Big Data Needs a Bath

Somehow, Mary Garnett’s name became inexplicably and erroneously linked with
my company, and we began to receive mail like this addressed to her attention.

As I recall, it was sometime in the late ’90s that Joe sold his printing business. Shortly thereafter, the new owners moved the plant to a new facility, leaving Joe with an empty building. By happy coincidence, my company was searching for a larger home, so we gave Joe a call.

His building was only a block from our existing plant, and since it was already outfitted as a print shop, with plenty of power and good climate control, his building was a perfect fit. Joe sold, we bought, Copresco moved in, and Joe and his wife moved to Florida. That’s the end of the story.

Well, almost.

When someone moves, it takes a few weeks for senders to get the new address and to update their databases. Depending upon how vigilant the addressee is in notifying his vendors, it may even take a few months. If the mailer is really careless (or disorganized, or inefficient) it may take even longer.

In the case of Joe and American Express, it has been two decades and counting. Yes, nearly 20 years since he vacated our premises, the marketing folks at Amex continue to pester Joe at my address, urging him to apply for a pre-approved American Express card.

I’ve been tempted to check the “yes” box and send it back just to see what happens, but I don’t think the experiment would be worth a stretch in prison for credit card fraud. For someone less scrupulous and less risk-averse than me this simple, but glaring, data correction error would be an open invitation to identity theft.

Unlike Joe, Mary Garnett never lived or worked at my address. Some of my readers know Mary. She held many positions at Printing Industries of America and organized many conferences. After PIA moved to Pittsburgh, Mary remained in the Washington, D.C., suburbs, where she still lives today. Since retiring she has filled her leisure with grandchildren, reading and some really neat hiking adventures in Europe.

Somehow, a long time ago in a database far, far away, Mary Garnett’s name became inexplicably and erroneously linked with my company, and we began to receive mail to her attention.

Let me repeat: Mary never worked here. She never lived here, or anywhere near here. I’ve done some writing and speaking for PIA in the past, which I suppose fell under Mary’s jurisdiction, but I don’t think I ever dealt directly with Mary for any of those gigs. Even if I did, I can’t imagine how her name would end up in a database with my company and address.

I may never figure out how this bizarre glitch first occurred, but the real mystery is: will it ever be corrected?

There’s no mystery in the thank you letter I received from the West Suburban Special Recreation Association after attending its annual fundraising dinner this year.

“Mr. and Mrs. Johnson” were addressed as recipient. Pretty ordinary, except there isn’t a Mrs. Johnson. “Dear Steve and Susan,” ran the salutation. Once again very ordinary, but even when there was a Mrs. Johnson her name wasn’t Susan. I did once bring a Susan as my date to this event, but that was several years ago.

This year, I attended alone. My letter should have simply been addressed to me. Ah, but big data is tenacious. It won’t go away. Once it knows that there was a Mrs., it will insist that I have a Mrs. forever. The “spouse” field must be filled in and, if there is no right answer, plenty of wrong answers stand ready to step in.

Bad data is an inevitable byproduct of the information age. The sheer volume of data collected is part of the problem, but it is the way it is managed that really causes glitches. Fact is, an awful lot of data is simply collected, but never managed or “cleansed” at all.

Customization and personalization are proven to get a reader’s attention. Just make sure that the attention is the good kind - leading to responses, purchases or donations - not the ill-will, poor publicity and lawsuits that result when dirty data is used.

There are tremendous opportunities for those who are willing to develop expertise in the field of data cleansing. Today’s big data needs its mouth washed out with soap!