The New York Times Sunday Magazine recently published a very thought-provoking piece about how retailers can use consumer purchase records to create very detailed and at times frighteningly accurate assumptions about people who can ultimately be used for sophisticated targeted advertising campaigns. To read the article click here.
The main example cited in the article is about how Target can identify women who are pregnant based on their purchases and purchase frequencies of certain items, and can even predict the baby’s due date within a small window of time. Knowing that having a child is a life-stage at which consumers are very likely to establish new shopping habits, Target uses their predictive data to send women advertisements that feature special offers on maternity and baby related items.
The article is quite long and touches on a great number of other issues as well. Of particular interest to me was the idea that there are key life events such as childbirth, marriage, divorce where consuming patterns change a great deal and in ways that one wouldn’t necessarily anticipate (according to the author of the piece, when a person gets divorced, there’s an increased chance they’ll buy new brands of beer.) Also, intriguing was the underlying idea that we are all creatures of habit much more than we’d probably like to think. For example, the article cites a Duke University study that estimated 45 precent of the choices we make in a given day are the result of habit rather than conscious choice.
There is a lot in this predictive analytics process for people interested in marketing and market research to consider. It raises questions about the ethics of learning very personal information about people based on information they might not even be aware they are providing. It also leads one to wonder to what extent predictive analytics will replace or coexist with traditional forms of market research that measure consumer behavior. We in the Bunker highly recommend reading the article, and we’d love to get your reactions to it in the comments section of this post.
If you think about it, isn’t the purpose of ALL research to be predictive? Who cares about the demographic characteristics of people who buy product A. You’re only discovering that so you can affect future purchase, right? that’s predictive analytics.
You’re right. Pretty much all market research could be accurately labeled as “predictive analytics.” The new element here, and the part that makes some people a little uneasy, is that retailers now have the ability to predict the behavior of individuals rather than groups. The traditional paradigm has been, “Let’s market this product to pregnant women,” but now that’s shifting to, “Let’s market this product to Jane Doe, since we are quite sure, through thorough analysis of her purchases, that she’s pregnant.” It’s one thing to be a member of target group, but to be singled out individually seems to many to cross a privacy boundary.
Those kinds of privacy concerns aside, I think another element in this — and this applies to traditional market research as well — is that some people simply resent the notion that their behavior falls into predictable patterns. We all want to believe that we’re too unique and complicated for anyone, especially some analyst we’ve never even met, to figure out. The kind of predictive analysis being done by Target and others confronts us with the idea that not only are we predictable, but the stores we shop at have us figured out to the point where they are one step (or more) ahead of us. For a lot of people, that’s a pretty uncomfortable notion.