To ‘Daughter Killed in Car Crash’: When Big Data Goes Bad

In early January, Illinois resident Mike Seay received a disturbing letter from OfficeMax.

At first glance, it looked like a typical piece of junk mail—except it was jointly addressed to Seay and “Daughter Killed in Car Crash.” And yes, Seay had lost his 17-year-old daughter, Ashley, in a car accident in February 2013.

Disturbed, Seay did the logical thing and called OfficeMax customer service, where a manager told him “this can’t be happening.”

But it certainly was happening, and OfficeMax ended up offering a standard-issue apology that threw a third-party provider under the proverbial bus:

“We are deeply sorry that Mr. Seay and his family received this mailing from us, and we are reaching out to Mr. Seay to convey our sincerest apologies on this unfortunate matter. This mailing is a result of a mailing list rented through a third-party provider. We have reached out to the third-party mailing list provider to research what happened. Based on a preliminary investigation today we believe this to be an inadvertent error; and we are continuing the investigation.”

Seay told the Los Angeles Times that he wasn’t a big OfficeMax customer, and that he wouldn’t have told anyone at the company about the car crash: “That’s not their business.”

Neither OfficeMax nor the unnamed vendor explained how information about a teenager’s death ended up in a dataset used for marketing purposes. But Pam Dixon, executive director of the World Privacy Forum (a nonprofit group based in San Diego), told the Times that companies collecting and merging enormous volumes of public records has resulted in datasets full of death, disease, and tragedy: “There were probably other people on that list that lost their children in car accidents, and we probably haven’t heard about them, and [Seay] just had the gumption to take it public.”

This incident reinforces how powerful marketing tools have become in our era of so-called Big Data. Now that we have software capable of collecting mountains of structured and unstructured data and knitting it all into comprehensive datasets, it behooves us to put the right filters and protections in place to ensure that information is used in the proper ways. Otherwise, innocuous mailings about discounts on office supplies can trigger quite a bit of heartrending trauma.


Image: windu/