justaskme

This is called yield management

bodory

... of the crew

San Gottardo

Err... no. It isn't. Yield management is not about cost. It is only about revenue. Yield management is to maximise the revenue from a given number of seats, where the "given" is the number of seats physically installed on board. The idea is to modify fares so as not to leave any seat depart unoccupied whilst getting as much revenue as possible for every sold seat.

Not offering existing seats for sale is NOT yield management. Yield management would be to for instance offer some lower fares (and attach restrictions, or limit the sale to certain segments or markets) to sell the remaining 3 seats.

The entire logic of not opening up seats for sale to reduce staffing by one FA reminds me of one Air France executive (I believe it was the former group CFO) to argue that Business Class is a terrible product for airlines because the seats take so much space in the cabins. A purely cost-driven view.

Goldorak

Exactly.

Fully agree

orbitmic

I can only repeat that what justaskme explained is exactly what other AF people explained to me in the past.

My guess is that this works in the following way: let's assume, for the sake of argument, that historic bookings suggest that a given route will indeed only get between 0-5 bookings in a given period and that this corresponds to 1 crew member. I guess that the advantage of basing bookings on a maximum P5 allows both to plan crew affectation accordingly (I understand P crew are specially trained so their pool is limited), to have RM plan F bucket availability on a basis of a cabin size of 5 (otherwise, if you start from the cabin size of 9, the RM algorithm will always perceive the class as underbooked and thus systematically release cheaper F, so that in effect, nobody will pay P), and have enough availability for the existing demand in the immense majority of cases. Should some flights near the full reduced capacity, you can always decide to upgrade the availability to the real maximum (8 or 9). The very fact that AF is moving to a standard Premiere cabin size of 4 suggests that this 5 maximum effectively corresponds to maximum needs in the vast majority of cases.

I understand the surprise of people asking why not start from 8/9 from the start and if it is not needed it won't be taken anyway, but this presents a number of disadvantages:
- Of course seats can be booked till the last minute. If you start from 8/9 capacity you need to have the corresponding crew ready to operate the cabin till the very last minute as you can't assume, 2 hours before the flight when you expect 3 people onboard that you won't get someone snatching another 3 seats within the hour;
- This is made worse by the obvious fact that crews need to be available not only for outbound but for return flights! So even if at the time your CDG-LAX sales are closed you are still on an occupancy of P1, you will still need two crew to fly the route anyway because you cannot assume that the LAX-CDG on which the same crew goes two days later and which is also showing P1 occupancy won't go beyond 5 passengers within 48 hours!

I think that the criticism of AF is a bit too harsh on that one. To me, assuming that the calculation of expected occupancy is correct, the policy actually makes sense. This is simply the same as some airlines knowing in advance that they will be flying a 4 class aircraft on a usual 3 class route and deciding not to sell F because they do not think that demand is there and that therefore it is not worth engaging the costs of crewing and catering the F cabin effectively knowing that you are very unlikely to make it pay for itself. Here, you can think of those extra 4 seats as a separate F cabin and AF thinking that it simply won't pay for itself to open it for sales.

AF has long made the choice of not offering cheaper rates in P with an aim of filling cabins. We may regret this choice (I certainly do) but it is theirs. In that context, it simply miscalculated P capacity in its current fleet, offering too many P seats considering its fares (high) and product quality (low onboard on the old P) which means that it effectively has a very low probability of filling 9 seats at the prices it wants. If it can't it might as well cut the loss and assume a virtual lower capacity.

bodory

Bottom line : this process may be appropriate in itself, it is consistent with a misunderstood P strategy.

And I was then wrong in my initial assumption :

Thanks justaskme for clarifying.

stimpy

I think if we discussing this in the same room it would be more clear to each other. From my side this sounds normal And as orbitmic wrote above other airlines take similar measures to manage their inventory and resources.

One question I have is, doesn't the AF reservation system allow overbooking in P like they do in Y? In that case AF is not actually restricting sales, right? And they can react if 8 people book P for a given flight.

orbitmic

No, AF is one of the few airline which does strictly no overbooking in P. It's a choice they made long ago and have stayed with since.

SFO777

I'm not aware of any airline that overbooks F.

San Gottardo

Wirelessly posted (Mozilla/5.0 (iPhone; CPU iPhone OS 8_1_2 like Mac OS X) AppleWebKit/600.1.4 (KHTML, like Gecko) Version/8.0 Mobile/12B440 Safari/600.1.4)

Lufthansa has started to overbook F on certain flights up until 4 weeks before departure. They say that their forecasting models are good enough to manage that. No idea if it's true, but so far I haven't heard or read about situations where they had to refuse a passenger.

But AF seems to have a problem with filling its seats so overbooking isn't really an issue for the vast majority of flights.

justaskme

It s not as simple as you think. You forget that Yield management works together with the flight plan in order to choose the right aircraft with the right configuration ( cabin and cargo ) to a specific route on a specific date taking into account operational costs ( which do include crew expenses ) , historic bookings, customer trends, other airlines flight plans, political situation , calendar issues etc.
if you take a look at airline industry you will see that F demand is down and almost all airlines keep reducing the seats.
On certain markets AF doesn't see the need to put 8 or 9 seats because the demand is down.

orbitmic

BA treat F as any other cabin and over-reserve accordingly. QR does exactly the same and in both cases there have been reports of people being downgraded as a result. As San Gottardo mentioned already, LH does the same. I would be surprised if UA and AA did not also treat F as any other cabin but I don't have direct knowledge about them so if you tell me that they strictly don't I am very happy to take your word for it.

Still, my understanding is that not overbooking a given cabin is very much the exception rather than the rule.

Junior_2J&K

It's AF990 on May 17. So in fact, while the P line could not get me into 3L, the AF Twitter folks could ! So I'm now happily back into 3L.

Anyway, sorry for launching such a debate ! Though it is quite an interesting topic. Thanks justaskme for your insights. Do you work for AF by any chance?

San Gottardo

Err... did I say it was simple?

I fully agree that all of that has to happen.

My point is only: what you describe is not called yield management. Yield management only looks to maximise the revenue of the seats on the plane that is taking off.

What you describe is more like network planning with elements of fleet planning, where indeed all of the things like operating cost, operating constraints, demand drivers, etc. But not yield management.

Take an A320 from Paris to Madrid: playing around with the number of seats sold in the various booking classes (J, C, Y, M, B, K, etc) - that is yield management. Putting an A321 instead of the A320 because it's the first day of vacation is not yield management.

Similarly, not selling seats in "F" class because they can probably be sold in the more expensive "P" class is yield management. Not selling seats on a plane to save the costs of an FA is not yield management. Yield management takes cost as a given.

Goldorak

I was not criticizing AF and I certainly thank Justaskme for providing us some explanations and information (please continue to do so ). But I was trying to understand this practice that was very obscure to me. I have the feeling I understand a bit better after your post Orbitmic (so thank you for that ) but I'm still not 100% convinced (but it doesn't matter at all !)
I'm thinking also about another potential problem with this practice. One of the benefit to fly with a P ticket is the flexibility to change the date of your flight(s) including at the last minute. But if the inventory is at 0, you might not be able to change your flight for the new date you want while there are still some seats available in the plane. Or does AF release seats "automatically" in this case ?