When data gets into the hands of people who don’t understand it, completely false stories end up being told. Maybe you saw the CNN article this week with the eye-catching headline “American Airlines cuts 31,000 flights from its November schedule.” It’s important to know that this is a gross misunderstanding of how American creates its schedule. This was a non-event, and I wouldn’t be writing about it if the record didn’t need to be set straight.
American, like all of the US network airlines, files a placeholder schedule that goes out for sale 331 days in advance. It’s a rolling schedule, so each day, one more day comes up for sale into the future so it’s a constant 331 days available for booking at any given time.
Because of this, American — along with others that follow this process — doesn’t put much thought into the schedules that are that far out. The airline just uses a previous schedule from a previous season that is a true placeholder and isn’t grounded in reality. When it gets closer to departure, American replaces that placeholder with something accurate. This will happen over time in fits and starts.
For example, you can book all the way out through July 15, 2023 right now, but that schedule for next summer is not real. The long-haul markets are usually solidified first, but even that hasn’t happened yet. As I’ve mentioned, this isn’t just an American issue. Take a look at United. It is also selling flights for next summer, but whether or not they’ll operate… your guess is as good as mine.
United makes a point of doing a big reveal of summer plans in the fall when it comes to international long-haul flying. Last year, it made its announcement on October 14. Until that happens, what you see is nothing worth mentioning. That’s why I could only laugh when the blogosphere breathlessly announced that United would begin Washington/Dulles – Berlin next summer. Yes that route is filed, but it was supposed to fly in 2022 and never did. The airline is probably using a carbon copy of that old schedule. It’s too early to know what will go.
For the domestic schedule, the window is even tighter. Historically, airlines have tried to finalize schedules about 100 days out. That really slimmed down to 30 days or so during the pandemic, but now it’s back to around 100 days again, for the most part.
The best way to illustrate this is with Cirium schedule data. This is the very data that CNN used to claim a huge 16 percent cut in Nov, but… well, let’s go to the chart.
American Airlines Departures By Day as Of August 13, 2022
So, you want to guess where this schedule veers off into ridiculousness? That’s right, it’s on December 1. (That huge dip beforehand is Thanksgiving Day, a normal occurrence.) From December 1, the schedule is insane. It has a flat daily schedule with a weird spike on Saturdays. The flying has no true day-of-week component to it like in previous months, and it is so far above what was actually flown this summer that it’s a joke.
Yes, November is lower than other months, but before Thanksgiving, that’s normal. And after Thanksgiving, I’d expect we’ll see more flights added back as demand shapes up. The point is that November looked like December until that cut a couple weeks ago, and while yes, thousands of flights were eliminated, none of them were ever planned to actually fly.
We can look at the other Big 3 airlines as well since I know you’re curious. Here’s United.
United Airlines Departures By Day as Of August 13, 2022
You can see something similar with United here in that nothing in 2023 looks even remotely close to realistic. But there is more here. Even November looks funny for United. You can see that pre-Thanksgiving period is higher than October with more pronounced variance by day-of-week. I’m not convinced they’re done there yet, and they definitely haven’t put any finesse into the winter holiday schedule. In Cranky Network Weekly, we track these things and only consider United accurate through October right now.
The real question to be asking here is… why is there not a more accurate placeholder put into place? Before the pandemic, schedules did change regularly but not to the sweeping extent that happened starting in March 2020. With networks remade, you would think the airlines would try to put something that’s at least a little closer to reality. American and United are probably struggling to find the right time to create a new placeholder, but I would hope that time is coming.
Delta, on the other hand, has already done work along those lines. Here’s that chart.
Delta Air Lines Departures By Day as Of August 13, 2022
You can see here that the 2023 schedule is clearly not right. It’s far too uniform compared to the 2022 plans, but it is much closer than what American and United have done. Delta has made it a point to create a better placeholder. That doesn’t mean there won’t be more changes — there absolutely will be — but hopefully it will be less disruptive than what American and United have.
So, you can ignore CNN’s story entirely. Yes American cut 16% of flights from its selling schedule, but that was never the schedule it planned to run. It was just a placeholder that has now become accurate. Maybe American will eventually realize that a more accurate placeholder would help stop bad stories like these from ever being published.
And now, a shameless plug. If you want to know what’s really going on, subscribe to Cranky Network Weekly!
Very interesting analysis, Brett, and thanks.
Serious question… Is Thanksgiving typically the lowest capacity day of the year (or close to it) for domestic US flights, in terms of flights/seat-miles scheduled, or are there others?
I’ve flown on Thanksgiving more than once, not least because even with the (much lower) number of seats, prices are often $150-200/RT less than the Friday or Saturday after Thanksgiving, and frankly, the choice was between flying on Thanksgiving or not seeing family at all for that holiday, given the cost of air travel.
With Christmas, it generally feels like there is much more of a variety in people’s trip timings and lengths (some people see family the last week of the year, some more on the front end of Christmas, etc; though flights on Christmas Day due tend to be a bit cheaper), but with Thanksgiving, clearly people don’t want to be traveling on the holiday itself, and there’s a rush to get home on the Sat/Sun following Thanksgiving.
Kilroy – I think it tends to be down there. Christmas Day is low as well.
It never used to have this kind of variation, but that meant fares were a lot cheaper. Now that airlines are better at tailoring capacity to demand day-by-day, it makes it easier to keep fares up.
Makes sense, thanks. As a consumer it’s definitely easier to justify $600 R/T fares for better days when supply is restricted enough on the holidays themselves to keep fares from being too low, and to limit cheaper options. Can’t blame the airlines for maximizing revenue.
Though many of them still have to work, I’m sure that the airline/airport employees also appreciate having more holidays off as well, given the lower number of flights scheduled on holidays, and (hopefully) the bit of slack in flights on Thanksgiving or Christmas gives airlines a little breathing room in the event of weather or systems meltdowns immediately before the holidays.
I’ve flown on Thanksgiving morning (7 AM departure to get home from college), and the Continental hub operation at CLE (dates that trip!) was as busy as ever. I imagine the schedule falls off a cliff in the afternoon on Thanksgiving (and maybe even lets a lot of airline staff — at least airport staff as opposed to flight crews — get home to their families).
Even in aviation, CNN is fake news.
Regardless of how one feels about CNN, I’m not sure it’s CNN as much as it is media outlets in general that feel pressure to have generalist reporters write up (or parrot from others, in many cases) articles/analyses about subjects that they know little about, and don’t want to spend the time to really research and learn about.
That’s the real value in more niche-focused journalism and in long-form journalism (lengthy articles or series of articles, often seen as investigations or exposés of specific topics/subjects, such as political corruption, environmental/social harms, or the influence of big business). That’s also the value in having journalists (and getting news from journalists) who are (if not “experts”) much more focused in specific industries/areas/topics.
The trade-off, of course, is that those types of in-depth reporting often aren’t as accessible or interesting to a more general audience with a shorter attention span. Also, those types of reporting sometimes assume a greater level of reader familiarity with the topic, and tend to be expensive (in terms of requiring a great deal of time/resources relative to page views or ad revenue gained). For those interested in the subject, however, such reporting can be invaluable.
I thought it, but said someone would react without just letting your satirical response to stand. I was right.
Tough to tell satire from reality anymore.
I think half of the actual news headlines i read are from The Onion. Sadly, however, they usually turn out to be real. Sigh.
I got you, but it’s just a matter if that onion is raw or not. If so, that will lead one to tears.
So an airline posts a schedule up to a year in advance, and happily takes money, knowing that many of the tickets it is selling are not remotely realistic and what it has promised will not be delivered. The world starts to realise they are people sold a lie, and they react badly to this, with stories appearing on places like CNN.
If American doesn’t want these stories appearing, then make the placeholder schedule highly realistic – or don’t put tickets on sale. Otherwise, American (and other airlines) needs to expect to look bad in the news.
Taking money for a service you know you cannot or will not provide is called fraud by many people
David, the customers will still get a seat and the product that they paid for in almost every case. The airlines know that only a certain % of seats will sell that far out so they have plenty of cushion to reduce frequency or downgauge an aircraft on any particular route. If anything their flight might shift time slots just a bit.
Looking at the graphs CF has produced, it seems like American’s real schedule is about 20% less than their placeholder schedule. That does not equate to “customers get the product they paid for in almost every case”
If I buy a ticket from Washington to Berlin for next summer on a non-stop overnight flight, I expect to get on a non-stop overnight flight on the same route on the same evening, with a departure and arrival time similiar to what was recorded in the original ticket – maybe 2 or 3 hours either side of the original agreement, subject to *unpredictable* things like weather, aircraft going tech, etc.
That’s what the marketing suggested was going to happen… and that’s what I expect – I’m not interested in having to read through a contract of hundreds of pages of legal-speak trying to figure out what kind of get-out clauses there might be. If an airline knows now that it cannot deliver what it’s selling, then that seems fraudulent to me, and I know I should not trust that airline in future.
Ryanair in Europe may give a lousy in-flight experience…. but in summer 2022 they have flown the schedule they promised… and plenty of people are switching to fly with them instead of more comfortable network carriers, simply because passengers know they will reach their destination as promised
David – if you purchase seats on that non-stop flight to Berlin and it doesn’t go, is changed to a connection or the flight times are significantly changed you can get a refund. This is so common it isn’t even an issue (as noted in the article). Probably half my flights get changed on AA after I book them. Usually minor (moved by 30 minutes or less) but sometimes equipment changes, connections instead of non-stops, etc. It is MY DECISION if I accept the change, contact AA (lifetime elite) to possibly change to another routing or request a refund. Anyone that flies knows this occurs.
Sure you can say you paid XXXX for a non-stop and aren’t getting what you paid for but if you check the contract of carriage that airline has every right to make these changes. Also, even with “Mayor Pete” threatening to fine airlines for selling seats on flights they know they can’t deliver that won’t impact this process. These are put out based on an airlines best estimate (agree it could be refined) and they, in good faith, make changes as necessary. The uproar in DC is more over near term meltdowns (mainly Delta this summer of the 3 majors) where they clearly had staffing issues that caused significant disruption. Yes AA and UA have had issues but theirs were mainly weather related.
“Taking money for a service you know you cannot or will not provide is called fraud by many people.”
I know this is way off topic, but look up a Chinese real estate company called “Evergrand” & the current scandal within the countries banking system to see what real fraud looks like. It can very well bring their entire financial system down with it.
Your argument seems to be “B does bigger fraud than A, so we should forgive A for doing only small fraud”. The person who paid money to A still loses their money.
David – it is only fraud if there is willful intent. Try proving that with a process that has been around for years and is common practice in the airline business. Again, over 99% of people get on the flight they booked or one that is materially the same. Those that don’t have rights and protections. Sure you may snap up a cheap fare and it cost more to cancel and rebook but that is life – you can always accept the airline’s routing change and get where you were going (maybe with a time or travel change) for what you paid. That is all you really bought.
Let’s go to the chart! Sounds like you were doing an impression of legendary NYC sportscaster Warner Wolf whose catchphrase was “let’s go to the video tape!”
Can you explain why American’s schedule is filed 331 days out? Now 100 days makes sense, but 331 days appears a bit random to me, although there’s go to be something to it.
331 days = 1 (leap) year minus a 5-week month
SEAN – I assume this was just the way it was done long ago in the early days of the GDS. It’s effectively an industry standard, though there are outliers.
Isn’t it a Sabre standard specifically? I thought at least some Amadeus airlines file 355 days out or something (an issue for redeeming AA miles on partners like QF who in the past released some fixed number of business class seats for miles sales at 355 days, so they were gobbled up before AAdvantage members had a shot, back before award availability was revenue managed better).
Alex – Probably. BA is 355 days out and they are on Amadeus. I don’t think it’s still a Sabre restriction though. I mean, I can look up availability up to 1 year out in Sabre today, but it doesn’t show me nearly as much once I get beyond 331.
Yeah. Given that it’s airlines, the fact that the technical capability may have changed two or three decades ago doesn’t mean airline procedures will have been updated to match! I can’t imagine they lose much revenue over that 24 day difference.
Oh, I see – learn something new every day.
It’s also not just American but it’s all the legacy. Southwest, Frontier, etc., fill their schedules closer in.
Super excellent article CF. In fact, it would be nice if a person could just find when schedules get ‘real’ by each carrier on an ongoing basis.
As aviation enthusiasts, we always find fault with what the general media says about aviation because they never understand the nuances and details.
That said, the practice of publishing a placeholder schedule doesn’t seem right to many consumers because other industries don’t intentionally file a product which they know will not be delivered as sold. It is one thing to make minor tweaks – such as 10-15 minutes different – but it is quite different to no longer operate certain routes or pull down X% of your system capacity – which airlines are still doing post-covid.
However, the DOT is getting more and more adamant about making sure that airlines don’t “bait and switch” someone and most U.S. airlines have long allowed people that are not happy with advance schedule changes to get a refund – but that may mean starting over with a higher fare.
What the general media says about schedule changes pale with the nonstop rants about flight cancellations which are at or pre-covid outside of severe weather days and driven by ATC activity, not weather per se.
I don’t believe that airline schedules are a total nuance, what about season tickets for sporting events. If I am a season ticket holder for a team they generally what me to renew my tickets before I know the next seasons specific schedule. I will also note that 90% of an airline seats are actually sold within 90 days of travel. While the schedule is out there 300+ days in advance in most cases there are very few bookings that may need to be reaccomodated.
With season tickets, you generally know the opponents you will face – unless the composition of a league changes – even if you don’t know the specific schedule.
I agree that there are very few passengers that book 300 days in advance that aren’t part of group or tour packages and there are even fewer passengers that choose not to accept a refund – which is generally offered – but airlines do operate their schedules differently than most industries.
It is also noteworthy that WN and some of the ULCCs don’t sell its schedules as far out but have still done massive schedule changes within 90 days just like the big 3.
However, just like with cancellations and delays, the public and general media get fixated on a story – regardless of content – and keep harping on it long after the facts have changed to the point that the story is no longer a story.
The best way for the airline industry to address the inaccuracies is to run good businesses and treat their customers right. Noise will always exist but if it comes from a small number of people that don’t really represent any significant base, even that noise will die out.
Many industries sell you one thing and deliver something else in exchange. How often do you actually receive the exact product you order online? 80% of the time? How often have you ordered a product with a certain delivery time, only to have it delayed, sometimes by months. The rental car companies are probably the worst offenders of not delivering what they offer. Actually receiving the car you reserve is the exception, not the rule.
Airlines put their schedules out 331 days in advance because their competitors do the same and their customers demand it. I scheduled specialty markets and would sometimes delay putting out a schedule until I had a better idea of which markets we would actually fly (after evaluating the current year results), and would receive multiple customer complaints for the delay. (customers would complain to their travel agent, to the airport, to the station manager, etc. who would forward the complaints to me.) Never mind that only about 2-3% of seats actually get booked that far in advance. Sometimes those buying that far in advance are individual customers, but often the buyers are consolidators and vacation package planners. So there is actually is a lot of outside pressure for an airline to put out a schedule as far in advance as they can (331 days) even if that schedule is entirely accurate. As the saying goes–a bad plan is better than no plan at all.
I meant to say even if that schedule is not entirely accurate.
I’d be curious to see the S-curve (I imagine it would be roughly that shape) showing the % of seats that have been purchased (Y-axis) for X or more days (X-axis) before the flight (or before the start of a the first flight in a roundtrip itinerary, or however one wants to view it)… It may not be public, but I’d be surprised if major airlines DON’T track and review something like this on a regular basis.
Not to defend the airlines or the practice of using placeholder schedules, but let’s not lose perspective… With perhaps a few limited exceptions (VFR travel and travel around heavy demand periods, such as holidays), I suspect that only a small, small fraction of US domestic flights are booked >6 or even >3 months out. It’s pretty common for those who book even a month or two out to have minor schedule changes (times that change by 10 or 20 minutes, for example) applied to their itineraries. As much as we focus on (and cite) the exceptions to this, much of the time such changes don’t really impact travelers in significant ways.
In the case of travelers being moved from nonstop to 1-stop itineraries (likely after paying a premium to book a nonstop routing), or where they are rescheduled to flights that are more than a few hours’ different, especially on long-haul and international flights… Yes, I agree, that can be a significant inconvenience, and I’d argue that those travelers should be able to receive the opportunity to make wholesale changes to their itineraries at no cost to them, or full cash refunds if the changes/options don’t work for them and they no longer wish to make the trip. While those situations aren’t fun to be in as a passenger (I’ve been in them myself, and have used significant schedule changes as leverage when requesting exceptions to policy), I don’t think that those situations affect more than a very small % of travelers.
Kilroy – For the most part, travelers do have those options available to them. If a stop is added, most airlines will allow a full refund. Others may require there to be a minimum time change to permit that, but it’s not uncommon. And many airlines will also allow you to change to use your credit for any other trip with the change fee waived (if a change fee even exists).
Tracking that curve is the essence of inventory management at the airline. The current sales curve is tracked against the historic curve and a desired curve. Adjustments to inventory (how many seats in the various class buckets) are made if the trend is above or below the desired curve.
Has Southwest reverted the their pre-Covid practice of posting a schedule about 8 months about that’s more or less set in stone(they might add flights but never remove them) or are they now more like the legacies?
Dan – It’s a mix. It’s not as rigid as it once was, but Southwest prefers to not make changes very often still.
The news agencies only care about clicks. Telling a complete story lessens the drama and hence clicks. They can’t possibly do that in today’s news environment.
The media has been very clickbaity with aviation stories the last few months.
This is another example of the old Mark Twain truism, “There are three kinds of lies – lies, damned lies, and statistics.” (bent a wee bit to fit the situation)
It seems to me that the details matter a lot here.
–Since few people buy tickets far in advance, the simple fact of decreasing schedule capacity doesn’t necessarily present any problem at all.
–A modest change in schedule times is inconvenient but I think acceptable. For myself, I look at it as a “cost of doing business” for booking way far out.
–A big schedule change, or a change in the number of stops, feels unfair.