Passenger Numbers

English Boards Community Support
1

I have a question regarding passenger numbers.  Let's say I have 4 round trip flights scheduled from A - B on a particular day and I notice that only 6 passengers per day are booking the outbound flight and 5 passengers per day are booking the return flight.  So the booking look something like this:

Day 1 booking          Day 2 booking          Day 3 booking

A - B  6                     A - B  12                   A - B  18

B - A  5                     B - A  10                   B - A  15

A - B  6                     A - B  12                   A - B  18

B - A  5                     B - A  10                   B - A  15

A - B  6                     A - B  12                   A - B  18

B - A  5                     B - A  10                   B - A  15

A - B  6                     A - B  12                   A - B  18

B - A  5                     B - A  10                   B - A  15

So I think, ok obviously not enough demand to support 4 round trips per day based on my ORS rating (which was at the top but I know I was losing some bookings to another airline with connections).  So I cancelled 2 round trips from flights that only had one day of booking so far thinking that the remaining passengers who would have booked the now cancelled flight will book the existing flights thereby doubling the booking.  Here is what I did:

Day 1 booking                Day 2 booking that was expected               Day 2 booking actual

A - B  6                           A - B  18                                                      A - B  12

B - A  5                           B - A  15                                                      B - A  10

A - B  6                           A - B  18                                                      A - B  12

B - A  5                           B - A  15                                                      B - A  10

A - B  6 (cancelled)

B - A  5 (cancelled)

A - B  6 (cancelled)

B - A  5 (cancelled)

So if my representation is clear then I think you understand where I am going with this.  What happened to the 6 A - B passengers and 5 B - A passengers who would have booked on one of the 2 cancelled flights?  I assumed they would book one of the remaining flights but it seems they didn't.  If the demand was for 6 A - B passengers and 5 B - A passengers in one day for 4 flights, why would the demand not be 12 A - B passengers and 10 B - A passengers in one day for 2 flights?

I know I must be missing something.

2

You have proportionally smaller share of flights in the ORS system with 2 flights instead of 4, so given your rating you got a proportionally less share factor of the demand on that route.

Additionally some pax may have been connection bound with connections no longer available with canceled flights.

3

You have proportionally smaller share of flights in the ORS system with 2 flights instead of 4, so given your rating you got a proportionally less share factor of the demand on that route.

Additionally some pax may have been connection bound with connections no longer available with canceled flights.

Thanks for your reply.  I clearly do not understand something about how bookings are handled.  I didn't know that the proportion of flights you have compared to the whole made any difference.  I thought it was ratings and any passenger required connections.

So that means that even if I had all 4 flights leaving at the same time (meaning connections would not be disrupted) the extra passengers would still not take one of the remaining 2 flights because the proportion of my flights was less?  How then does rating play into that scenario?

Sorry that I am having difficulty wrapping my head around it.

4

Awnings matter but it’s not a winner takes all model. Top ratings get a proportion of all bookings which are then equally distributed among all flights with the same rating. A lower level rating gets another proportion of bookings plus any overage not filled on bookings from the top rating, which then again is equally distributed among all flights on that level, and so on.

And yes it may take a year or more to somewhat master the logic and workings of the distribution.

5

Awnings matter but it's not a winner takes all model. Top ratings get a proportion of all bookings which are then equally distributed among all flights with the same rating. A lower level rating gets another proportion of bookings plus any overage not filled on bookings from the top rating, which then again is equally distributed among all flights on that level, and so on.

I think I understand.  So there is a *percentage* allocation to the top rated flights, all else being equal, which presumably would equate to the proportion of flights that airline has to the whole.  That's very different than I was thinking.  I was thinking more of the winner take all model you mentioned.  I thought that, again all else being equal, if 200 passengers want to travel direct A - B they will all book the highest rated flight until it is full before the bookings move to the next one.

Makes it a bit more difficult to understand how changes you make to your flight profile will actually equate in terms of bookings when you can't say for sure that removing a flight will increase the load on an equivalent flight.

Doesn't that also lead to a situation where you would get more passengers by having multiple smaller planes rather than fewer larger ones?

6

Doesn't that also lead to a situation where you would get more passengers by having multiple smaller planes rather than fewer larger ones?

Yep! It's definitely one of the identified problems and one of the major motivations behind the new demand allocation system (which I don't think we've heard any news on for a few weeks...)

To help visualise the process, try thinking about it like this:

You have two airports, A and Z; and two hub airports, L and M. 100 passengers want to travel between A and Z per day

Passengers can make the following connections between them:

(1) A - Z (direct, ORS rating 99)

(2) A - Z (direct, ORS rating 92)

(3) A - L - Z (short transfer, rating 70)

(4) A - L - Z (shortish transfer, rating 65)

(5) A - L - Z (moderate transfer, rating 60)

(6) A - L - Z (longish transfer, rating 55)

(7) A - L - Z (long transfer, rating 50)

(8) A - L - M - Z (two transfers, rating 10)

Obviously (1) is going to be the most attractive. It won't get all the traffic, but it should get a large proportion. So, you might get traffic distributed like this (just as a general hypothetical type statement - real life results may vary!) :

(1) - 45

(2) - 22

(3) - 10

(4) - 7

(5) - 6

(6) - 5

(7) - 4

(8) - 1

Looking at that distribution, you can see some big problems:

- Demand between the two points isn't enough to fill a very large aircraft; so, if you want to grow your market share, you might fly a small one at high frequency (expensive, not very profitable), or a larger one at low frequency (cheaper, more profitable - but only so long as you don't run into trouble because...);

- Your market share is subject to changes from factors outside your control; things your competitors do will naturally have an effect on your loads. Let's reimagine the distribution above as this:

(1) A - Z (direct, ORS rating 99)

(2) A - Z (direct, ORS rating 92)

(3) A - Z (direct, ORS rating 92)

(4) A - Z (direct, ORS rating 92)

(5) A - L - Z (short transfer, rating 70)

(6) A - M - Z (short transfer, rating 70)

(7) A - L - Z (shortish transfer, rating 65)

(8) A - L - Z (moderate transfer, rating 60)

(9) A - M - Z (moderate transfer, rating 60)

(10) A - L - Z (longish transfer, rating 55)

(11) A - L - Z (long transfer, rating 50)

(12) A - M - Z (long transfer, rating 50)

(13) A - L - M - Z (two transfers, rating 10)

Given that, you might get traffic distributed like this:

(1) - 30 

(2) - 13

(3) - 13

(4) - 13

(5) - 6

(6) - 5

(7) - 5

(8) - 4

(9) - 3

(10) - 3

(11) - 2

(12) - 2

(13) - 1

Again, obviously very hypothetical - but you can see how everyone's hurting a bit more. 

To complicate it even further, you might consider some other things:

- The dude running A - L - Z, A - M -Z etc doesn't necessarily care about the proportion of A - Z traffic he gets; he's feeding off (say) A - L - B through A - L - Y, too, and if you're not flying those routes directly that could amount to an awful lot of people;

- Very long connections, or connections where one or several flight segments individually get low ratings (such that the route as a whole has an ORS rating of -1 or below), may mean that people would just prefer to use ground transport instead; those passengers might as well not exist, right up to the point where someone provides them with a better option. There's one server where ground transport is turned off, but I don't know if that means those people just don't go anywhere instead.

- Some flights will be full; the people who can't get themselves on those flights will then necessarily travel some other way. This can mean that your flights do well despite a terrible rating; it can also mean that you get, for example, no bookings one day and tons another, if your competitors are only flying, or only making particular connections available, some of the time.

- Economic variation (the AGEX thingy) can mean that one week your pool is 80 people, and the next week it's 120 (though the change is usually fairly gradual).

Hope that's enough to get you thinking! I think in practice direct flights tend to get a larger proportion of passengers than that, presuming they're available, but since most people will be flying to and from hubs most of the time trying to exactly match demand to seats is a tremendous amount of work, so you get spillover here and spillover there and everything gets very opaque.

7

And I would add that if there are 5 flights at A (99) rating, those would get equal proportion of those 45/30 passengers, so each would get 9/6 passengers respectively. But then there EIS yet another factor that if there are many more 99 rating flights that rating level will get a higher rproportion of pax overall, so you may see 75/45 instead of 45/30.

8

To finally answer your question:

If you cancel a flight which already has, say 6 passengers booked on it, these 6 passengers will not automatically be added to your next flight (as a re-booking). They will "disappear" in the system, e.g. not travel at all.

9

@ijv - Thanks for that!  Very informative and definitely some points to think about.

@rubiohiguey2000 - Understood.

@voeni - Thanks but that actually wasn't my question. :D I do know that the cancelled passengers disappear.  What I was asking is that on the following day's booking I would have expected to get the NEW booking of 6 passengers for both the active flight and the NEW booking of 6 passengers that would have booked on the now cancelled flight.  That's why I was expecting 12 new bookings on day 2 (for a total of 18) instead of 6 (for a total 12).  If the passengers on the cancelled flight still existed and rebooked then I would have expected 18 new bookings on day 2 for a total of 24.

10

I'll take the blame  :D