The email last week came as a surprise. I was invited along with three others to chat with Southwest CEO Bob Jordan about the airline’s failures during the holidays. Considering Southwest had been very quiet throughout the meltdown, I was pleased to finally — emphasis on “finally” — see the airline putting itself out there, assuming more small calls like this were being arranged with others. What I learned is that Southwest isn’t willing to say it knows exactly what went wrong just yet, but it has a basic idea and will be working with consultants to reconstruct its failures quickly.
Bob did not mince words in understanding the impact of this event. He repeatedly talked about how the airline “messed up” and how it couldn’t do that again. That sounds easy to say, but not every company would admit it.
What’s harder to say is exactly what happened and how the airline could make sure it wouldn’t happen again. Bob spoke for 20 minutes and then the four of us were each allowed to ask one question. I asked Bob if he’d comment on the SWAPA union’s detailed dive into what it says went wrong. He wouldn’t, saying it was too early to know everything. He started to say he was sure that some of it was right and some was not, but he stopped himself saying he didn’t even know that. The airline wants to do a full autopsy before commenting, but it is finally at least speaking about what it does know.
SWAPA’s piece was detailed and full of conclusions immediately after the event. It had plenty of grenades to throw at management… which is one of the reasons I tend to discount union communications. There’s always some kind of negotiating motive behind them. That’s not to say that what comes out is always wrong, but it’s important to always keep the source in mind. Without question, comparing what the union had to say and what Bob had to say gave me a better sense of what happened.
Though Southwest will admit to having some idea of the problem, it believes that there is more to uncover. The airline has contracted with Oliver Wyman — where, it should be noted, COO Andrew Watterson worked for a dozen years — to do a full reconstruction and review of the mess. Though I have a general disdain for using consultants, this is actually a use I support. It’s a specific, limited task that could benefit from an outside perspective. Bob said this will not take months and will happen quickly. I really hope that’s the case, because the airline seems to want to reserve full judgment until this is done, and that just can’t keep waiting forever.
Fortunately, Bob didn’t just completely defer to this review before commenting on what happened. Here’s what everyone seems to agree on so far.
The Storm was Bad
The storm that rolled throgh on December 22 was very cold, very windy, and problematic in many ways. According to Bob, “we saw things that we just have never seen before with the temperatures. You had jetways that froze, you had deicing fluid that froze. You had aircraft that had ice on the engines in the morning.”
One of the questions I’ve had along with many others is why this was so much worse for Southwest than it was for, say, United which has hubs in Chicago and Denver. There has been talk about network design and all that, but one of Bob’s comments about the storm grabbed my attention. He said, “we had crosswinds that shut us down in Chicago for awhile.”
A look a historical wind data for Chicago shows that starting just after noon, the winds started to ramp up with gusts consistently in the mid-30 kt range. The wind was coming out of the WNW for most of this time, which is just slightly north of due west.
At O’Hare, where United has its hub, the six primary runways are all now east-west oriented which meant there was very little crosswind component. Midway’s runways that would have been in use run NW and depending upon the actual wind degree, would have faced a much greater impact. Little bits of bad luck like that can add up quickly to snarl the operation for one airline and not another.
Overwhelming the NOC
Southwest’s network operations control (NOC) is the operational heart of the airline. As the mess progressed, this drove what Bob called “an historic level of work into the NOC” to get the aircraft in place. And that turned into another “historic” level of work on crew re-scheduling.
Bob made it clear that the technology did not fail. Instead, it just wasn’t up to the task to be able to deal with the crushing amount of work that it needed to do. I liken it to shoveling in a driving snowstorm where just as you clear the snow, another layer falls on top, creating more work and you never get fully caught up. (Keep in mind I live in Southern California, so my metaphor may be woefully inaccurate.)
Cancellation upon cancellation happened close-in, and the tool that SWAPA referred to as SkySolver in its note — now apparently called GE Crew Optimization according to Bob — is meant to be able to run solutions to re-crew future flights. The key word here is “future,” because if there are problems in the past that haven’t been solved, Crew Optimization can’t handle it. And Southwest had many close-in cancels that Crew Optimization couldn’t touch.
Because of this, Southwest needed to bring in an army of people to manually solve all those previous problems, which could then be fed into Crew Optimization to create future assignments. But… there was no manual workforce available to actually deal with this, so Southwest quickly trained up over 100 employees in other functions to be able to work the issues and get good data into Crew Optimization so it could do its job. The airline then cleared out two-thirds of its schedule so that it could get all the pieces in place over a three day period, ready to restart on Friday, December 30.
If this sounds crazy, it is. There needs to be software to handle past problems, and… there is. Bob says that there is a release of Crew Optimization that the airline will start testing this week that will actually solve past problems. I followed up with Southwest to understand whether this is an actual new release or it’s just new to Southwest, and a spokesperson told me “we are working with GE on an update to the software that is tailored to our needs. It wasn’t previously available….”
Begging Customers to Return
Once the operation was up and running again, Southwest turned to figuring out how to deal with the hordes of rightfully-angry customers. As we all know by now, the airline put out vague guidelines that it would process refunds quickly, offer reimbursements, and then it pushed out a 25,000 point bonus to everyone stuck in the mess.
The communications were chaotic at best, and if I’d been given a second question I would have asked Bob if he was pleased with how communications were handled and if not, what he would have changed. Unfortunately, I didn’t have that opportunity. What I do know is that now that this is over, at least customer compensation has been humming along quickly.
Bob says they’ve processed 93 percent of refunds, turning them around within 3 days,. The remaining refunds are those which have come in during the previous three days, so it’s a quick turnaround. Reimbursements for expenses are tougher because they involve reviewing receipts through a more manual process. But Bob says Southwest is now processing 30,000 of those requests a day.
The 25,000 points were deposited almost immediately, and while Bob couldn’t give specifics with earnings coming up soon, he did say that many of the points handed out have already been redeemed, so people are coming back. Of course, this is nearly free travel we’re talking about, so it’s not a surprise. But for Southwest, it just wants to get people back on board any way it can.
The airline has also returned all but one percent of the mountain of bags that were stuck during the event, and what remains are the really hard ones with no tag or no address.
In other words, it sounds like Southwest is doing what it can on the customer side now that the deed is done, but how does it make sure this whole trainwreck doesn’t happen again? On the one hand, Bob really wants to wait for that Oliver Wyman report to come out, but on the other, he knows there are things that don’t need to wait.
What Southwest is Putting in Place Today
Here is what Bob outlined as already being part of the plan to make sure this doesn’t happen again. Until the Oliver Wyman report is done, consider this a partial list.
- Improving the internal “advanced warning dashboard” to include more metrics like number of open crew lines so they can get a better sense of when something bad is happening.
- Keep this new temporary workforce they built to manually process crew schedule changes. They activated 25 of them during the FA A NOTAM outage but they didn’t need them. At least they’re more ready to deal with disruption considering the seemingly-inadequate system they have in place.
- Moving forward on implementing the new release of Crew Optimization that will allow the system to process past problems
- Looking at every process top to bottom, for example, found they didn’t have enough engine covers in the outstations, so they’ll fix that.
- Doing a full deicing review, because it was bad. Denver they use open baskets because it gives better visibility but when it’s cold, that isn’t worth it since frostbite concerns require much more frequent breaks and reduce productivity.
- The Board of Directors has set up an Operations Review Committee, which seems like something it should have done after one of the many previous failures. I don’t see how this is going to help do anything except cover the board’s asses.
Some things are harder to fix than others. For example, we heard all about how Southwest had to manually call crews to get them reassigned. But Bob explained, “we have automatic notifcation in place [but] it takes a contractual change, I believe to require acknolwedgement.” I followed up and asked for clarification on exactly what contract is the issue, and a spokesperson explained “the collective bargaining agreements for Pilots and Flight Attendants contain scheduling rules that, if altered, would require negotiation with those respective unions.”
It is somewhat frustrating to not have more answers, but at least the initial steps seem to make sense based on what we publicly know so far. And regardless of what happens in the Oliver Wyman report, Bob said, “I can’t imagine we don’t boost our [technology] investment based on what we find.” He spent time defending what Southwest spends on tech today, but it sounds like he has an appetite for more.
Could this result in bigger changes around the airline? It’s possible, but I’m guessing probably not. In response to a question about whether a network rework should be considered, Bob left us with this.
I don’t think it’s the network. But there could be things about how we flow crews or how we assign crews or how we schedule trips… you know, those kinds of things.
Just how much Southwest changes because of this remains to be seen.