Did Southwest's Tech Fail? — Jason Michael Perry

Many news sources cite Southwest’s lack of investment in technology at its scale. I’m an outsider here, so I can only speculate how this happened, but I find the finger-pointing at technology to be a bit like a Jedi mind trick.

What makes Southwest Airlines so unique is the back-of-napkin concept they proudly celebrate during drink service. Southwest runs a continuous loop model that moves a single plane and multiple flight crew across the US by migrating the team from city to city until a plane fully loops around to its start destination. If you fly Southwest often, as I do, it is not uncommon for planes to stop at intermediate destinations with a continued end stop. Usually, a plane loops every 2-4 days, and you can easily find that if you fly a regular schedule, you’re pretty likely to end up with the same plane or crew.

Nearly every other traditional airline of scale has functioned on the hub and spoke model. Airplanes venture out from a core hub, like Atlanta, and ultimately the majority of airplanes cycle between these hubs that same day. Growth for an airline requires building out infrastructure to add additional hubs around the country or world.

As you can imagine, spoke and hub reward geographic focus – so airlines with this model tend to favor destinations within short proximity of their hub(s). You can focus on one core operational city to house the majority of flight crews, giving you a greater concentration of people you can tap when a person is sick, or you need to add an unexpected route or plane into the mix.

Southwest’s model strays greatly from this and creates an amazingly efficient system with high reward and high risk. In the Southwest model crews can live anywhere, but the system can stage them to the desired destinations to serve a route. Planes no longer needing to cycle home to a hub can venture out for days before returning to their “home” destination.

While hub and spoke airlines have required larger and larger planes to expand the reach of its hubs, Southwest has grown on smaller plans making many, many more stops across a larger loop. This model is amazing and one of Southwest’s core differences – with this approach, it can offer extremely competitive prices and reach destinations that are simply unaffordable for other airlines. So yes, this is how the system works and by design. However, it also means that one airplane needing unexpected service or a call out from the flight crew can cause a cascading delay as it interrupts a plane’s next 2-4 days in its cycle. Something that you can refer to as a cascading delay.

So my read. Southwest’s unique scheduling structure has required some level of custom development or frankestiening of technology solutions to do what they need to do. These systems also operate at an acceptable level of risk or built-in slack. As they are the only airline doing this at such scale, it has become riskier to manage these operational needs and harder for humans to intervene when cascading issues occur.

This recent week introduced the triple whammy. Flight crew shortages due to unexpectedly high volumes of sick pilots and flight attendants; Plane delays and cancellations stemming from one of the largest and most nationwide weather disruptions in recent history; and, as mentioned earlier, a model that, when bent too much, becomes unwieldy and nearly impossible to manage or fix manually. At this point, you have two choices manually fix the system (which has grown so complex it would take weeks to unwind) or hit reset, restage all planes and crew and start the cycle a new.

The reset was the quickest option, and that’s what Southwest did, and I doubt we experience another blip like this again. After all, the system operated like it should and failed when it reached outside of the parameters it was built to handle.

If I’m right what happens next is regulation or self policing that will mean an increase in the slack the system could handle.

So is the technology problem? I doubt it. Southwest got caught running its system way leaner than it could support, and it failed. Can tech make it better and help them look at ways to self-heal when these issues happen next time? Prob. But the true question is when is lean too lean for an airline.

Let’s see what happens.