Jamincan From Canada, joined Aug 2006, 792 posts, RR: 0 Posted (7 years 3 months 3 weeks 5 days 9 hours ago) and read 3366 times:
I was thinking about how one could objectively measure the importance of an airlines hub for an overall airline network, and decided, as a starting point, individual stations in a route network could be compared by calculating the ratio of seats per week at that particular station to seats per week over the entire network. For example, if the airline has two routes, A-B and A-C, both flown with a 100 seat aircraft daily, A would be 1.0 (1400/1400) and B and C would be 0.5 (700/1400). You could say that A would be twice as important to the network as B or C. This statistic would basically measure the proportion of service each station captures of an airlines overall service.
Just for reference, looking at Porter Airlines, the percentages come out as follows:
Obviously Porter is heavily centralized in Toronto, with YOW and YUL being two important stations.
It's fairly easy to look at Porter's schedule and calculate this, but much larger airlines would obviously require far more effort. Is it possible to find readily available statistics that indicate the number of seats an airline has to a particular station in a year, and the overall number of seats an airline flies?
PanAm747 From United States of America, joined Feb 2004, 4242 posts, RR: 8
Reply 2, posted (7 years 3 months 3 weeks 5 days 8 hours ago) and read 3293 times:
I don't think available seats indicates the importance or profitability of a route by itself. One must take into account the average amount being paid for by these seats, as well as the competition on the route from other airlines that might keep fares artificially low.
I kind of see where your question is going, but I don't think your model will be sufficient to explain all the variables in explaining routes. For example, MIA-PAP for AA is almost a license to print money, but MIA-JFK has to be competitively priced, even if both airports have (relatively) the same number of seats on AA.
Pan Am:The World's Most Experienced Airline - P(oor) S(ailor's) A(irline): San Diego's Hometown Airline-Catch Our Smile!
Rwessel From United States of America, joined Jan 2007, 2740 posts, RR: 2
Reply 3, posted (7 years 3 months 3 weeks 5 days 2 hours ago) and read 3149 times:
I'm not sure I really like the number you're computing. The value of a hub really is in the number of connections it allows to be made. All 700 A-B and all 700 A-C could well be O&D flights, in which case A is not a hub at all. A's value as a hub is in any B-C flights it enables.
For the sake of discussion, assume that 300 of the 1400 seats moving through A are B-C traffic. The total value *supported* by A would be 400*value_AB+400*value_AC+300*value_BC.
OTOH, then the total value *contributed* by A would be more like 400*value_AB/2+400*value_AC/2. Note that I'm leaving half the value of the A-B and A-C routes to the B and C stations. And note that I'm not assigning A as contributing any value to the B-C route (although it does support it). That's not different than the airplane itself - it doesn't contribute one bit to your income, but you better have one if you want to call yourself an airline.
I think the distinction is important - consider what happens if D is another hub, which could, conceivably support all the B-C traffic if you closed A, but none of the A-B or A-C routes.
Ultimately the value of a hub is the concentration of traffic it creates which may make service to a particular destination practical or more profitable. IOW, the value of a hub is how much in increases revenue (or income, depending on what we’re measuring), relative to *not* having it.
Jamincan From Canada, joined Aug 2006, 792 posts, RR: 0
Reply 4, posted (7 years 3 months 3 weeks 4 days 16 hours ago) and read 3017 times:
You guys make some good points. Factors such as yield and connections are certainly important in determining the relative importance of individual stations within an airlines network; however, that data is far more difficult to get.