Life is full of unpredictable disasters: supply chain disruptions, power outages, hurricanes, toxic chemical spills—the list is long. How do you plan for these unforeseeable events?
At the end of last year, Southwest Airlines suffered the most significant and costly disruption in the history of United States airlines, stranding tens of thousands of passengers. Could the airline have predicted that significant snowfall would ground planes that week? No, but they could have guessed that, sooner or later, something would disrupt air traffic and that their antiquated software was in no position to “cold start” and get operations back on schedule. The postmortems of the disruption are still ongoing, but they could have been anticipated and reduced with a premortem.
Premortems anticipate problems before they arise. The classic postmortem asks: Why did the patient die? With a premortem, you imagine the loss before you start—instead of determining the cause of failure after the fact. Companies with global supply chains do premortems. Imagine your supply chain fails. You need to know: Where did it happen? How did your backup suppliers fail? What other contingencies should have been in place?
It’s worth noting that better forecasting complements—but does not replace—premortems. For instance, hurricane forecasts can inform you a week in advance that a storm is likely so you can shut down operations and execute an emergency response. But you still need a plan for what to do if disaster strikes. The goal of a premortem is to imagine the disaster and what steps you’ll take.
So, do a premortem the next time you roll out a new service or make an infrastructure change. Imagine that the rollout failed miserably: What went wrong? Needed more capacity? Slower-than-expected switchover? Was the necessary hardware not available? New software failed? Stakeholders had concerns? Were security weaknesses exposed?
Determine which risks are your biggest concerns, devise a plan to detect potential problems early, and how to respond to them. With a plan in place, if you have a widespread outage, you can get information to the right people and restore services as quickly as possible.
Learn more about new Calix capabilities to manage critical events in our eBook "How To Make Your Broadband Operations a Subscriber Experience Differentiator: Embrace Predictive Analytics and Machine Learning."
Connect with Lyle Ungar, computer and information science professor at the University of Pennsylvania, on LinkedIn.