Originally published in The Street
By Art Zeile
The tech world is still pondering precisely what caused a massive point-of-sale computer outage at Starbucks (SBUX – Get Report) last week that left millions of customers in the United States and Canada in various states of caffeine withdrawal.
The grande glitch, which began on Friday night and ended on Saturday, forced some managers to briefly close their stores. Others stayed open, and baristas handed out free drinks.
Crisis management experts will likely give the company high marks for empowering its quick-thinking employees to fall back on the beverage giveaways, which along with the store closures, almost certainly cost Starbucks tens of millions in revenue.
Tech analysts, however, are still scratching their heads. Is it conceivable that the global coffee giant does not have a total redundancy in its systems, known in industry jargon as an n + n system? Or did its plan for launching that back-up system somehow fail? And if so, did the breakdown occur in the technology or was human error or both?
The stock fell nearly 2% on Monday after a very positive reaction to first quarter earnings on Friday. Investors appear to be sorting out the opportunities from the risks.
The Seattle-based company cited an internal failure during a daily refresh of its system, and stated that a cyber attack or external breach was not responsible. The outage also affected Starbucks’ Evolution Fresh and Teavana stores.
Starbucks has used MICROS Simphony for its enterprise point of sale (POS) software since at least 2011. MICROS Systems Inc. was acquired last summer by software and service giant Oracle (ORCL – Get Report) .
MICROS, on its Web site, says its hospitality management platform allows customers to self-store data at one central location or at one of its four worldwide data centers. Is Starbucks’ data system based on one central location, which would, obviously, make the network more vulnerable? Has this happened before? Was data lost? If so, where did it go?
Sources online have speculated that the primary POS table was somehow deleted, which left employees unable to log into cash registers. It would be hard to believe that something as critical as the primary POS table would not have a backup. Best practice for mission critical applications is to maintain more than one backup for that matter and in different physical locations.
At a minimum, a complete and transparent autopsy of the incident is needed. Investors should demand it.
Art Zeile is the founder and CEO of Hosting, a Denver-based managed-cloud-services firm. Follow him @artzeile and via his company.