Every time you shop at a retailer, you’re sharing valuable information that can be leveraged to figure out what products you like, what promotions they should be offering you, and perhaps…whether you’re pregnant.
It’s precious for retailers to grab parents at the crucial moment that their family is about to grow. If a store can identify this life event and develop a relationship where they can serve all of a customer’s baby needs, it can cultivate a habit where a customer frequently returns for high dollar, high frequency, and high margin items like diapers and baby formula. The business of predicting baby bumps can be a lucrative one. The problem is, it’s also really creepy.
Big box retailer Target attempted to thread the profitability/creepy needle by appointing statistician Andrew Pole to derive a “pregnancy score” from user data. Effectively, Pole was to develop a model based off of past purchases. He determined that there were 25 products that when purchased together, such as unscented lotion, vitamin supplements, and washcloths, were highly predictive that a customer was pregnant. They were so effective at predicting this event that there were reports that Target was sending advertisements for baby products to women whose families didn’t even know they were pregnant yet including a teenage mother who had not broken the news to her father.
These types of campaigns need careful evaluation before they are released into the wild. There was significant media backlash due to the sensitive nature of the targeted demographic. It turns out people don’t want to feel like someone is spying on them, something that the campaign’s architects understood well, “…as long as a pregnant woman thinks she hasn’t been spied on, she’ll use the coupons…As long as we don’t spook her, it works.” Their solution was to subtly place these coupons into more generic looking advertising circulars, intentionally slipping in products in which a customer would be uninterested in to throw them off the trail that the coupon book was perfectly tailored to them.
Partially due to these efforts, Target’s revenue figures have improved dramatically. In 2002, when the analytics department was founded, revenues were $39.9 billion, and in 2017 they almost doubled to $71.9 billion. While the uptick in sales is easy to measure, the risk posed to the brand is more difficult to quantify. There is a growing subset of customers who are actively searching how to avoid loyalty programs and companies that are leveraging this type of information to market more effectively. While the number may be small currently, after big data breaches have come up in the news, the concern is at the forefront of the public’s mind.
While it’s a near certainty that big data efforts like these are here to stay at large retailers like Target, it’s important to keep in mind the human side of how customers will react. People are beginning to realize the subtle nudges that retailers are giving us to manipulate our behavior in ways optimal for them, but disadvantageous for us. The strides made in machine learning and big data will only accelerate this trend, so next time you approach the checkout counter, beware that the machine is watching and knows more than you might know about yourself.