Target Using Predictive Analytics to Increase Value Capture
Target has identified 25 products, which when purchased together, predict the likelihood of a woman being pregnant and her associated due date.
For years now Target has been using analytics to increase the wallet share or value captured from shoppers. Using predictive analytics, Target and many of its competitors have been able to categorize customers based on anticipated buying patterns and influence those customers with advertisements or coupons. For example, Target identified adults with young children, and sent them toy catalogs before Christmas. But customer shopping habits become ingrained and are incredibly difficult to change so targeted advertisements meant to alter buying habits have had limited success. However, when customers go through major life events like the birth of a child, their shopping habits do change and thus brand loyalty is up for grabs. Target and their competitors know this and birth records are public so after the birth of a child parents are barraged with product ads. Target wanted to identify these women earlier, to access new parents and establish brand loyalty before other marketers, so they utilized advanced predictive analytics to do so.
How Target creatively uses data to increase value capture: Target assigns each customer a guest ID number which tracks any and all information available like name, credit cards used, email address, purchase history, if they’ve clicked on email advertisements, where they shop online, etc. Target also purchases demographic information like where the customer lives, their job history, and where they went to school.
From here, Target analyzed buying patterns associated with baby registries. They found that 25 products could be used to predict the likelihood of a customer being pregnant and her associated due date. For example, vitamins like calcium, magnesium, and zinc are associated with the first trimester and unscented lotion is usually associated with the second trimester or when a customer purchases cocoa butter lotion, a purse large enough for diapers, and a blue rug, she is likely having a boy. Target did not disclose the analytics used to identify these 25 predictive products but complex regression analyses were likely used to determine which variables (products) were correlated with pregnancy. With this predictive data, Target could then send coupons to mothers-to-be for products before she even knows she needs them. By doing this, Target could capture more value from customers by creating brand loyalty at a time when shopping habits are changing.
A word of caution: Target got better and better at predicting pregnancies and a few years ago they were able to identify that a 16-year-old in Minnesota was pregnant before her father even knew. As the story goes, a father walked into a Minneapolis store with a coupon book for his 16-year-old daughter full of coupons for diapers and maternity clothes. He was angry and accused Target of encouraging his daughter to get pregnant. When the store manager called the father to apologize, the father had spoken to his daughter and learned that she was in fact pregnant. This brings up concerns of privacy. Many Target shoppers would be uncomfortable with Target knowing so much about them. To avoid offending shoppers, Target has adapted their strategy to send targeted coupon books that include many of the items they know the shopper will need but hidden among items Target knows the shopper won’t need like lawnmowers and wine glasses.
Target doesn’t disclose sales figures broken down by division but since the company started this analytics initiative in 2002, sales have increased from $44 billion to $72 billion and the former CEO Gregg Steinhafel has attributed some of this growth to the “heightened focus on items and categories that appeal to specific guest segments such as mom and baby”.
Student comments on Target Using Predictive Analytics to Increase Value Capture
I remember when the article came out exposing Target’s analytical abilities. As you mentioned, I remember many people being very concerned about their privacy and feeling uneasy that a company could know so much about them. However, I actually have the opposite opinion of retailers using my purchase history like this. If Target, or Amazon, or some other retailer can send me coupons and ads for products that are better tailored for me, I’m all for it. I think this type of data analysis is actually to the benefit of the shopper.
Could not agree more with the other comment, but I wonder if that’s specific (generally) to our generation. We all (mostly) dislike advertisements, but as retailers gather more one-to-one relationships with customers and can better target their promotions, I believe we’ll see a shift in feelings towards advertisements.
Thanks for your comments. I also like when I receive more specific advertisements and agree that the differences are probably generational. The wealth of data Target has is a real competitive advantage because they can use analytics to test how best to deliver the advertisements. They could easily create an A/B test and send (A) very specific coupons to one group of pregnant shoppers and (B) specific coupons mixed with irrelevant products to another group of pregnant shoppers. They could then evaluate the results to see if they had a better response for one of the two groups. Since they have demographic information like age, they could refine their analysis further to see if there are differences between age groups.
Such a fascinating story – I totally missed it when this story came out. This is a such an impressive use of massive amounts of data and separating the signal from the noise to find true patterns It’d be interested to know how Target found, and confirmed, these 25 predictors? Were they looking to predict pregnancy (I imagine pregnant customers have a high customer lifetime value)? It’s one thing to have the hypothesis, I wonder how they were able to confirm that they were right.