Southwest Airlines’ Reliability Problem Starts With Their Schedule

Southwest Airlines had a major operational meltdown over the holidays, stranding tens of thousands of passengers. It also has focused the Department of Transportation on what they can do, and Southwest has estimated the cost of refunds and other passenger compensation will approach $1 billion. Yet, while this happened, other airlines in the U.S. did not see more than the normal number of disruptions.

Quickly IT was blamed — Southwest says their systems could not keep up with the situations of all the aircraft and crew in a way that would allow them to recover quickly. This is undoubtedly true to some extent, but an IT solution is not all Southwest needs to avoid another event like this. The answer lies more in how they schedule their aircraft and crews.

Out And Back Versus Linear Flying

Most airlines schedule their planes to fly out and back. This means when a plane leaves a station, it returns to that station after its first landing. The aircraft that flies United’s Chicago to Seattle route returns to Chicago. It could go on to San Francisco, another United hub, or Los Angeles, or even Houston or Newark. But it goes back to Chicago. By scheduling mostly this way, United helps ensure that bad weather in Chicago doesn’t affect their other hub operations. American and Delta also predominantly schedule out and back trips, as do other low-cost airlines like JetBlue, Spirit, Allegiant, and Frontier.

Relatively few Southwest flights fly this way. Instead, their trips are linear, and may leave Baltimore for Chicago’s Midway airport, then go on to St. Louis, then on to Dallas’ Love Field. To be fair, I don’t know if a single Southwest plane flies this exact routing. This kind of scheduling is efficient in terms of how long the plane is in the air versus on the ground. But it also links most of their system together, so that once one piece of the chain breaks, many other flights are affected.

Customers who fly Southwest know this to be true, because even when you are first in line to board you often see the plane is half full already. This can annoy those who pay to be early in the boarding process so they can get the seat of their choosing.

Aircraft And Crew Utilization

It’s not only airplane movements, but crew movement that make up part of the complex puzzle that is airline scheduling. Flight crews, cabin crews, and airplanes each have different constraints to be optimized. Just because the plane can fly six or seven hops per day, doesn’t mean the same pilots can fly all of the trips. For cabin crew, it could be even different. In a linear flying structure, crews have to be positioned and maintained to keep the line moving.

Moving from a linear structure to more out and back flying would be a massive change for Southwest. It would affect their aircraft utilization (how many hours per day each plane flies), and likely change where they base their crews as well. This would be highly disruptive but would help to create a puzzle that actually has a solution when things get bad. The IT issues Southwest blamed is that they couldn’t keep up with all the aircraft and crew movements. This was true, but it is because Southwest gave its IT system an immensely hard problem to solve.

How Often To Optimize

Scheduling is about optimizing a big puzzle, with many components each having different constraints. Aircraft need maintenance time. Crews need rest. Spare parts and trained mechanics can’t realistically be placed at every station. Each airport has different availability of gates and parking. Some airports have constrained operating hours, and a few have slots that define specifically when the take off or landing must happen.

So, say you run an airline with 500 planes, and you’ve optimized the schedule to meet all of these constraints in an efficient way. Now, you take delivery of the 501st aircraft. Do you re-optimize the entire network, or do you keep the first 500 doing what they do and create a single new line of flying for the new plane? My sense is that Southwest has grown from a small, highly successful airline into a huge airline without seriously thinking about re-optimizing the whole network. Maybe I’m wrong, but you don’t fix what isn’t broken and that inertia is hard to overcome.

The Puzzle Can Be Even More Complicated

Realities of consumer demand can make the schedule problem even more complicated. What works on weekdays may work less well on weekends, which tend to have fewer business travelers. This means what the schedule looks like by day of week can change, and as demand stabilizes post-pandemic this may create even more optimization opportunities.

In addition to day of week demand differences, connections can also complicate the scheduling puzzle. In big hub operations like American at Dallas or Delta in Detroit, aircraft are timed to arrive at the hub so that customers can connect without long wait times. Southwest carries a lot of connections today, and it isn’t clear to me how much of that is purposely created by the schedule and how much is incidental.

Back To IT As A Solution

Southwest likely needs investment in more IT but this should not be conflated with the unique causality for their operational reliability. IT can solve tough puzzles quickly, but the problem the IT must solve goes a long way to what the IT needs to do. One of the biggest IT mistakes a company can make is to automate an inefficient process. First you fix the operation, then you build the automation to help manage unexpected weather and other disruptions.

Southwest is a great U.S. airline and has been a model for airlines around the world. The problem that caused the holiday meltdown is complicated, and without peeling the onion to a much finer level will a real solution appear. Southwest has smart people and is up to this, but do they have the conviction to do it?