When assumptions about a user can go too far

A few years ago, Target made headlines by announcing a pregnancy to a teen’s father. In this recap of the incident from over three years ago, Target statistician Andrew Pole claimed “he was 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 a pregnancy.” He continued by saying that “what Target discovered fairly quickly is that it creeped people out that the company knew about their pregnancies in advance.” To avoid creeping customers out, Target started including the coupons for baby and pregnancy items in coupon books with other, non pregnancy items like wine glasses and lawn mowers. ““And we found out that as long as a pregnant woman thinks she hasn’t been spied on, she’ll use the coupons. She just assumes that everyone else on her block got the same mailer for diapers and cribs. As long as we don’t spook her, it works.”

Evidently, according to another article on Target, “new parents are a retailer’s holy grail.” Why? because once consumers’ shopping habits are ingrained, it’s incredibly difficult to change them yet there are moments when changing such habits is possible. One of those moments, in fact “the moment, really — is right around the birth of a child, when parents are exhausted and overwhelmed and their shopping patterns and brand loyalties are up for grabs.” Like every other campaign, timing is everything. While birth announcements are public, birth is often too late to target new parents. Marketers would like to reach “women in their second trimester, which is when most expectant mothers begin buying all sorts of new things.”

Join Amazon Mom? I think not.

Join Amazon Mom? I think not.

Target seems to target pregnant women based on actual product purchases and finding patterns in those purchases. Amazon, on the other hand, seems to be so eager to get new moms that it will base its pregnancy campaign on product views. Note this recent note I got from Amazon: “Hagit Katzenelson, we noticed you’ve recently shown an interest in Baby products. Join Amazon Mom and enjoy FREE Two-Day Shipping with Prime, along with 20% off diapers, wipes and other family essentials.” As an aside, no, I am not pregnant and I honestly have no idea what “Baby products” Amazon is basing their pitch on. The only thing somewhat related are some classic children’s books I looked at a few weeks ago. In their eagerness to reach pregnant women, Amazon seems to be making some rather shaky assumptions.

Mining data to predict future purchases isn’t new, and, Amazon’s misstep aside, I assume it’s done more effectively now than when these articles were written in 2012. So why am I writing about this now? I’ve thought a lot about what Eric Meyer calls “Inadvertent Algorithmic Cruelty” and how an algorithm that its creators intended to cause happiness can cause grief when not all scenarios are taken into account. In their eagerness to sell (retailers) or prompt sharing (Facebook) companies either don’t consider the negative scenarios or place their goals above their users emotions. I doubt the latter is the case, but, like I’ve said before, it’s challenging to think of the negative scenarios. In pregnancy and birth, like in anniversaries and annual summaries, not every story has a happy ending. To get a taste of how sensitive this topic is, note the responses to a gaffe by Shutterfly last year when it mistakenly sent out an email to many of its customers to congratulate them on their new arrival. Though this was a clear mishap, and was intended to customers who made “baby-related purchases” on Shutterfly, it could have just been an overreaching algorithm classifying a certain, neutral product (say, a personalized book) as “baby-related.”

The takeaway for marketers and product managers is the same: before launching a campaign or product, spend a moment thinking about the assumptions made and what could go wrong. Consider the potential hurt to certain individuals and look for criteria to use to exclude them. And Amazon, your algorithm might need a few tweaks.

One thought on “When assumptions about a user can go too far

  1. Pingback: What it all boils down to

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s