Hacker News new | comments | ask | show | jobs | submit login

It's been discussed on HN before [1], but it's worth mentioning in this thread that there's a YC-backed company – Volantio – working to solve this specific problem.

Most airlines' position is that they have to oversell flights, otherwise planes will fly with empty seats unnecessarily.

But from passengers' point of view, there aren't always enough people who'll voluntarily accept an offer to transfer to a different flight once they are at the gate awaiting boarding [this sentence was edited in response to kortilla's comment below].

Volantio's product applies sophisticated analysis to each flight's booking levels well before departure time (days or weeks before), and will send offers to passengers to offer incentives to transfer to less-full flights, allowing plenty of time for them to change their plans.

So far the product seems to be having an impact for passengers [2], gate staff [3] and airline management [4].

[1] https://news.ycombinator.com/item?id=16836991

[2] https://www.independent.co.uk/travel/news-and-advice/qantas-...

[3] https://www.volantio.com/alaska-airlines-improves-denied-boa...

[4] https://www.business-standard.com/article/companies/indigo-p...

Disclosure: Volantio was a pivot from Adioso (YC W09), of which I was a co-founder. I'm no longer an active employee or significant shareholder. However I remain friends with the team and am pleased to see them doing well and solving an important problem.






> But from passengers' point of view, there aren't always enough people who'll voluntarily accept an offer to transfer to a different flight once they are at the gate awaiting boarding

There are; they just won't accept any offer. Most of the airline offers I have been offered in this situation were paltry: reschedule and some airline "credit".

Give me a partial or full refund, and I'll be much more willing to volunteer. Otherwise, you're seeing passengers value their time higher than some credit and lost time.


I've seen some very lucrative offers. Like $500 credit, upgrade to first class tomorrow, AND a 4-star hotel room. That might not be the norm, but that's a pretty amazing deal.

I was held at checkin for an Economy Plus BA flight from San Diego to London and they asked for volunteers to get the next flight (same time the next day) and you would get $800 cash (prepaid credit card) and a hotel room for the night.

A few people took the offer, I didn't. As a result they denied me boarding and gave me the exact same deal, except it was a $1250 pre-paid credit card. I never expected to be quite so happy about not being allowed on a plane.


At least in the US, involuntarily getting bumped from a flight requires mandatory compensation in some instances (if the arrival delay is >1 hour): https://www.transportation.gov/individuals/aviation-consumer...

There was a story a little while back that claims that customers may not actually get those credits. Basically, this guy only got it because the media picked up on it. https://www.cbc.ca/news/business/air-canada-reneges-on-compe...

I was pulled off an Ethiad flight in Abu Dhabi back in 2011. They gave be a night in a 4-star hotel, 3 free meals and $300 in credit which never got used.

They were asking everyone and most people refused. I should have bartered harder to get an upgrade to business.


Wouldn't exactly say so. They always escalate their offers, starts off being quite meager, but eventually a few hundred bucks credit plus a hotel isn't bad.

On the flip side, I've seen an offer which looked extremely tempting, except I don't think my employer which was paying for the flights would have seen it that way...

> Most airlines' position is that they have to oversell flights, otherwise planes will fly with empty seats unnecessarily.

With most fares nowadays being non-refundable, why should the airlines care whether the seat is empty or not? It's paid for. If it's empty, that's a bonus because they save a little bit of fuel, time at the gate, and the passengers adjacent to the empty seat have a better experience. Empty seats that are paid for should be a win-win.


Because if a seat is sold twice but only used one that's more profit than if every seat is only sold once. Most businesses maximise profits, not passenger experience (and those that work on passenger experience do so because they've decided that's their way to increase profits and they'll do so in a cost effective way).

Airlines say that this means all the seats on a flight can be a bit cheaper as the cost of operating the flight is shared amongst more people but I'm skeptical of that. Businesses like free money and only pass that on as lower prices due to competition when they really have to.


> Because if a seat is sold twice but only used one that's more profit than if every seat is only sold once.

Well, yeah, but that badly misses the point. Selling any product twice is more profitable than selling it only once, all else being equal. But in most industries, selling the same product twice is considered unacceptable. In fact, it would be considered fraud. The story that the airlines tell to make people think it should be acceptable in their industry is clearly bogus.


One valid reason for this practice is the existence of fully flexible fares, that can be switched at late notice. So there is always some uncertainty about exactly how many ticket holders will be wanting to board a given flight.

As for the rest, I dunno, maybe give them a break. Margins run at about 1%, give or take, so two empty seats matters. Heavy financial losses and bankruptcies are common for airlines.

On the practical side, people like to get to their destinations at their preferred time, so it’s good that airlines try and serve people by getting as many people as possible onto their preferred flight. And for environmental reasons people like to know that capacity/fuel isn’t being wasted.

Having worked with airlines for a long time, I’m no great lover of their business practices, but on this topic we can do a bit better than just presuming them to be greedy, exploitative monsters.


> One valid reason for this practice is the existence of fully flexible fares

Airlines are under no obligation to offer fully flexible fares. If they screw up the business model so badly, maybe the answer is to simply stop offering them.


Flexible fares matter a lot, both to customers (mostly business travellers with uncertain schedules) and airline economics.

They don't screw up the business model, the business model relies on it. Without these fares, everyone would be paying more and flexible passengers would be inconvenienced.

But the presence or absence of flexible fares doesn't change the fact that the precise number of passengers trying to board is uncertain, and that will always create a complex optimization problem if you care about both efficiency and customer satisfaction.


No, it's not complicated at all. Just sell "standby" tickets and "guaranteed seat" tickets.

The problem is that right now every ticket is effectively a standby ticket, and this is buried in the fine print.


I keep explaining to you, in good faith, based on 10+ years in this industry, why things aren't all they seem, and you keep inventing new reasons why it really should be that simple but that you seem to think nobody has thought of yet.

A lot of people have thought a lot about these problems for a long time now.

Many of them are both clever and well-intentioned.

I understand that last sentence is where we differ.

So I guess we're done here.


> I guess we're done here.

You've chosen to invoke an Argument By Authority, so I guess you're right.


> If they screw up the business model so badly, maybe the answer is to simply stop offering them.

The problem is that there is lots of competition backed by heavy state funding that can weather out bad events (e.g. a 9/11 scenario with a general downturn, or something as simple as a volcano eruption, or mundane such as strikes of any kind of personnel).

The other smaller airlines have to keep up with the offering of the big giants even if they're unsustainable or risky or they risk directly going out of business.


Allowing airlines to oversell has the effect, in a competitive market, of reducing ticket prices for everyone else.

Put another way: banning oversell would mean that airlines routinely travel with more empty seats. This inefficiency would show through in higher ticket prices.

> In fact, it would be considered fraud.

Fraud involves deception. This doesn't apply in a market environment where it is widely known that they oversell.

The downside is that a free market isn't very good at adjusting to unlikely individual events.

I favour the legislative model in which airlines are permitted to oversell but there is statutory compensation due to anyone who is bumped. This provides the counterbalance to airlines taking it too far in risk to passengers.


I think the Right Answer is for the airlines to sell two different ticket classes: guaranteed seats, which may not be oversold, and standby seats, which can be. That way everyone knows what they're getting.

Standby seats will just jam up the airport and security. Most shoppers will search for the cheapest flight, the airline will sell it as a bargain with the requisite disclaimers that no one reads, and when 100 standby ticket holders show up to the airport all hell will break lose.

Do remember you can't pass security without a confirmed seat. They can't load your luggage without you being confirmed and through security. If every flight on every airline tried this as SOP it would turn out really badly.

If you think I'm kidding check the news around a major holiday with inclement weather. Hoardes of people with standby tickets waiting in terminals, camping out in whatever floorspace they can find. Sounds pretty damn dystopian.


> Standby seats will just jam up the airport and security.

No more than they do now. The only difference between what I'm proposing and what is currently done is that the people who get bumped if too many people show up "volunteer" ahead of time.


You absolutely can and many do pass security without a confirmed seat.

That's the practice in the article, passing ticket holders along and telling them the seat will be assigned at the gate, well after baggage is checked and security lines have been waited.

Selling standby seats at the same/similar overbook ratio as they currently do wouldn't lead to any more of a dystopia then we currently have.


I wonder what the effects would be if the airlines had to put a clear statement directly above the "buy now" button, like "You might not be allowed to take this flight if too many people turn up for it, see our compensation policy if this happens here". It shouldn't make a difference at all if it's actually "widely known" to all participants, but I suspect it actually isn't.

While I seriously doubt that overselling is really widely known, it is also quite an assumption that repeating a widely known fact has no effect on customer behavior.

As something between data and anecdata to support that statement, I've seen measured conversion rates indicate that customers cancel an online shopping deal after they have already agreed to it, after being redirected to PayPal. Those numbers were way too high to indicate just users who couldn't remember their passwords, so we assumed (here's the non-data part) that they jumped off when they realized that they have to log into PayPal and pay real money now. And those things are widely known, certainly so to a customer who has paid with PayPal in the past.


> Allowing airlines to oversell has the effect, in a competitive market, of reducing ticket prices for everyone else.

No. You have to calculate for compensations for the people who have to be left behind because the flight was oversold. As you can read in the comments, those compensations can come in quite expensive. And rebooking passengers to other flights means that you can't sell that specific seat at the right price.

> Put another way: banning oversell would mean that airlines routinely travel with more empty seats. This inefficiency would show through in higher ticket prices.

Again: No. You have 100 seats, you sell 100 tickets. People have to show up to fly with you, no refunds if you don't show up. If this type of calculation does not cover your costs as the airline, you have a bigger problem in general.

> Fraud involves deception. This doesn't apply in a market environment where it is widely known that they oversell.

I find it highly deceptive that somebody is trying to sell me a seat which he has already sold and is now hoping that one buyer does not show up. To be honest, selling the same good twice and only delivering once could very well be a definition of fraud. Try that on some kind of marketplace and you will find yourself in quite some trouble in no time.


> You have to calculate for compensations for the people who have to be left behind because the flight was oversold.

Sure, but factor those in and you still end up being able to offer cheaper ticket prices by overselling. It is clear that this is true from existing market behaviour.

> You have 100 seats, you sell 100 tickets. People have to show up to fly with you, no refunds if you don't show up.

That is one way of running an airline. However, given that some proportion of passengers change their plans in practice, overselling will cause a reduction of ticket prices in a competitive environment. This is just textbook economics and I'm not sure how I can explain this concept better to you.

> If this type of calculation does not cover your costs as the airline, you have a bigger problem in general.

The airline industry disagrees with you.

> I find it highly deceptive that somebody is trying to sell me a seat which he has already sold and is now hoping that one buyer does not show up.

You can find it to be whatever you want, but it is legal and not fraud. Wishing it to be fraud doesn't make it so. You could seek to make overselling illegal. If this were to happen then it would be illegal, but still not fraud.


> You can find it to be whatever you want, but it is legal and not fraud.

It's fraud and it's legal. Which is why the EU cracks down on it pretty hard with large settlements, fines, and the passing of passenger bill of rights. In this thread alone you see passengers offered $100s - $1000s in compensation. Do you think those airlines wanted to pay out compensation? Those fines are statute according to the terms of passenger rights, which you will find printed on the back of every ticket.

Just like speeding, most don't get caught. But just because they don't does not make it any less illegal.


it's flagrant scamming , I don't know who's paying you to lobby for them but your goals aren't aligned with your fellow man

I don't know who you work for . But you care too much about corporate profits and not enough about humans

I'm not employed in any related industry. I do travel for work, so if anything I'm more impacted on the consumer side as a frequent traveller.

However, the consumer has voted, again and again, for cheaper airline tickets over anything else. If that's what they want, then banning overselling would be contrary to that as it would cause ticket prices to rise.

I live in the EU and am quite happy with the current statutory compensation arrangements here.


I'm ambivalent about overselling, but I think your argument is flawed. Consumers can only "vote" for what's on offer, and you cannot individually escape a prisoner's dilemma/race to the bottom scenario without making yourself worse off.

I agree and this is what I meant by "The downside is that a free market isn't very good at adjusting to unlikely individual events" above.

Since banning overselling would increase ticket prices, there will inevitably be some passengers who would suffer as a result. Statistically we'd find that some passengers can no longer afford to go on holiday, etc. These passengers surely want a cheaper ticket so they have _some_ opportunity still.

Instead, I think the most reasonable solution is to allow oversell but enforce adequate compensation by statute, which is what we have today in many places.

The difficulty is in deciding what constitutes adequate compensation. Too much and we'll be back to the "no oversell but inefficient and therefore high ticket price" situation. Too little and traveling would be a mandatory gamble that passengers might be wiped out (eg. a weekend away ruined with no compensation to go again another time, etc). A balance is needed. Banning overselling completely however is I think too far in the wrong direction.


In most industries, selling the same product twice is considered unacceptable.

Not necessarily.

There are a lot of businesses which sell you a service assuming some "average" load. Web hostings, insurance companies for example.

That's even before I start whinging about my monthly train ticket.

But in pretty much any profession there's a chance of a promise not been met. If that happened I would expect an apology, and some kind of a remediation. For extra credit a small token of appreciation would also be nice.


If they know the s’est is going to be empty, they can sell it again.

Sure. The problem is that they sell it again without knowing that it's going to be empty.

(Some) Airlines make money selling to passengers. For example - Ryanair’s entire business model revolves around selling you extras.

No passenger, no profit.


Intrigued by how this works days or weeks before (except in the case of an aircraft swap being known in advance) given that the extent to which passengers will turn up in numbers greater than originally estimated is generally only known at check in stage, probably <24h before departure.

Still, even communication on the morning of the flight beats finding out at the airport (Gold standard for poor service goes to airBerlin, who somehow managed not to inform me that they'd actually cancelled my entire flight days before until I turned up for it in Greece with a handful of other bemused holidaymakers, most of whom unlike me hadn't booked direct. I'd have accepted a much less expensive and more flexible date for an alternative flight though.)


Here's a presentation they did to the Skift travel conference a few months ago, which explains it well:

https://vimeo.com/296887680


>But from passengers' point of view, few people are willing to voluntarily accept an offer to transfer to a different flight once they are at the gate awaiting boarding, even if the offer is quite generous.

This is blatantly false. I fly all the time and I see this happen about every other time at a neighboring gate. Gate agent announced flight is oversold, makes an offer. Several people volunteer. End of story.

You don’t hear about these events because they (used to) happen hundreds of times a day all over the US with no hassle.

Then the stupidity of United and the ignorant backlash of the public ruined it. Infrequent fliers associated overselling with the United incident and completely ignored the functioning 99.9% of cases that benefited all fliers overall.

Volantio’s product is ok, but it’s not a substitute to the problem overbooking was solving, nor is it better for passengers who were flexible with last minute changes (because the offers are worse since rebooks days in advance don’t solve the last minute flex fare changes).

Ugh, sorry about the rant but this whole oversale thing is a perfect example of the outrage culture damaging a mostly functional market due to sheer ignorance.


Fair enough, I've toned down that sentence.

That said, though I can sympathise with the frustration, I've learned that sentiments like "ignorant backlash of the public" and "outrage culture damaging a mostly functional market due to sheer ignorance" don't get us to good solutions.

As you said, the problem is a real one, and Volantio has shown that it can be significantly reduced through smart application of technology well before departure time.


"Then the stupidity of United and the ignorant backlash of the public ruined it. Infrequent fliers associated overselling with the United incident and completely ignored the functioning 99.9% of cases that benefited all fliers overall."

The United incident had nothing to do with overbooking.


While true, that doesn't really contradict the idea that "infrequent fliers associated overselling with the United incident".
jayalpha 5 days ago [flagged]

Why the downvote?

1. Unites flight was not overbooked. They wanted to put some of their own employees on this flight. This has nothing to do with overbooking.

2. Law stats that you can deny boarding if a flight is overbooked. Let someone on board and then throw him off the plane was definitely not within the scope of the law, even should the flight have been overbooked. Try to tell this to the authorities, "we let more people on board than seats available and had to remove some..". Good luck with that.

So based on 1 and 2 the airline was wise to settle with the guy ASAP since throwing him off the plane had nothing to do with overbooking, nor was it in any way covered by law.

Proof that you are an idiot. Downvote this :-)


Agreed. It’s pretty rare for the airlines to not find the volunteers they need. Often they need to bump up the comp to get them, but they get them.

In fact, I’ve volunteered a few times. If I’m flying back home to visit family, arriving one day later isn’t a big deal when the airline pays for a hotel room, meals and offers enough comp that I could fly back home twice.


Well done, Adioso was fantastic.

Indeed - I still use it (among others) more or less any time I'm looking for flights.

It was tomhoward's cofounder Fenn Bailey that also first introduced me to "microcaching" with nginx via some blog post many years back, a trick which has been of much use over the years.


Microcaching is fantastically useful. Way back when, when Edgeio hosted Crunchboard (the Techcrunch job board), we added it to deal with the load it imposed. I don't remember if I'd read about it anywhere first, or was just desperately trying things to avoid it hammering our database.

[For anyone not clear on what is meant: Microcaching is the idea of doing caching with really short time intervals, maybe as low as a second. The point is to set the TTL for the cached data so short that users will rarely if ever notice stale data even for dynamically updated content. The upside is that even a 1 second cache time for a heavily requested page means only 1 request per second per frontend for that resources hits your backend, no matter how much your traffic increases. Super-short TTLs often allow you to cache resources that you otherwise couldn't, and still get a hard upper bound on the load they'll impose]


Awesome, great to hear!

That micro-caching trick was so cool. It wasn't actually for Adioso, but for the Falls Festival, a huge New Years Eve outdoor music festival here in Australia. The coolest application we heard about for that technique was when Google did a doodle for Robert Moog's 78th Birthday [1], and the Moog company's humble little website got hammered.

[1] https://www.google.com/doodles/robert-moogs-78th-birthday


Ha, found (via the wayback machine) the original article[1] and also an HN discussion about it back then too[2]. Fenn didn't mention who it was for, "just some friends" - I guess I figured out the connection and linked it to Adioso in my memory.

[1] https://web.archive.org/web/20111014232647/http://fennb.com/... [2] https://news.ycombinator.com/item?id=3109283


Many thanks. BTW, Adioso is not dead, it's resting. We spun it out of Volantio so I could rebuild it only without all the mistakes. Stay tuned.



Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: