How UA tries to optimize the load / yield? Role of overbooking
#1
Moderator: United Airlines
Original Poster
Join Date: Jun 2007
Location: SFO
Programs: UA Plat 1.995MM, Hyatt Discoverist, Marriott Plat/LT Gold, Hilton Silver, IHG Plat
Posts: 66,856
As the overbooking discussion seems to have gotten far away from the Skip-the-Waitlist topic, Skip the waitlist (STW) questions
Pulled 10 posts out from the StW thread to create an overbooking / maximizing yield discussion
WineCountryUA
UA coModerator
Pulled 10 posts out from the StW thread to create an overbooking / maximizing yield discussion
WineCountryUA
UA coModerator
Last edited by WineCountryUA; Dec 9, 2019 at 11:48 am Reason: new thread
#2
Join Date: Feb 2015
Programs: united
Posts: 1,636
How UA tries to optimize the load / yield? Role of overbooking
I am thinking that UA is going to significantly improve cash usability in the near future by increasing attractive cash UG offers to prevent too many PlusPoints upgraders from exploiting the system.
In Kirby's dream revenue model, all seats will be sold and no non-cash UGs will be given, but "elites" will still feel valued and appreciated.
A good way of viewing the ideal FF program from an airline exec's POV is that since you want some slack in your system, you have an opportunity to fill that slack in a way that will reward FF'ers for their loyalty and spend.
What IS true is that you don't generally want to be giving away seats that can be sold for a lot of money. Standard awards, capacity controls, impeding the usability of GPU's, exempting popular premium cabins (such as p.s. service) from upgrades, and waitlisting are all mechanisms for doing this. But the basic goal is the same: if a premium cabin seat can sell for >$1,000 (whether in the form of a paid ticket, an instant upgrade from a full fare Y, or a TOD upgrade), it should be difficult for a FF'er to snag it. On the other hand, if a seat is likely to be unsold, and a FF'er wants it, you don't mind if she takes it in one way or anther (RDM's, instruments, even a CPU sometimes) and is happy the airline gave it to her.
Just about every change in the FF program that gets discussed on this board is an attempt by UA to get the program closer to that theoretical ideal.
#3
Join Date: Aug 2003
Location: Northern California
Programs: I want to be free! Free!
Posts: 3,455
This isn't really right. If you are an airline exec, you definitely want to be close to full load with paying customers, but you don't actually want to be at full load. The reason is because IRROPS happen, and you want to be able to accommodate your customers in those circumstances.
A good way of viewing the ideal FF program from an airline exec's POV is that since you want some slack in your system, you have an opportunity to fill that slack in a way that will reward FF'ers for their loyalty and spend.
What IS true is that you don't generally want to be giving away seats that can be sold for a lot of money. Standard awards, capacity controls, impeding the usability of GPU's, exempting popular premium cabins (such as p.s. service) from upgrades, and waitlisting are all mechanisms for doing this. But the basic goal is the same: if a premium cabin seat can sell for >$1,000 (whether in the form of a paid ticket, an instant upgrade from a full fare Y, or a TOD upgrade), it should be difficult for a FF'er to snag it. On the other hand, if a seat is likely to be unsold, and a FF'er wants it, you don't mind if she takes it in one way or anther (RDM's, instruments, even a CPU sometimes) and is happy the airline gave it to her.
Just about every change in the FF program that gets discussed on this board is an attempt by UA to get the program closer to that theoretical ideal.
A good way of viewing the ideal FF program from an airline exec's POV is that since you want some slack in your system, you have an opportunity to fill that slack in a way that will reward FF'ers for their loyalty and spend.
What IS true is that you don't generally want to be giving away seats that can be sold for a lot of money. Standard awards, capacity controls, impeding the usability of GPU's, exempting popular premium cabins (such as p.s. service) from upgrades, and waitlisting are all mechanisms for doing this. But the basic goal is the same: if a premium cabin seat can sell for >$1,000 (whether in the form of a paid ticket, an instant upgrade from a full fare Y, or a TOD upgrade), it should be difficult for a FF'er to snag it. On the other hand, if a seat is likely to be unsold, and a FF'er wants it, you don't mind if she takes it in one way or anther (RDM's, instruments, even a CPU sometimes) and is happy the airline gave it to her.
Just about every change in the FF program that gets discussed on this board is an attempt by UA to get the program closer to that theoretical ideal.
#4
Join Date: Jan 2019
Location: middle of nowhere, formerly TYO/EWR
Programs: UA 1K
Posts: 213
I think dilanesp's argument (which I agree with) is that this "slack" can be taken up dynamically close to departure. As the day of departure approaches, you can be increasingly sure there won't be IRROPs (and also increasingly sure you cannot sell that space last minute) and adjust appropriately by filling that space with upgrades, hence pleasing the FFs and keeping high valued customers happy.
In this model this whole STW thing seems a bit silly, but I suppose it is just another way to keep the highest valued customers happy. A lot of people talk about +Ps as though United giving them out costs them something directly, or that they want to limit the number people have. The only reason it "costs" United anything is via the extremely indirect path that increasing the number means decreasing the success rate, which means FFers will be increasingly displeased with United. The same thing is more or less true for normal MileagePlus miles as well.
In this model this whole STW thing seems a bit silly, but I suppose it is just another way to keep the highest valued customers happy. A lot of people talk about +Ps as though United giving them out costs them something directly, or that they want to limit the number people have. The only reason it "costs" United anything is via the extremely indirect path that increasing the number means decreasing the success rate, which means FFers will be increasingly displeased with United. The same thing is more or less true for normal MileagePlus miles as well.
Last edited by A Little Cow; Dec 8, 2019 at 4:01 pm
#5
FlyerTalk Evangelist
Join Date: Oct 2001
Location: Austin, TX
Posts: 21,413
This isn't really right. If you are an airline exec, you definitely want to be close to full load with paying customers, but you don't actually want to be at full load. The reason is because IRROPS happen, and you want to be able to accommodate your customers in those circumstances.
I suspect, if you offered an exec the chance to sell to 105% on every flight, or to sell to 99% on every flight, where they'd meet their revenue goal on a per-ASM basis in both cases, they'd take the 105%, even if it meant having to go to IDBs.
Or, to put it another way: the last couple of seats on a flight are way too valuable to keep them in reserve in case things go sideways.
#6
A FlyerTalk Posting Legend
Join Date: Apr 2013
Location: PHX
Programs: AS 75K; UA 1MM; Hyatt Globalist; Marriott LTP; Hilton Diamond (Aspire)
Posts: 56,461
Which is why most carriers routinely oversell their flights.
#7
Join Date: Feb 2015
Programs: united
Posts: 1,636
This might be true systemwide, but it's emphatically not true on any given flight. Any airline will happily sell beyond full load, because that's a problem that can be solved by offering people money to go away. The liability incurred from the denied boarding vouchers is much smaller than the incremental revenue from selling an extra seat.
I suspect, if you offered an exec the chance to sell to 105% on every flight, or to sell to 99% on every flight, where they'd meet their revenue goal on a per-ASM basis in both cases, they'd take the 105%, even if it meant having to go to IDBs.
Or, to put it another way: the last couple of seats on a flight are way too valuable to keep them in reserve in case things go sideways.
I suspect, if you offered an exec the chance to sell to 105% on every flight, or to sell to 99% on every flight, where they'd meet their revenue goal on a per-ASM basis in both cases, they'd take the 105%, even if it meant having to go to IDBs.
Or, to put it another way: the last couple of seats on a flight are way too valuable to keep them in reserve in case things go sideways.
So you need some slack in the system. Indeed, overbooking depends on that slack- airlines can overbook because when they guess wrong, they can BOTH pay compensation AND get the overbooked passengers on other flights. Some slack is ultimately required.
#8
Join Date: May 2010
Location: AVP & PEK
Programs: UA 1K 1.9MM
Posts: 6,355
The whole purpose of overbooking is that is gets overbooked EXACTLY with the same amount of seats that UA predicts (and has decent data to back that up) passengers won't turn up for. Obviously no airline would want every passenger to turn up on an overbooked flight.
I don't blame them for overbooking either.
The goal is to fly out with every seat sold & occupied.
They don't want to pay compensation, and they don't want to fly with empty seats.
I suspect UA (and most other airlines) have a pretty good idea which flights they can overbook and to what extent, based on past experience. This will likely vary from route to route, depend on time of day and time of year, etc.
#9
FlyerTalk Evangelist
Join Date: Oct 2001
Location: Austin, TX
Posts: 21,413
I mean, we are getting really into the weeds here, but I actually don't think that's literally true. An airline that is 105 percent sold on all flights- I don't mean simply overbooked, I mean 105 percent of the number of actual seats on all flights show up- will be in deep trouble, because it isn't simply a matter of paying compensation, that airline will simply be unable to transport many paying customers.
#10
Join Date: Dec 2006
Location: NYC
Programs: UA GS (1MM), DL Gold, Marriott Plat, Hyatt Globalist, Avis CC
Posts: 526
Having worked in that department (and not speaking on behalf of them), I can tell you it's amazing how accurate the predictability is and how high the no-show rate is on certain markets on certain times/days!
#11
Join Date: Feb 2015
Programs: united
Posts: 1,636
The whole purpose of overbooking is that is gets overbooked EXACTLY with the same amount of seats that UA predicts (and has decent data to back that up) passengers won't turn up for. Obviously no airline would want every passenger to turn up on an overbooked flight.
I don't blame them for overbooking either.
The goal is to fly out with every seat sold & occupied.
They don't want to pay compensation, and they don't want to fly with empty seats.
I suspect UA (and most other airlines) have a pretty good idea which flights they can overbook and to what extent, based on past experience. This will likely vary from route to route, depend on time of day and time of year, etc.
In the real world in which airlines operate, you want to have just a bit of slack in the system. NOT a lot of slack, but just a bit. Enough that you don't create a huge customer service problem (and shunting 5 percent of your paying customers onto other airlines would be a terrible customer service problem, even if the compensation paid is theoretically less than the value of the last seat on the plane). And in that area of slack is exactly where airlines want FF programs to operate.
#12
Join Date: Oct 2009
Programs: UA 1K, Hilton ♦ , Hyatt Carbonado, Wyndham ♦, Marriott PE, "Stinking Bum" elsewhere.
Posts: 5,000
It is a true statement that you want to sell every seat in an ideal world. But in the real world, there are IRROP's (weather and maintenance), there are situations where the overbooking algorithm doesn't make a correct prediction, there are deadheading employees, etc.
In the real world in which airlines operate, you want to have just a bit of slack in the system. NOT a lot of slack, but just a bit. Enough that you don't create a huge customer service problem (and shunting 5 percent of your paying customers onto other airlines would be a terrible customer service problem, even if the compensation paid is theoretically less than the value of the last seat on the plane). And in that area of slack is exactly where airlines want FF programs to operate.
In the real world in which airlines operate, you want to have just a bit of slack in the system. NOT a lot of slack, but just a bit. Enough that you don't create a huge customer service problem (and shunting 5 percent of your paying customers onto other airlines would be a terrible customer service problem, even if the compensation paid is theoretically less than the value of the last seat on the plane). And in that area of slack is exactly where airlines want FF programs to operate.
I am guessing that a 1% IRROPS risk doesn't justify holding seats empty "just in case". My bet is that UA wants to oversell every flight and leave no slack in the system.
#13
FlyerTalk Evangelist
Join Date: Oct 2001
Location: Austin, TX
Posts: 21,413
I believe that I am a pretty typical UA customer. I've flow around 2.8 million bis miles with UA and another 2+ million on other carriers. I did a rough estimate that I've flown on around 2,000 flights. In 2,000 flights, I've had maybe 20 IRROPS situations. That is 1 %.
I am guessing that a 1% IRROPS risk doesn't justify holding seats empty "just in case". My bet is that UA wants to oversell every flight and leave no slack in the system.
I am guessing that a 1% IRROPS risk doesn't justify holding seats empty "just in case". My bet is that UA wants to oversell every flight and leave no slack in the system.
So, the average UA customer has experienced a >1% cancellation rate, and that doesn't even consider the possibility of misconnects, or delays great enough to make you want to find an alternative.
I agree with your basic premise: UA would rather oversell every flight and deal with the consequences -- but if your statistics are correct, you're definitely not typical.
#14
FlyerTalk Evangelist
Join Date: Sep 2003
Location: Honolulu Harbor
Programs: UA 1K
Posts: 15,024
I would guess the passengers booked beyond the plane's capacity are buying some pretty expensive last-minute fares and the cost to offload BE (or other) passengers likely to volunteer for $ works out to mitigate a lot of the $ risk of overbooking.
#15
Join Date: Feb 2015
Programs: united
Posts: 1,636
One other point to make- the "overbook every flight and shunt 5 percent off to other airlines" strategy could only work if OTHER AIRLINES HAVE SLACK. In other words, if that were actually the correct business strategy, but UA's competitors followed it as well, you'd literally end up with 5 percent of the customers being unable to travel at all, not simply being shunted off to other airlines.
Again, how do you prevent this from happening? By having just a bit of slack in your system.