Glitch: The Hidden Impact of Faulty Software -- Converging Forces
On July 15, 2009, 22-year-old New Hampshire resident Josh Muszynski swiped his debit card at a local gas station to purchase a pack of cigarettes. A few hours later, while checking his bank balance online, Muszynski discovered that he had been charged $23,148,855,308,184,500.00 for the cigarettes, as shown in Figure 1.1.1
Figure 1.1 Josh Muszynski's bank statement
That's twenty-three quadrillion, one hundred forty-eight trillion, eight hundred fifty-five billion, three hundred eight million, one hundred eighty-four thousand, five hundred dollars. Plus the $15 overdraft fee charged by the bank. This is about 2,007 times the U.S. national debt.
In a statement from Visa, the issuer of Muszynski's debit card, what had happened to Muszynski, along with "fewer than 13,000 prepaid transactions," resulted from a "temporary programming error at Visa Debit Processing Services which caused some transactions to be inaccurately posted to a small number of Visa prepaid accounts." In a follow-up statement, customers were assured "that the problem has been fixed and that all falsely issued fees have been voided. Erroneous postings have been removed this incident had no financial impact on Visa prepaid cardholders."
No financial impact. It's hard to believe that this incident had no impact on Muszynski, considering that he had to go back and forth between the debit card issuer and the bank to address the situation.
Although the issue was settled after a few days, Muszynski was never formally notified of the cause of the error aside from the information he discovered by doing an Internet search. He saw that the $23 quadrillion cigarette charge and overdraft fee had been corrected only after he continuously checked the balance himself.
When I spoke with Muszynski about the situation,2 he said, "I was surprised that they even let it go through at first. Then, after it happened, I was even more surprised that I was the one who had to notify them about it." Muszynski is still a customer of the bank, but now he checks his account balance every day. As another result of this incident, he has shifted his spending behavior to a more old-fashioned approach: "I now pay for things using cash. I used to rely on my debit card, but it's just easier and safer for me to go to the bank and take cash out for what I need."
Now that's something every bank that's making a killing in ATM transaction fees wants to hear—a customer would rather stand in line for face-to-face service than risk being on the receiving end of another glitch.
Glitches Are More Than Inconveniences
Barely a day goes by where we don't hear about one or more computer errors that affect tens of thousands of people all over the world. In the content of this book, when we address glitches, it is from the perspective of software development and its impact on businesses and consumers. More specifically, it's about the way that software is developed and managed and how inconcistent approaches and methodologies can lead to gaps in the software code that compromises the quality of product or services that is ultimately delivered to the customer.
Unless these types of glitches affect us directly, we tend to shrug and write them off. For consumers who are grounded in an airport for several hours or who are reported as being late on their mortgage payments, these types of computer errors have a much longer and far more public impact thanks to the rise of social media tools such as Twitter and Facebook. A quick Google search revealed that the story of the $23 quadrillion pack of cigarettes appeared in more than 44,800 media outlets, including newspapers, blogs, radio, and television.
For the businesses servicing these customers, the cost of addressing these types of issues goes beyond brand damage control. This is because what can initially appear to be an anomaly can actually be a widespread error that is staggeringly expensive to fix. These unanticipated costs can include extensive software redevelopment or the need to bring down systems for hours, if not days, to uncover the underlying causes and to halt further mistakes.
Although charts and tables can try to estimate the cost of downtime to an organization, it's difficult to quantify the impact because of many variables such as size of company, annual revenues, and compliance violation fines.
According to a research note written by IT industry analyst Bill Malik of Gartner, a Stamford, Connecticut-based technology analyst firm, "Any outage assessment based on raw, generic industry averages alone is misleading." The most effective way to gauge the cost of downtime, according to Gartner, is to estimate the cost of an outage to your firm by calculating lost revenue, lost profit, and staff cost for an average hour—and for a worst-case hour—of downtime for each critical business process.3