When humans meet data management

Hard data is more than 0s and 1s; real people, with their real-world behavior, are behind those numbers

I recently attended a discussion on enterprise data management strategies with an elite group of IT execs and academics.

One of the more humorous stories told involved a visit to a supermarket, part of a large chain, by one of the academics as part of some research he was doing. This particular store had shown up as an anomaly in the chain’s performance metrics because it seemed to be selling an unusually high quantity of Butterfinger candy bars. But as it turned out, the store wasn’t actually selling lots of Butterfingers.

Instead, the market’s cashiers had discovered that the bar code on the Butterfinger was larger and easier to scan than those on other candy bars, so to save time and speed customers through the crowded lines, they’d keep a Butterfinger handy and simply scan it in place of the other candies.

On further inspection, the professor also discovered large bins of discounted Butterfinger bars on sale at this store, because the scanning situation was triggering the chain’s replenishment system, causing delivery of excess, impossible-to-sell quantities of Butterfingers. So not only was the shortcut leading to bad merchandising and performance data, it was also wreaking inventory havoc.

The moral of this story, one would think, is to teach employees the importance of taking data management seriously and not putting garbage into the system. But not so fast! While the cashiers’ shortcut was certainly messing up the company’s merchandising and inventory data, it was also resulting in higher customer satisfaction and sales volume. Arguably, the cashiers were doing the right thing.

The real moral of the story: Don’t forget that behind all data are real people – who are going to act like people, with an infinite variety of idiosyncrasies and corner-cutting habits.

The same professor had another story, involving a large airline on which he’d flown so many miles that they constantly showered him with upgrades and other perks. He bought a ticket with the airline for a complicated one-week itinerary that involved stopping in several cities all across the country. The trip was going well until severe thunderstorms threw a wrench in things.

Concerned he’d miss an important presentation if his next flight was grounded, he decided to rent a car and drive one leg of his journey. He made the presentation on time, but when he got back to the airport for his next flight he found the rest of his itinerary had been cancelled. It was like pulling teeth to get it reinstated.

Why? The same company that listed him in its database as a highly valued customer also assumed he was a deadbeat trying to cheat the system by skipping a leg of a ticket. One group within the enterprise trusted and valued him, while another was treating him as a criminal – and both were interpreting the same data.

There are lots of issues here about synchronization of siloed business processes and who should really “own” enterprise data. But my biggest takeaway is simply the reminder that enterprise data is a representation of human beings: employees, customers, vendors, and partners. And while human beings do make mistakes or occasional try to shortcut the system, they’re also pretty smart and often do the right thing. Data is just one representation of reality, and you may have to do some detective work to find the truth behind it.

To get this column delivered to your e-mail inbox every week, sign up here.