What the Southwest Airlines Meltdown Means for You
By now you are all aware (some of you painfully so) of the historic failure of Southwest Airlines’ systems this week. What started as a simple, routine weather event during a holiday week cascaded into a literal collapse resulting in almost 2500 canceled flights. The actual chain of events have been extensively documented in the Wall Street Journal and elsewhere, so I won’t repeat that excellent coverage. Rather, I want to talk about how this event holds lessons for you and your business, no matter what industry you are a member of.
Technology risk is a hidden demon. It lurks in the shadows rather than cry out in monthly reports. It grows over time, unseen, until that “just right” perfect storm coaxes it out of its lair. And when it comes, it does not come quietly or in isolation. It roars across the enterprise like a wildfire. In my nearly 40 years of modeling businesses, I have seen more than a dozen of these kinds of events, most of which doomed the firm.
At ground zero of the disaster, you will find a common pattern: some structural flaw in the company: a subsystem that is too weak, a process that is out of place, an asset that is more fragile than its heft would suggest. Ironically, the hardest-working people in the company, determined to get the product out the door by all means necessary, find ways of working around such flaws, ultimately masking the underlying symptoms.
What if we could find these not-so-obvious flaws before they take down the company?
The answer is not terribly complex: Digital Twins.
Simulation models of companies—Digital Twins by another name—can give us visibility into how the firm and its processes and assets perform under a wide variety of circumstances, both real and imagined. Some might refer to this as stress testing or scenario planning. In any case, we are experimenting with a digital copy of the company under controlled, laboratory-like conditions in slow motion.
- “What happens if we lose our best customer?”
- “What happens if the Interstate by the plant closes for two days?”
- “What if any two machines on Line 12 go down?”
- “What if our defect rate swamps the QA department?”
- “What if demand doubles next year?”
I am certain that most people working in businesses could easily create a long list of potential stressors to throw at a model.
“What happens when we have a winter storm on a heavily traveled holiday week?”
That’s the kind of experiment that should have been run well before today’s headlines, and one that is easily obtainable by firms willing to explore and experiment. Sadly, many firms are solely reliant on the judgment and experience of human Subject Matter Experts (SMEs) and fail to ask such questions out loud.
I have observed another aspect of events like these: lack of data-driven evidence. It’s one thing to say that executives “should have known” about this problem or that, but in too many situations the people attempting to elevate a risky issue do not bring the data with them to that argument. Their assertions just become one of many voices that fail to stand out. Simulation exercises, however, can visually show the consequences of certain actions (or inactions) thereby making a far more persuasive case to a busy executive.
All companies have weaknesses. Smart companies find them and fix them using analytics.
