Data mining of large data sets, frequently called Big Data, has become a big part of analyzing consumer-related data that companies collect. Basically, data mining is a computational process of extracting information patterns from data sets and transforming it into more organized uses for the future. Usually this information is used to increase sales, reduce costs, or otherwise improve operations or profits. Obviously, if retailers could predict based on past purchasing patterns what consumers might buy in the future, this would be extremely valuable in target marketing.
The statistical team at Target, the giant national retailer, used data mining techniques to determine whether and when a woman was going to have a baby long before she ever purchased diapers from their stores. The marketing plan was then to target ads toward the woman for purchases she might be thinking about for the future. A conversation with one of Target’s statisticians revealed how this worked.
Apparently, Target assigns every customer a Guest ID number and links this with your name, the credit cards you use, addresses, phone numbers, and e-mail addresses they are able to ascertain. They collect all the demographic data they can gather or perhaps buy from others. Their statisticians then decided to analyze the women’s purchases who had signed up on Baby Registries in the past and then study their purchasing patterns before their babies were born. They discovered through data analysis that certain purchasing patterns did emerge. For example, the analysts noticed that women started buying huge quantities of lotion during the beginning of their second trimester of pregnancy. They also discovered that during their first 20 weeks of pregnancy, women stocked up on calcium, magnesium, and zinc supplements. The analysts also detected that as the woman was approaching her delivery date, she bought larger quantities of scent-free soap, cotton balls, hand sanitizer, and washcloths.
After their analysis, the Target team concluded that there was a set of 25 products, which when analyzed together yielded what they termed a “pregnancy prediction score.” They concluded they could closely predict that a woman was pregnant and when the baby was due to be born and could target their ads to the woman based on what they expected her needs at that time would be. To their benefit, by getting the women to shop at Target during an expensive and habit-forming period of her life, she could be made a customer for life.
The Father from Minneapolis
As a result of one set of mailings going out, an angry father showed up at his Target store one day demanding to speak to the manager. In his hands were a fistful of coupons sent to his daughter who lived at home in Minneapolis. His complaint was that his daughter, who was still in high school, was receiving mailed ads for maternity clothes and cribs. He angrily asked the manager “Are you trying to encourage her to get pregnant?”
Not sure what was going on, the manager apologized to the angry father and called him on the phone several days later to apologize again. At this point, the father started changing his tune. He uncomfortably told the manager that he’d had a talk with his daughter and discovered that there were some things going on around his house of which he was unaware. He confessed that he’d discovered that his daughter was due in August and he said that he owed the manager an apology himself.
As a result of the New York Times reporter investigating Target’s data mining approaches and wanting to know more, Target decision makers finally figured out that some of what they were doing may not be appropriate. They eventually concluded that not everyone wanted them to know this much about their personal lives. They discovered that as long as women didn’t know they were being spied on they would use the coupons sent them. But, once they discovered how Target knew to send them these coupons—because of their reproductive status—some of them were quite upset. The company decided to slow down and think about this target marketing scheme a bit more.
Realizing that the company had the ability to create custom advertising booklets, a process that they had used in the past with other products, the company decided to put its ads for baby items in booklets where the baby products were paired with other items such as wine glasses or lawn mowers so that the advertising appeared to be random and not profiled toward an expectant mother. This advertising program was then used. Target discovered that not too long after this revised advertising approach was begun, its sales in the Mom and Baby category increased impressively.
Others Use of Big Data
Many companies today mine Big Data to gather insights into consumers’ likely purchasing habits. For example, Facebook recently reported that it collects data on 1.6 billion consumers, including their “likes” and social connections, to identify behavioral patterns that they might use in their product design and marketing. Increasingly, companies are mining social media data, Internet searches, purchasing behavior, and publically available data sets to improve their products, track trends, and place ads. Recently, Microsoft reported that it had conducted a study suggesting that a consumer’s search queries might provide clues that a person has cancer even before such a diagnosis is made.
Because of the concern for consumer users’ privacy and the public furor it has caused in the past, Facebook announced that it was starting to use an internal review process to assess the ethical impact of each of its research efforts. It has established a five-person group of employees, including experts in law and ethics. Then, if a manager thinks that a research project deals with a sensitive topic, the proposal will get a review by the group to consider its risks and benefits and to consider consumers’ expectations about how their personal information is being used.
Questions for Discussion
1. What are the social and ethical issues in this case? Who are the affected stakeholders and what are their stakes?
2. Target’s practice of predicting a woman’s pregnancy and then exploiting it commercially was legal. Was it ethical? If you were the woman in the case and you learned what Target had done, what terms would you use to describe their practice?
3. Would you feel deceived, tricked, or exploited if a company predicted a purchasing expectation of yours that you felt was private or intimate? What, if anything, would you do about it?
4. How do you evaluate Target’s new approach to disguise the ads by mixing them in with other, unrelated ads?
5. What is your assessment of Facebook’s efforts to protect consumers’ privacy? Or, is it really to protect the company should questions be raised?
6. Have companies gone too far in their statistical analytics and use of Big Data and data mining such as that described in this case? Is it ethical for companies to use data mining to target market customers based on the patterns they detected?