Originally Posted by
hobo13
Actually, I think you are wrong. The reason that you sell GLSTK fares is that they are typically sold months in advance. That's money in the bank -- a free loan to the airline! It also assures them a 'baseload' for the flight.
Most flights don't even open the lower buckets until a couple months out, not 9-11 months like they used to. The float on that money isn't nearly as lucrative as it used to be.
Originally Posted by
hobo13
Selling full fare Y tickets at the last minute is much more of a statistical process. You will never accurately predict precisely how many Y's you'll sell 3 days before a flight -- just a mean, and a std deviation. So if your model says that you will sell 6 Y fares 3 days out, with a stdev of 3, how many seats should you hold back? 6? 7? 8? 9? And what about when the 10th guy shows up and wants to buy a Y ticket? Do you turn him away because your model didn't predict him? Of course not, you smile, take his $1200 and then go bump some K-fare chap by giving him his E[$400]=$200 voucher!
This suggests that you always sell the Y fare. No matter what. Because the voucher is always going to be less than the full Y fare. But airlines don't do that. They understand that just giving someone a voucher doesn't make them whole as a customer. If you cannot accommodate them then you've got a much bigger problem. And with LFs as high as they are these days that accommodation issue - on any airline - can be quite complicated. So you don't overbook as aggressively on the last flight of the day or even on the flight before one that is seeing higher than expected loads.
Originally Posted by
hobo13
Ever heard of stat arb? It's pretty amazing stuff.
Yes, I have and it is. That's why those models are so useful.