Already this year has witnessed a newsworthy software bug in Google’s heavily-hyped Nest smart thermostats that drained the devices’ batteries overnight, garnering icy attention from its users. Are the software sins of the past going to be repeated in 2016?
The result of a December software update, Nest users woke up to chilly homes in January—not the worst thing, but potentially dangerous or property damaging in some situations. Google isn’t in danger of being sued, because according to New York Times tech columnist Nick Bilton, users were frozen out of legal action due to an arbitration clause in their service agreements. Still, just think about the impact this event had on its customer satisfaction and public reputation.
Sadly, these kinds of update-related glitches are nothing new, nor are business disruptions from “mysterious” software bugs. Last year was rife with them. Every week seemed to bring another story of an enterprise brought to a virtual standstill. Major ones made the headlines, but for every one that made the news there were likely hundreds or thousands—large and small—that avoided unwanted media attention.
What can we learn from them?
Top five software glitches of 2015
NYSE shut down – On July 8, computer systems at the New York Stock Exchange went down for nearly four hours, rattling already anxious investors. An expert in Wall Street trading systems said it appeared the result of a faulty system upgrade, and the exchange also reportedly told investors the same.
United Airlines grounded – The same day as the NYSE system woes, computer problems caused United Airlines to ground all flights worldwide—nearly 5,000 of them—for more than an hour. What was the cause? Not hacking, according to airline officials. The FAA attributed it to a glitch in the software automating the airline’s operations.
Starbucks has bitter refresh – A daily system refresh caused point-of-sale systems at Starbucks across the U.S. and Canada to crash in April, forcing baristas to hand out thousands of free drinks. Unlike other major 2015 glitches, this one didn’t hurt the company’s customer reputation much, but it likely cost the coffee-giant $3 or $4 million in lost revenue.
Air Canada’s sweet deal – In August, a computer glitch caused another airline to offer a ridiculously too-good-to-be-true deal: ten one-way business class Air Canada trips from anywhere in western Canada and the U.S. west coast for only $90 apiece. Social media turned the accidental offer into a monumental hit, before Air Canada stepped in, cancelled it and refunded the buyers, claiming it was a “computer loading error.” The error echoed ones that happened to United and Delta Airlines in 2013, but those airlines accepted the loss.
Challenged Alaskan rollout – It was announced last year, that a troubled software rollout in Anchorage would go ahead, despite being dramatically late and over budget. By 2015, the project to replace the city’s aging ERP system had exceeded its original $9.6 million budget by $25 million and missed its July 2012 launch. A $750,000 audit confirmed the project was salvageable, but at significant further cost to the city.
It’s easy to shrug and assume that glitches like this happen as a matter of course, or are simply a cost of doing business in an always-on, connected world. But that couldn’t be farther from the truth. There’s simply no reason for companies or their stakeholders to settle for sub-par business execution.
Yes, upgrades and innovation are introduced at a rapid pace today, but there are ways to deploy them while ensuring that every business process and enterprise app continue to run exactly as they should. How? By the end-to-end testing of every business process, at the pace changes and upgrades are introduced. As they told us in school, “Check your work!” The same holds true for today’s enterprise applications.
Unfortunately this can’t be done manually, at least not cost-effectively or practically. But, as usual, technology has stepped up to meet the challenge. Today’s automation platforms allow business process testing at almost any scale and at the frequency needed to catch everything – even across the most complex enterprise. Automated business process testing can safeguard against such software-related disruptions and failures, and keep your business off the lists of top glitches in the years to come.