The year was 2012, the company was Target, the story was a daughter’s conception. If you don’t remember the flurry of news items around the incident allow me to provide the short version.
Every time you shop, online or brick and mortar, you provide the retailer valuable data about who you are and what you like. Browser cookies, loyalty cards, custom coupons, credit cards, IP addresses, and more all allow retailers to assign the activity to a unique customer ID. Calling or emailing customer support, making a return, just browsing, you can bet these activities are logged as well.
Retailers then use this data to improve your shopping experience with more relevant promotion, discounts, and messaging. They also identify important segments of the population as core shoppers and target those segments to ensure they win their business.
Target, had identified new mothers as one of these core shoppers. Realizing that new mothers spending habits increase and that their routine changes allowing Target to win more of their business, they worked with their data analysts to see how they could segment this group.
Using the vast accumulation of shopping data Target had collected a statistician on their team, Andrew Pole, was able to identify the shopping profile of an expecting mother. Some of these patterns included purchasing scent-free lotion and soap instead of the standard brand, or loading up on supplements like calcium, magnesium, and zinc. Pole’s analysis identified approximately 25 products that allowed him to assign each shopper a “pregnancy prediction” score. He could also estimate her due date, allowing Target to send more timely coupons.
Target did exactly that.
This lead to a very angry man demanding to speak with a manager of one of their Minneapolis stores.
“My daughter got this in the mail!” he said. “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?”
The manager, while somewhat confused having not been involved in the data crunching or marketing campaign, apologized. He even called a few days later to repeat his apology. While on the phone he learned, the daughter was in fact pregnant and was due in August. Target’s marketing department knew about the pregnancy before the grandfather.
This story circulated the news, not because of the amazing big data crunching, but because of the humor. This type of data analysis happens every day with every person. Thankfully, retailers only use this data to improve our shopping experience, advertising networks use it to send more relevant ads, and social networks improve their suggestions. If you are annoyed, you can opt-out by using private browsing windows or paying cash.
This all changed yesterday when congress passed H. Res 230 allowing our ISPs to collect this data without our consent. While retailers can only collect the interactions, you have with their store, your ISP can collect the entirety of your browsing history. They are then allowed to sell this data to the highest bidder.
Hopefully, most ISPs will only sell aggregated anonymized data. Hopefully this data will only be used by retailers and ad networks to improve our shopping experience even further, but I worry about the other possibilities.
This data could fall into the wrong hands, through direct sale, a purchase through a shell corporation, or through a data breach, which seems to become more prevalent each year. If Target can determine if a woman’s pregnant with their limited data set, imagine what our adversaries could do.
The Iranian Revolutionary guard could run a big data analysis and individually identify citizens who are susceptible to radical Islamic propaganda. They could then target these people online and garner financial donations or worse, they could convince them to commit a terrorist attack like the one in San Bernardino.
North Korea could monitor the search data for anyone living around a military base. If they see a spike in searches for terms like “Korean Weather”, they will be ready for a troop surge in South Korea.
Russia played a part in the past presidential election, we can’t be sure what that part was, but it was clear they were involved. With this type of information, they could exert even more influence on our politics.
Big data is an exciting field with huge possibilities, but we need to remember that there is always risk. I worry about the impact of this decision, but I also remind every company, marketer, and analyst to be careful with the information they collect. Purge old records that are no longer needed, utilize advanced encryption, and limit access to avoid a breach. While it’s tempting to store everything, just in case, be judicious when collecting data and avoiding storing personally identifiable information whenever possible. If we use our best judgement, hopefully we can avoid major issues.