After one of the most catastrophic meltdowns in the history of the aviation industry, Southwest Airlines is resuming its full schedule today, and is promising operational reliability. Will the airline be able to deliver?
In this post:
Southwest Airlines’ awful meltdown
Over the past week, Southwest Airlines has canceled well over 10,000 flights, and has stranded over a million passengers. While the situation was precipitated by bad weather, the extent of the meltdown was entirely within Southwest’s control.
The problems ultimately came down to Southwest’s horribly outdated crew scheduling software, which is supposed to assign pilots and flight attendants trips. The software works fine when there are minor operational issues, but it’s not advanced enough to reschedule employees on trips when there are severe operational issues.
Not investing in this technology was pure negligence on Southwest’s part, and this is embarrassing for an airline of this size (Southwest carries more passengers domestically than any other airline). Executives at the airline had talked about plans to invest in this technology in the future, but seemingly kept pushing it off. And then this happened.
Perhaps as bad as the meltdown as such was how poorly Southwest executives communicated throughout the whole saga. For the first few days there was nothing public coming from Southwest’s CEO. The airline spent several days doubling down on how operational issues were due to weather and the fact that the airline operates point-to-point flights, even though neither is to blame for the extent of Southwest’s issues.
Southwest’s “great reset” happens today
For the past several days, Southwest Airlines has canceled a majority of its flights day after day. The logic is that the airline was focused on being able to fully reset its network. Since the crew scheduling software at Southwest is so bad, this was all a manual effort on the part of Southwest’s crew scheduling department.
So rather than continuing to fail day after day, the airline greatly cut back its network for several days, with the goal of having a day where operations fully return to normal. That day is supposed to be today, Friday, December 30, 2022.
Southwest has assured passengers that today should be the day where things reset. And indeed, so far things are looking good. According to data from FlightAware, as of 6AM ET on Friday, Southwest has canceled only 39 flights (0% of the schedule), and has delayed 59 flights (1% of the schedule).
While I imagine the number of cancelations and delays will increase somewhat, hopefully it turns out to be an smooth day for the airline. Personally I’d feel pretty confident that Southwest will operate a reliable schedule today, unless there’s some surprise breakdown (whether with technology or weather).
Hopefully this is the end of the Southwest meltdown. Though of course even if the airline operates reliably, there are still a lot of stranded travelers around the country. After all, flights around the holidays have been fully booked, so there haven’t been many flights on which to accommodate people who are stranded.
Bottom line
It’s a big day for Southwest Airlines (and aviation in the United States in general), as today should hopefully mark the end of mass cancelations. Southwest has been operating a small schedule for the past several days, in anticipation of being able to reset its operation. Today is that day, so hopefully the airline is able to deliver. Things are looking good as of now, but it’s still early…
Do you think this is the end of Southwest’s operational issues?
I have very little confidence. Remember back in oct 2021 the entire SW system was affected for almost a week by bad weather in Florida that nobody even talked about outside of that region of Florida.
Ah ha! Now we know why Southwest continued to offer "free" checked bags after every other airline charges something for them; and why they have that unusual free-for-all seating system; and why they don't play well with other airlines or routing systems. Because they can't! They can't charge for bags. They can't charge or keep track of seat assignments. They can't grant other systems access to their data because their own system is SO decrepit....
Ah ha! Now we know why Southwest continued to offer "free" checked bags after every other airline charges something for them; and why they have that unusual free-for-all seating system; and why they don't play well with other airlines or routing systems. Because they can't! They can't charge for bags. They can't charge or keep track of seat assignments. They can't grant other systems access to their data because their own system is SO decrepit. These systems are incredibly complex to design, build, and maintain -- it will take A Very Long Time for Southwest to recover, but perhaps they can thus leapfrog over others, by building from scratch? A rosy scenario, indeed.
You can't rebuild in a week. So nest crisis will be another crisis.
"as of 6AM ET on Friday, Southwest has canceled only 39 flights (0% of the schedule)"
Obviously cancelling 39 flights is not zero percent of the schedule but rather the software truncating or rounding down. Surprised Lucky didn't comment on what nor did anyone else.
It wouldn't take but someone quoting Southwest as stating they cancelled "0% of the schedule" for this to turn into a false statement.
Speaking of software and IT, and apologies for straying from the main topic, but what about United Airlines' 40 year old SHARES reservation system - and its myriad of other systems hanging on by band-aids? One of those other systems is a crew scheduling system iirc. Are there any updates on this 40 year old SHARES system and a possible replacement? I would hope United would be a bit worried that a similar situation could happen to them.
SHARES is only one component in a very complex technology stack that United probably uses. From what I know, United has probably been the the most proactive in terms of investing in technology of any US airline. We as customers can see the end product of this investment in how useful both their mobile app and website are compared to other airlines (*cough* AA *cough*). They also have a much more modern crew scheduling system...
SHARES is only one component in a very complex technology stack that United probably uses. From what I know, United has probably been the the most proactive in terms of investing in technology of any US airline. We as customers can see the end product of this investment in how useful both their mobile app and website are compared to other airlines (*cough* AA *cough*). They also have a much more modern crew scheduling system compared to Southwest's that allow pilots and other crew members to simply log in via an iPad app while also automatically handling IRROPS situations. If anybody has more detail into what this looks like from a crew members perspective at United or any other airline I would also be very interested!
Out of curiosity, I went to see who the Chief Technology Officer of the company, and good lord! they don't have one. How does a company this big doesn't have a dedicated CTO is beyond my comprehension? That clearly shows the technology is just an "after thought" for Southwest. Sad state of affairs!
https://www.southwestairlinesinvestorrelations.com/corporate-governance/senior-executive-leaders
They do have a CTO, the title is "Vice President Infrastructure and Services at Southwest Airlines".
Thanks Brian! Glad they have one. I was expecting more like an executive role where they have their own divisional budget, cost-centers etc., looks like they are rolled up under someone else and it can be challenging to have proper budget allocation, given there can be competition from other VPs.
@yp - you're right, but you wrote rather in corporate-speak.
Put more bluntly: their head of technology is not on the Senior Executive Team or (confirming your original comment) "not in the C-Suite".
Whereas a "Chief Administration and Communications Officer" is.
On a slightly related topic... I am one flight short of qualifying for A-list next year. My 25th flight was cancelled in the great SWA melt down. I called them today about it and they said any flight scheduled between Dec 20th and Jan 2nd WILL COUNT TOWARDS STATUS OR CP QUALIFICATION EVEN IF IT WAS NOT TAKEN!
I'd love for somebody to explain--rather than just assert--why it's harder to recover from a large storm with a mesh than with hubs. I don't find that intuitive.
In a big storm, pretty much everything shuts down. Once the airports dig out, a hub airline is left with a bunch of planes at its hubs, a bunch of pilots at its hubs, and many flights scheduled to depart from hubs.
On the other hand, a...
I'd love for somebody to explain--rather than just assert--why it's harder to recover from a large storm with a mesh than with hubs. I don't find that intuitive.
In a big storm, pretty much everything shuts down. Once the airports dig out, a hub airline is left with a bunch of planes at its hubs, a bunch of pilots at its hubs, and many flights scheduled to depart from hubs.
On the other hand, a mesh is left with planes everywhere, pilots everywhere, and flights scheduled to depart from everywhere. This seems more complicated, but not like a structural disadvantage with adequate software. Either way the stranded assets are in a distribution approximating the way the flights depart.
What am I missing?
It’s been largely reported that SWA indeed does not have adequate software capable of performing at the scale of their current operational size or complexity when faced with disruption events of this magnitude.
Right. Totally. But in lots of places, including this comment thread, people have remarked that it's not just an IT issue, but an inherent limitation with their network structure (e.g. Krugman's NYT editorial yesterday). My intuition was that an adequate computer system should be able to solve that problem. I don't understand why that isn't the case.
Perhaps easiest to grasp by:
At a hub, if a crew times out, the aircraft is accessible as soon as a new crew shows up (which at a hub, is often)
If at an outstation, the aircraft is unusable until either that crew has completed rest, or a different aircraft turns up with deadheading crew.
For an airline with hub and spoke, timeouts would usually occur roughly 50% of times at hub.
Perhaps easiest to grasp by:
At a hub, if a crew times out, the aircraft is accessible as soon as a new crew shows up (which at a hub, is often)
If at an outstation, the aircraft is unusable until either that crew has completed rest, or a different aircraft turns up with deadheading crew.
For an airline with hub and spoke, timeouts would usually occur roughly 50% of times at hub.
For an airline with star network, timeouts at outstations would tend to be a higher proportion of the total.
I suspect the main reason you are seeing this (somewhat valid) argument being pushed is because the airlines with hub and spoke (which inconveniences passengers, is harder to manage, and increases pollution) don't want it challenged.
grichard- Believe the weakness with the star network design is the crew rest situation. Once a plane lands and there is no replacement crew (due to previous cancellations) the plane is parked until a crew arrives and this delay is more likely to set off a cascading event. Without IT to calculate crew rest, it is nearly impossible to manage it manually. A hub and spoke arrangement is better able to recover, it may lose...
grichard- Believe the weakness with the star network design is the crew rest situation. Once a plane lands and there is no replacement crew (due to previous cancellations) the plane is parked until a crew arrives and this delay is more likely to set off a cascading event. Without IT to calculate crew rest, it is nearly impossible to manage it manually. A hub and spoke arrangement is better able to recover, it may lose a city but the rest of the network is okay. Of course, if AA or DL or UA had an IT collapse on the order of the WN disaster this week, their hubs wouldn’t save them either. I watched a great 10 minute rundown of the WN meltdown on Juan Brown’s blancolirio YouTube channel yesterday evening, it explains everything better than I can.
There should be opportunity to grab some of the recently laid off software engineers from big tech/FAANG companies to turn this around and rearchitect their systems hopefully
Regular schedule? We shall see.
The problem wasn't the inability to find programmers it was a focus on dividends and the stock price instead of investing in systems. They have been using 20+ year-old tech to schedule crew.
FYI, all of the US majors used companies like Deloitte and Adobe to build their scheduling software, apps, etc.
Pretty much, a former SWA pilot did a reddit thread the other day, and he mentioned this all started when Gary Kelly took over when founder Herb Kelleher stepped down from the CEO position.
Kelleher was very operational focused when he ran the company, hence their good profitability and efficiency that Southwest is known for. Along with Kelleher was usually in the trenches helping his employees and to inspect how things were going and...
Pretty much, a former SWA pilot did a reddit thread the other day, and he mentioned this all started when Gary Kelly took over when founder Herb Kelleher stepped down from the CEO position.
Kelleher was very operational focused when he ran the company, hence their good profitability and efficiency that Southwest is known for. Along with Kelleher was usually in the trenches helping his employees and to inspect how things were going and staff morale.
It's why current and former Southwest who were there during Kelleher's reign will talk fondly of him along with his quick wit and humor.
Kelly had an accountant background and not so much an operational one. So when he took over, that's where his focused lied along with the COO that was appointed at the time. And basically it started the ongoing deterioration of operations and staff morale. Culminating in the diaster that is currently happening. There's hope that things might change for the better as their current CEO is an operations focused person, but it'll take a few years to fix the IT problems at Southwest.
You would be a fool to do so. One, the initial layoffs especially at big tech is not competent engineers. Rather it's dead beats that have been on the company's list of who to ditch first with less chance of lawsuits than straight up firing them. You don't want to pick up other company's losers. Second, companies like airlines generally rely on really really incompetent outsourcing instead of staff IT.
To the extent Southwest's problems can be solved with tech, they will be solved. Eventually. It'll take several months, in an absolute best case scenario (i.e. the CEO authorizes immediate no-holds-barred investment into tech with job postings for $2M+ annual comp packages for software engineers and operations research PhDs), to throw money at the problem - hire competent technical people - and test/roll out modern software infrastructure.
But some of their problems have no solutions....
To the extent Southwest's problems can be solved with tech, they will be solved. Eventually. It'll take several months, in an absolute best case scenario (i.e. the CEO authorizes immediate no-holds-barred investment into tech with job postings for $2M+ annual comp packages for software engineers and operations research PhDs), to throw money at the problem - hire competent technical people - and test/roll out modern software infrastructure.
But some of their problems have no solutions. Foremost, the non-hub/spoke model is a hardline weakness in IRROPS that no amount of investment will solve.
I'm drunk right now by the way, having used my $10 monthly Chase GoPuff credit on a 750ml bottle of Gentleman Jack. I mildly regret this purchase, I guess Gentleman Jack tasted premium to me when I was in graduate school, but I should have sprung for something actually premium.