's business and not the number of customers they acquire (Danna 377). This philosophy is the foundation behind customer relationship management. It is argued that "business-to-business and business-to-consumer enterprises have to build better and more profitable relationships with their customers in a customer-centric economy" (Danna 374). Customer relationship management is basically getting to know each customer. Typically this is done by creating a profile on every customer. A customer profile can contain data regarding product and usage, demographics, psychographics, profitability measures, history
of contacts with the customer, and marketing and sales information based on customer responses (Laudon and Traver 375). It seems like a lot of time and money could be spent on creating this kind of profile on every single customer, which compared to the benefit may not be cost effective. Arguably for some organizations the benefits may outweigh the costs, but for others the Pareto Principle may be the approach in sifting through customers and deciding which customers to target. The Pareto Principle, applied to customer relationship management, simply states that eighty percent of a firm's profit results from twenty percent of their customers. So, if a firm is able to identify the most profitable twenty percent of their customers and create a customer profile on each in order to engage in one-on-one dialogue with each of those customers, the firm would be considered as having an effective customer relationship management system (Danna 377). This notion of gathering data on customers in order to engage with them one-on-one is the tenet behind the marketing strategy of personalization. Personalization segments the market not by groups of customers, but segments the market by individual customers. The goal of personalization is to provide customers "with what they want or need without requiring them to ask for it explicitly" (Mulvenna, Anand, and Buchner 122). Personalization is a system of learning the patterns, habits, and preferences of customers. In order to achieve personalization within a customer relationship management system, the techniques of data mining and web mining are utilized. The number of companies using data mining techniques in order to personalize their marketing strategy is growing at a fast pace. Only two years ago in 2001, over half of Fortune's top 1000 companies had plans to utilize data mining technology (Danna 375). Data mining and web mining involve not only gathering information, but also extracting predictive information
from databases looking for patterns, as well as the cause of those patterns. This process requires sophisticated and complex analytical tools, but more importantly this process requires people who have advanced skills in data analysis and business knowledge. The analytical tools necessary to perform such complex analysis can be obtained through companies specializing in personalization software and as the popularity of data mining and web mining increases, so does the simplicity of software (Brandel 68). Methods What methods are used to gather the vast amounts of information used in data mining and web mining? In discussing methods used for mining data, it is useful to differentiate between data mining and web mining. Another way to look at the difference in methods is to think of the methods as offline and online. Data mining methods have been used for years to better enable organizations to segment their customers. Some of those offline methods include call centers, product registration cards, point-of-sale transactions, and mail-in surveys. In traditional data mining techniques the information gathered from these offline methods was used to segment customers into groups. As discussed earlier, marketing strategies are moving away from segmenting customers into groups and moving towards segmenting customers into individuals. Although the offline methods are still utilized, the Internet has provided a new resource for mining data. Web mining collects data through transaction and navigational activities performed on the internet. Users often browse web sites to comparison shop, to gather information for research purposes, to get current news and financial information, and for a variety of other reasons. The navigation that occurs during a users browsing is documented through an online method called click-stream. Click-stream records mouse clicks of users and records what users
do while viewing a web site. The information gathered from click-stream becomes the psychographic part of the customer profile. It is the online version of behavioral observation. Another method of web mining is in the form of cookies. Cookies are stored on a user's computer that helps to identify the user and the user's preferences. A user can easily deleted cookies stored in their directory, but many users don't even know they exist. A form of cookies
that is used by many researchers are data-augmented URL strings or Web bugs. The URLs contain information such as passwords, personal information, and survey information. The difference between Web bugs and cookies, which is an advantage to researchers, is where the data is stored. Cookies are stored on the users computer, whereas Web bugs are stored on the researchers server. (Miller and Dickson 153) As technology changes so will the methods of gathering information used in web mining technology. One of the emerging technologies are wireless devices. Wireless technologies are being utilized to embed sensors in products and processes in order to gain new insights in consumer patterns, behaviors, and preferences. It is predicted that sensors will be embedded in virtually all products including vehicles, vending machines, clothing, and food products. (Hendrix 44) Ethical Issues With all these existing and emerging technologies designed to gather information on consumers, the prominent question that is debated is "are these methods of gathering the information and the way in which the information is used ethical?" Ethics have been debated over for centuries and business ethics for decades. So, when new technologies are introduced to the business world, a debate of ethics is bound to be not far behind. Business ethics have many subdivisions. The ethics that this paper is concerned with are marketing ethics. There are said to
be three tenets marketing behind ethics. They include 1) both the buyer and seller must agree to and be adequately informed of what is being purchased and how much the purchase price is, 2) neither buyer nor seller is impaired through coercion, severely restricted alternatives, or other constraints on their ability to choose, and 3) both buyer and seller are capable of making a rational decision concerning the transaction (Stoll 122). These tenets will be used as a premise for this discussion regarding ethics in data mining. A summation of this idea is give by Mary Lyn Stoll in her article "The Ethics of Marketing Good Corporate Conduct". She says that in the evaluation of the ethical character of a person or organization "it is of the utmost importance that honesty, non-deception, and non-coercion are maintained (Stoll 126). Another approach to the concept of marketing ethics is to ensure that the value of right conduct does not become secondary to the generation of profit. Considerations of more than the impact on the bottom line, is a recommendation given to those who bear responsibility in designing and implementing a marketing strategy (Danna 386). The third tenet of marketing ethics regarding the ability to make a rational decision seems to the most controversial of the three tenets. Some researchers argue that rationality is not compromised by marketing tactics as long as the individual is consciously aware that their desires are being fostered by those tactics. On the other hand, some argue that an individual's rational is defined by their beliefs, desires, and attitudes. So, when these beliefs, desires, and attitudes are impacted by marketing tactics, an individual's ability to make rational decisions is suppressed. (Stoll 122-123) The core of this debate seems to be focused on the degree of influence marketing tactics have on individuals and to what degree their rationality is impacted. Because the Internet has become a part of so many aspects of our lives it is not hard to determine
that web mining activities used in personalization strategies are going to heavily impact our lives. Whether or not our rational is impacted is a discussion outside the realm of this paper. The concerns regarding ethical conduct in data mining and web mining techniques are not confined within limits of the previously mentioned tenets, but also encompass issues surrounding trust and privacy. Can consumers and internet users trust that 1) they are aware of the content of the information about them that is being mined and 2) is the information that is gathered about them is being utilized by researchers and corporations in an ethical manner? Some say that researchers and corporations conduct should at the very least inform consumers and internet users of the type of information that is gathered about them and the ways in which the information will be used (Danna 386). One suggested approach is to use the Kantian standard of the "Golden Rule", which cautions us not to do anything that we would want to have done to us. An excerpt from "On-line Market Research" summarizes the "Golden Rule" standard by stating, "Whatever the research modality, market researcher have an obligation to conduct research in a responsible manner, recruiting with respondent opt-in and opt-out, protecting respondent confidentiality, being respectful of people's time, and protecting individual privacy" (Miller and Dickson 152). Assuming the approach to an ethical marketing code is the "Golden Rule" standard, some of the questions to ask researchers and marketers would be are you willing to pay higher price than your neighbor for purchasing the same product? Or, do you think it is fair that you are not offered or given the ability to purchase certain products because of the neighborhood you live in? Often times consumers are forced to pay a higher price than their neighbor or are not given the same purchasing opportunities as other based on their demographics. Although some argue that price and product discrimination is necessary to keep buyers from shifting, this type of market
discrimination is considered by some to be unethical (Danna 381). Price discrimination is not a new concept and there are different levels of discrimination. Ethical concerns may not be raised when considering third-degree price discrimination, such as student or senior citizen discounts. Ethical concerns may not even be an issue when considering second-degree price discrimination, such as the different prices airlines charge based on various fare restrictions. However, firstdegree price discrimination, which requires a severe exploitation of price sensitivity between consumers, is often considered unethical. The first-degree price discrimination is made much easier the more advanced data mining and web mining techniques become. The more personalized the consumer profiles become the easier it will be for organizations to engage in first-degree price discrimination. The advocates of price discrimination argue that versioning content and price of products is necessary in order to maximize profitability. However, those who discourage price discrimination argue that it would restrict access to high quality products so that lower income consumers would have no choice but to settle for low quality products (Danna 383). Excluding consumers from a market all together is another way the data contained in consumer profiles is used. Based on their profile, some consumers may not even have the option to purchase a good or service, regardless of the price. This kind of discrimination is referred to as redlining. Whether the discrimination occurs in a physical location (redlining) or on a web site (weblining), some consider the practice unethical because it is not treating consumers as individuals who are capable of makin...