Big Data and Retail

How Target used predictive analytics to identify pregnant customers and why it matters for retail

In 2002, Target hired Andrew Pole as a statistician. Pole was tasked by Target’s Marketing Department to see if there was a way to use statistics and predictive analytics to determine whether a customer was pregnant.

Pole began his research combing through Target’s baby-shower registry, identifying female customers who had through registering, informed Target of their pregnancy. By observing these women’s purchasing behaviors as they approached their due date, Pole was able to identify useful patterns.

“…women on the baby registry were buying larger quantities of unscented lotion around the beginning of their second trimester.”[i]

As Pole and his colleagues studied the purchasing data, they were able to “identify about 25 products that, when analyzed together, allowed him to assign each shopper a “pregnancy prediction” score. More important, he could also estimate her due date to within a small window, so Target could send coupons timed to very specific stages of her pregnancy.”[ii]

This application of predictive analytics opened a revenue growth opportunity, as Target was able to effectively target women with coupons and ads for the products they would soon need as expecting mothers with the hope to convert a consumer into a loyal Target guest.


However, a year after Pole developed this “pregnancy-prediction model,”[iv] a father of a teenage daughter entered a Target angrily with coupons for baby/maternity clothes and cribs that his daughter had received. After the Target team member followed up with the father a month later, the father shared that he had learned that his daughter was in deed pregnant.


Create Value:

As consumers, we want companies to anticipate our needs, we want products to be available when we need them, and that companies will adapt to our needs as they change.[v] Predictive analytics allows retailers to fulfill these customer demands and expectations.

“Almost every major retailer, from grocery chains to investment banks to the U.S. Postal Service, has a “predictive analytics” department devoted to understanding not just consumers’ shopping habits but also their personal habits, so as to more efficiently market to them.”[vi]

This use of advanced analytics and predictive modeling is changing the face of retail. [vii]

From predicting trends, demand forecasting to maximize inventory, price optimization, and attracting new customers, [viii] predictive analytics is learning more about customers, sales patterns, inventory management and how to attract new customers at major life events when the customer is willing to change shopping habits – such as the birth of a child.[ix]



Capture Value:

Retailers are capturing value from predictive analytics by targeting customers through metrics such as demographics, age, income, and other variables to guide the customer to new products based on previous purchasing histories.[xi] By building a comprehensive customer profile, the retailer can ensure the customer receives the appropriate coupons, advertisements, promos etc. resulting in an increase in purchases and revenue while minimizing the amount of wasted spending on less targeted advertising.

Retailers are capturing consumer dollars by offering incentives at the right time for the right items while the customer is in the right mindset, maximizing the likelihood of a purchase and thus increased sales revenue.




As Target realized soon after news went public of the pregnant teenager, using data to predict a woman’s pregnancy can be a public relations disaster. [xiii]

Target shared that “we are very conservative about compliance with all privacy laws. But even if you’re following the law, you can do things where people get queasy.”[xiv]

In addition to the challenge of identifying the right customer data and ensuring predictive analytics is in deed predicting the right habits and behaviors, retailers must also be aware of the challenge of identifying consumer trends and supporting them without making a consumer feel that their privacy has been breeched.[xv]

After minimizing the creep level of predictive analytics, Target found consumers were willing to use the coupons. Once the consumer entered the store, Target was able to cue up additional targeted rewards utilizing GEO locating technology both upon entering and when walking throughout the store.



As predictive analytics continues to define the retail space, retailers will need to further develop means to differentiate their products and value to the customer. The ability to provide value through a targeted coupon at the right time will no longer be the differentiating factor as all retailers become more highly educated on their use of predictive analytics.

The ability to be the first to predict trends and execute on them seamlessly will become table stakes.

For now though, Pole shared with NY Times, “Just wait. We’ll be sending you coupons for things you want before you even know you want them.”[xvi]



















[Blog Photo]


OpenDoor: applying big data to home selling


Next Big Sound – moneyball for music?

Student comments on Big Data and Retail

  1. Wow this is really interesting I had no idea Target ever did this. All retailers are trying to predict what consumers will buy. Ad targeting is all about ROI so obviously they want to target you with ads on things they think will make you convert. However, my one concern with something like this is that if we get to a place where we use too many predictive analytics, will consumers miss out on interesting new things that they may want but are never exposed to? Or will new products/categories find it hard to break into the mix since they will likely be under-advertised?

    1. Thanks for the comment! As I responded to Julia below – I think Target has done a successful job to date of balancing utilizing predictive analytics to maximize advertising to the right targeted consumer while still being innovative on products in store. I completely understand the concern for a future where data pushes just advertising the guaranteed sales – but Target recognizes the extremely competitive nature of the retail environment and knows that if it doesn’t maintain a level of creative innovation it will become obsolete very quickly. For that reason I believe Target will still innovate as well as market those new products and categories to its consumers while also marketing those proven products.

  2. Thanks for sharing, Megan! Very interesting post. Predictive analytics are a powerful tool for improving ROI and conversions. However, I wonder the extent to which a company can become TOO data-driven. If all marketing decisions are made based on a model, what happens to creativity and risk-taking? Do companies become too short-term focused as they look for only those results that are highly measurable? How do you find the right balance?

    1. Thanks Julia for your comments! Not sure if you are in CMC – but we discussed this exact issue yesterday in our case on the Whiz Kids – essentially the Ford Motor Company fell into this is exact pit fall where they became so obsessed with the financial metrics and data proof to proceed on an idea that a lot of creativity was missed. What I would say from Target’s perspective here is that the predictive analytics are looking more to understand what the consumer will buy next in order to market/advertise effectively. The buyers however are making decisions on merchandise based on past sales data and understanding upcoming trends – meaning that innovation from a product perspective is still very common.

  3. Great post! Predictive analytics is no doubt the way of future retail; currently, ecommerce essentially serves as a mirror of your past behavior, with ads showing up for websites you recently visited (“Remember me?? Go back and buy this shirt you looked at 10min ago!”), which only seems effectual for certain situations and can be pretty irritating for the consumer. I wonder if predictive analytics faces the same pitfalls that financial forecasting does – it relies too heavily on past performance. What “black swan” events, like unpredictable fashion trends or spending shifts, can occur that will disrupt your model? Conversely, I wonder how much of this might become self-fulfilling; Zara claims to have sophisticated demand forecasting tools, but it’s become such a fashion staple nowadays that arguably it itself sets fashion trends for the season.

    1. Yezi – thank you for your comment. I know exactly what you mean regarding advertising based on past behavior. As I wrote this blog post I saw the exact same advertisement over and over again for the shipping company I had researched the day prior. It was quite ironic to write a post on predictive analytics while being specifically targeted with advertising based on my own previous behavior.

      The one pushback I have on your comment is that it assumes that what you are attempting to sell is similar to what I have purchased in the past. Target here is actually trying to use predictive analytics to get you to purchase something entirely new which I think is a fascinating use of the technology. As the final quote says from statistician Pole in the New York Times: “Just wait. We’ll be sending you coupons for things you want before you even know you want them.”

      Of course retailers may not always recognize the future unpredictable fashion trends. However, if they can successfully use predictive analytics to boost up their customer base, they will have access to additional data that may help them to at least recognize when such trends are approaching to react to them accordingly.

  4. Hi Megan, great post. One thing that strikes me as super useful from this data is not just the specific event prediction, but the general life-stage prediction. I think correctly identifying the life-stage of your customer is far more valuable than any single event. This helps you track the customer throughout their buying journeys and hopefully predict the next one. For example, though I’d coupon for unscented lotion, I’d also start retargeting with infant formula so as to drive the behavior I want when the baby is actually born. A month or two after, I may push infant clothing or even makeup / supplies for the mother. A few months after that, 1st birthday party supplies. And so on and so forth.

    1. Meghana – Thank you for your comment! I completely agree that targeting for additional items once the baby has been born is a great way of capturing the additional value. One of the big takeaways I found from reading the New York Times article specifically is that customers have a few moments in their lives where they are willing to change their shopping behaviors and the birth of a child is one of them. This is why predictive analytics identifying which customers are pregnant is so significant – it allows Target to encourage the shift in a customer’s behavior towards purchasing more of their regular basket at Target as Target provides all the items you need throughout your pregnancy. For those who had registered with Target’s baby registry – these customers were already receiving targeted coupons throughout the pregnancy and following the birth of the child as predictive analytics were not necessary here.

  5. Nice post Megan. The baby example is fascinating. A question I have is how Target is able to accurately track an individual’s different transactions over time? I understand that in the baby registry example the user has a profile tied with all their transactions. But what about the rest of Target’s shoppers? I can see challenges with building a comprehensive profile. I know CVS and other drugstores do it via their loyalty cards but I don’t recall Target having one.

    1. Chun – thank you for your question! You are correct, Target doesn’t have a loyalty program per se however it does have 3 financial products (Target Debit Card, Target Credit Card, and Target Visa Credit Card) which allows Target to track your purchases. The NYTimes article answers your question well so I am sharing this passage below. Essentially it comes down to creating a “Guest ID Number” which is developed over time through all of the customer’s actions in a Target store. Hope this helps to clarify!

      “For decades, Target has collected vast amounts of data on every person who regularly walks into one of its stores. Whenever possible, Target assigns each shopper a unique code — known internally as the Guest ID number — that keeps tabs on everything they buy. “If you use a credit card or a coupon, or fill out a survey, or mail in a refund, or call the customer help line, or open an e-mail we’ve sent you or visit our Web site, we’ll record it and link it to your Guest ID,” Pole said. “We want to know everything we can.”

      Also linked to your Guest ID is demographic information like your age, whether you are married and have kids, which part of town you live in, how long it takes you to drive to the store, your estimated salary, whether you’ve moved recently, what credit cards you carry in your wallet and what Web sites you visit. Target can buy data about your ethnicity, job history, the magazines you read, if you’ve ever declared bankruptcy or got divorced, the year you bought (or lost) your house, where you went to college, what kinds of topics you talk about online, whether you prefer certain brands of coffee, paper towels, cereal or applesauce, your political leanings, reading habits, charitable giving and the number of cars you own.”

  6. Thanks for sharing, Megan. As many other comments mention, I agree that predictive analytics will only gain in popularity and customer acceptance moving forward. Following up on Meghana’s comment, I wonder whether there is an ethical way to monetize this data beyond Target. For example, other non-competitor third parties selling things that Target doesn’t carry (i.e. daycare companies) would be very interested in targeting new or expectant mothers. Maybe in the future, there is a way to anonymously target customer segments with useful third party ads/promotions. Or perhaps that’s a bridge too far for a B&M retailer.

    1. Christy – thanks for your comment! This is a fascinating proposition. I can completely understand the potential value for a non-competing third party seller to gain access to identified pregnant women.

      The only way I can see Target agreeing to such a data sharing program would be if the guest opted in herself. In addition to the PR challenges Target has already faced from recognizing and targeting a pregnant teenager – Target would never sell their guest’s data. Especially given the known consequence that the guest will then be targeted for additional services she did not inquire into.

  7. Super cool stuff. Do we have any sense for the degree of coupon conversion? Of course it will be imperfect given that customers may not always use the coupon at the point of purchase, etc, but I’d want to know as someone working on this campaign. And I wonder how rich of an inventory of coupons needs to be maintained by Target to retain continued customer interest. Discount fatigue!

    1. Thanks for the question. I unfortunately do not know the coupon conversion. However the NYTimes article seems to infer that it is at least significant enough that Target saw value in continuing such targeted advertising and discounting.

      As for discount fatigue, Target’s use of predictive analytics here is to provide coupons and discounts for items that the guest needs and most likely is already planning to purchase. I would argue that that is different than other types of advertising where I am shown an ad for clothing/shoes etc that I do not necessarily need but may want. Additionally, I would reference Meghana’s point above regarding the potential to expand such couponing beyond the pregnancy to include items needed following the birth of the child. This allows Target to continuously coupon for new items that the customer may need which will hopefully reduce discount fatigue.

Leave a comment