Lessons from Target’s pregnancy prediction PR fiasco

Several years ago, Target asked its customer analytics group, Guest Data and Analytical Services, to predict which of its customers were pregnant based on customer purchasing patterns.[i]

Now, why would Target marketers want to know this? Because from a marketing perspective, when customers are faced with significant life events—such as getting married, having a child or buying a house—they are more open to changing their purchasing habits. [ii] That means that if you as a company are able to determine when customers are going through a life-changing event, and you send them the right advertisements, you have a higher chance of earning their loyalty as long term customers.

How was the analytics team at Target able to predict something like this? They looked at data from women who had baby registries with the store, merged that with demographics data, and identified a set of 25 products that, when a combination of them was purchased, would indicate an increased probability that the woman was pregnant. Even better, depending on the type of products purchased, Target was able to identify the pregnancy’s trimester, and thus to predict the baby’s approximate due date. This in turn was very helpful so they can market specific products depending on the due date.

This very interesting predictive analytics project was described by Andrew Pole, a marketing analytics manager from Target, who in October 2010 served as a keynote speaker at the Predictive Analytics World Conference in Washington DC (the video recording is available at www.pawcon.com/target). Unfortunately for Target, in February 2012, Charles Duhigg, a New York Times reporter, published an article about this project with an unfavorable tone. The article cited an anonymous example of one angry father who walked into a Minneapolis store with baby-related coupons that were sent to his teenage daughter and demanded to talk to the manager. He was not very happy that, as he put it, “She’s still in high school, and you’re sending her coupons for baby clothes and cribs? Are you trying to encourage her to get pregnant?”[iii] The manager quickly apologized, unaware of what was going on. A few days later when the store manager called to apologize again, the father had had the chance to have a conversation with his daughter, and it turned out that she was indeed pregnant, but hadn’t told him.

A media and public relations storm followed, as many people were outraged at the idea of a company figuring out a highly personal situation like being pregnant. Target responded to the PR storm by banning the reporter from accessing Target employees and corporate headquarters, and by issuing a defensive PR statement.

So, what lessons can we draw from this situation?

Think about ethical implications and customer sensitivities. Target didn’t breach any privacy laws. It used internal customer information and external publicly available data. However, the fact that you can legally learn something private about your customers doesn’t mean that it’s ethically appropriate. Before jumping into the next predictive analytics project, think about how the public might react if they find out what you are doing. Just because you can do it doesn’t mean that you should.

If you decide to do it, be subtle. If you are going to try to predict a very personal and sensitive situation like this one, try to be subtle in the way that you use the information and communicate with the customer. For example, when Target started the pregnancy prediction project, it faced a bit of a backlash when customers started receiving direct mail or emails that were fully focused on baby-related products. It seemed creepy, rather than helpful. After some testing, Target decided against sending standalone marketing pieces that had too much focus on baby products, and instead they started including baby products in their coupon books, together with other unrelated items. Customers were happy to use the coupons, and as long as it was not clear that the company knew something private, they were not upset.[iv]

Streamline a public relations review process. It is ideal to enforce a company policy that asks employees to go through Corporate Public Relations before presenting material at a conference or talking with a reporter; most large public companies have that type of policy. Even better, provide training on why this is important and how to react when approached by a reporter. There are advantages to a company allowing its marketing analytics employees to talk about what they do—for example, increased brand awareness, easier recruiting of analytics professionals, and career growth and motivation for its employees. But there are also potential disadvantages, like inadvertently releasing trade secrets to the competitors or exposing the company to a PR storm like Target’s. In my personal experience working in large companies, the processes in place to go through a Corporate PR review of material are very slow and cumbersome. Many times, it’s very difficult to identify who in the company needs to vet this type of material. Then, when you find them, they don’t have enough bandwidth to review material in a reasonable timeframe. So my recommendation is to streamline the review process and make it widely known internally.

Own the public relations situation and apply crisis management best practices. When something like this come out under the public light, don’t try to shut it down with tactics like banning a reporter’s entrance to your corporate office or prohibiting employees from talking with the reporter (reporters don’t like this, and you’ll hear about it in the press). Instead you need to implement the best crisis management practices. In general, the key actions to take are a) a company executive must take ownership of the situation; b) the spokesperson should tell the truth and should not take a defensive stand, but rather offer an apology; and c) the organization has to take steps to ensure the problem doesn’t occur again, and make the public aware that the organization is taking them. [v]

To sum up, here’s how to avoid running into a similar problem: think about the ethical implications of the analytics and marketing you plan to do; if you decide to go with a strategy that uses sensitive information, be subtle when you communicate with your customers; streamline a public relations review process; and own the public relations situation when you are faced with it.

And yes…. it’s fitting that the company involved in all of this is named Target.