The Airline Industry’s Problem with Absolutely Ancient IT

The Airline Industry’s Problem with Absolutely Ancient IT

To try everything Brilliant has to offer—free—for a full 30 days, visit

Watch Jet Lag: The Game at

Buy a Wendover Productions t-shirt:

Subscribe to Half as Interesting (The other channel from Wendover Productions):

Sponsorship Enquiries:
Other emails:

Writing by Sam Denby and Tristan Purdy
Editing by Alexander Williard
Animation led by Josh Sherrington
Sound by Graham Haerther
Thumbnail by Simon Buckmaster


You may also like...

46 Responses

  1. Aaron Disario says:

    I no longer consider Southwest to be a “low-cost airline.” Its fares are often right in line and sometimes higher than the traditional big carriers.

  2. LBGST zockt says:

    The fact that airlines don’t yet extort every passenger individually is the only good thing in this story

    • White Rabbit says:

      Other races in the universe are expanding their territories.
      Before it’s too late, we must secure a lot of celestial bodies and build defenses.
      We need to train everyone in the world to be scientists.

    • Celarc says:

      @Simon Chadwick Fun trick for hotel booking. Check your booking price, and if its to high for you, reset your browsers cookies and go back to the webpage. (Do not reload, input the original link again.)

      On certain booking sites, this is enough to have the algorithm generate a new price.

    • Simon Chadwick says:

      I don’t have concrete evidence, but I suspect hotel and car rental companies do this by tracking browser cookies and IP addresses. Sometimes I have secured better rates when using a VPN via another state or country.
      And don’t always believe the “Last room left at this rate” flags.

    • Flying Cat says:

      There’s no such command in the GDS

    • Ian Bowers says:

      And that’s only because of their complacency. Certainly not owing to any sort of benevolence.

  3. TikkaQrow says:

    I promise that the last 15 years the IT dept has done nothing except warn higher ups of an impending failure of the system, and the higher ups likely said ‘We’ll think about it’ while stretching the backbone of their company on box fans pointed at server racks in rooms where the AC failed 5 years ago.

    Always prepare for the worst, because the worst WILL happen given enough time, and however much it costs to prepare, I promise it’ll be cheaper than waiting for the worst to happen.

    • Edward Chan says:

      Murphy’s law!

    • Andrew Koster says:

      @altrag Executive performance is evaluated on three-month timescales, and the metric for success is “did the company acquire speculative value in the minds of stock market gamblers?”, so that explains the short-term thinking. It’s a very unfortunate set of incentives.

    • DaleC5195 says:

      As an IT guy this is all too familiar lol

    • altrag says:

      @Zahir khan Its never cheaper in the long run, but nobody (at least nobody with power over the direction the company takes) cares about the long run.

      CEOs can just “resign” and move on to the next company they can suck dry, major shareholders will divest and leave the bill for the smaller shareholders that don’t have any real influence.

      Everything in modern business – not just airlines but all publicly-traded companies, which is almost all large companies – are built around the quarterly report. _Maybe_ the annual report if they’re really “forward” thinking. If you can’t show record profits – not just profits but _record_ profits – every single quarter, your company is considered to be one step from failing.

      CEOs don’t get big bonuses for merely non-record profits – they may as well start filing chapter 11 already if that’s all they’re going to produce!

    • Aloha Tigers says:

      @Dino Scheidt

      I would rather NOT have to move every few years and be forced to live in a greedy landlord’s apartments. Too much money. I would rather buy land, build a house and retire.

  4. Marina Weinberg says:

    European passenger rail companies looking at the situation, after having done the exact same thing and now refusing to go back to interoperability in booking systems: finally, a worthy opponent!

    • Vertti Koo says:

      @Zephyros EU has big plans to double the high speed rail traffic by 2030. It benefits everyone.

    • Vertti Koo says:

      Yes. Now the rail companies thinks that they are airline companies 🎉

    • leftaroundabout says:

      @BigHenFor Maybe, but AI has a tendency to fail even more horribly than traditional IT systems under unforseen circumstances.
      There’s really not need to go all the way to AI. Booking isn’t _such_ a high-volume affair (compared to, say, weather data or financial transactions). It would be totally feasible to design this to work smoothly, it would just need to be done with modern frameworks, discipled modelling and testing practices, rather than COBOL or some other pre-1980 crud with layers upon layers of ad-hoc patches.

    • Jaakko Jokinen says:

      The EU mandated interoperability between banks by requiring open application programming interfaces. I suspect it’d be similar when it comes to rail (or flights). The EU mandates that rail operators must provide APIs, and then 3rd party operators build platforms that take advantage of the and provide better booking options for passengers.

      This is a great way for the EU to promote healthy competition.

    • BigHenFor says:

      Shouldn’t AI make it easy to have interoperable schedules? So programming should be less onerous?

  5. Benjamin Mock says:

    Ah yeah… The old Story of “technical dept”. Technicians always warn and management never listens.

  6. Pinkertonfloyd says:

    This isn’t just an issue with Airlines, as someone who’s worked in enterprise IT for Large companies and Governments for a few decades now, everything may look nice, but almost all system backends are running on either an ancient system, or ported over from such (The old software just moved to newer computing systems with a new-ish user interface, but the backend is still talking in what most would call “dead” protocols and database languages. Cobol, Fortran, etc.) Same goes for those “fancy” systems that healthcare use, it’s all based on a 1960’s protocol (M, or MUMPS)
    The problem is they become so used, that nobody wants to be the person that changes it out in fear that it’ll break things more, and or deal with the downtime involved.
    On top of that most “new” systems are being developed by “Cloud” software companies, rather than software you can own/run, customers are being “locked” into a company that can… once under contract, be able to hold your data for ransom and charge whatever they want. (This is the main reason behind the “cloud”, it’s all subscription “locking” in customers, not about saving money or reducing support needs or cost).

  7. Alex's Meme Dump says:

    If you think airlines’ IT is bad wait till you see the hardware that air traffic controllers are running. Call up Denver Center in Longmont and ask for a tour. You’ll probably get three videos out of it

    • Steve Pittman says:

      @romanpfarrhofer Banks have this exact problem – the system sorta barely works and upgrades are both expensive and risky, so they do nothing until they are forced by circumstances to do so, and even then they do so reluctantly and while freaking out non-stop. ‘Woops, our IBM System/36 – the one from like 1982 that still has 8″ floppy drives – broke and no one make replacement parts anymore, I guess it’s time to upgrade to a 30 year old system!’

    • Andrew Koster says:

      @fuck duncan But that’s not how computers generally work. The hardware will eventually break, and when it does, it will need to be replaced with new hardware that currently exists on the market, and that new hardware will have limited compatibility with your legacy software, so that software will need to be changed. There’s no avoiding it.

    • Aloha Tigers says:


      THANK YOU. Someone gets it

    • harzer99 says:

      @fuck duncan You cannot maintain old systems forever. Spare parts become more and more expensive the less parties are operating a system. And with software and programming language you quite quickly run into the isse of them not being supported anymore. Good luck finding someone that fixes a security issue in some Java module that was created by a group of volunteers 20 years ago and has not been worked on in the last 10 years (this stuff happens regularly).
      Maintaining IT infrastructure means updating it. You have stay in this sweet spot of matureness of the product and support availability.

  8. Jean Carlo Castro says:

    I worked for about 5 years with a travel agency, and got really familiar with Sabre GDS. The tool itself was quite marvelous as a new agent, but as the years went on I started to notice the huge amount of limitations it had. For example, the entirety of a corporate account for one of the biggest oil companies in the world could be messed up quite easily by one single agent unticking a box while looking up a PNR. Happened many times, and I swear there was never a change on the GDS to avoid delaying thousands of passengers from booking their flights.

  9. Richard Pauli says:

    The most widely ignored software design rule is to “fail gracefully” Then the rule becomes “Lessons not learned, will be repeated “

    • Shehla Sikander says:


    • leftaroundabout says:

      The most widely ignore software design rule is to fail *early* – better a small, localized crash that can be fixed by the developers actually responsible and while they remember all the context, than keeping it limping along with a bunch of squelched warnings and eventually cause a chain reaction of other failures.

  10. GingerKiwi says:

    My mom worked for Air Canada in reservations and bookings in the late 60’s and early 70’s. They had one of the most up to date systems at the time. It sounds like not much has changed since then, and that she could literally do the same job 50 years later with almost no additional training. (She still remembers most of the 3 letter airport codes too!)

    • Roman Odaisky says:

      @F L The more than 10 countries that aren’t the US use codes that start with various letters, and there, unfortunately, the codes don’t match at all. Heathrow/Gatwick/Stansted/City = LHR/LGW/STN/LCY = EGLL/EGKK/EGSS/EGLC

    • Rahul Jain says:

      @Wesley Nishi nah, that just means you can start typing the airport code even before the customer says it!

    • Wesley Nishi says:

      Must be hard when working fof Air Canada when almost all destinations start with “Y” !

    • Gaby5011 says:

      ​@F L K is for the US, C is Canada’s, CPGrey has a good video about it!

    • F L says:

      @WorldTravel1518 I could see four letter codes to accommodate newer airports, but what’s the point of four letter codes that all start with a “k”?

Leave a Reply

Your email address will not be published. Required fields are marked *