Hub Scheduling Comparison

[ELP_WN_Psgr]

Since you are so good, and seem to have som much fun doing this (perhaps you are susceptible to shameless flattery) could you show WN at BWI, LAS, PHX, MDW and HOU at :15 intervals?????

 

[BoeingBoy]

mweiss,

"but not how long you sat on the ground"

Don't really know how to answer that. I know about the trips I fly, but don't know if they're representative of all 737 trips, much less of most of the system trips an all airplanes.

Obviously, the international and long-haul are the best in this respect - get on an airplane, fly a long way, get off the airplane, go to the hotel.

I'm assuming that the 737 is the worst, with shorter legs (longest is 3 to 3-1/2 hours or so), and the aircraft has the lowest utilization of any of the fleet types.

In May, three trips had 11:55 gnd time for 21:15 sked flying, and one trip had 9:13 gnd for 20:17 sked flying. Average gnd time for the first three was 1:11 and for the last was 1:09.

In June, one trip had 10:01 gnd for 22:51 sked flying, and three trips had 10:35 gnd for 19:30 sked flying. Average gnd time for the first was 1:11 and the last three was 1:10.

In July, one trip had 7:35 gnd for 15:40 flying (this was a 3-day), and three trips had 9:05 gnd for 19:52 flying. Average gnd time for the first trip was 1:05 and for the last three was 1:08.

In August, one trip has 12:14 gnd for 22:19 sked flying, and three trips have 8:42 gnd for 20:45 sked flying. Average gnd time for the first trip is 1:12 and for the last three is 0:57.

In September, one trip has 10:55 gnd for 23:05 sked flying and 1:06 average gnd time, and three trips have 9:14 gnd for 20:19 sked flying and 1:19 average gnd time.

According to the bidding program I use, for June the average trip on the 737 in CLT called for sitting 0:53 for each hour flown, in July it was 0:51, for August it is 0:52, and for September it is 0:52.

Note that the above does not include the duty time before the first departure or after the last arrival of the day. This is sitting between flights during the day.

Hope this helps.

Jim

 

[BoeingBoy]

ELP_WN_Psgr,

The LAS, PHX, and BWI are fairly easy. I didn't look at MDW or HOU so those would require going back to the base data and starting from scratch (picked LAS and PHX since they seemed to have more flights than other WN cities, and BWI since it's just down the road from PHL).

I'll put them up when done.

Jim

ps - for everyone, these are just mainline flights.

 

[BoeingBoy]

ELP_WN_Psgr,

Here's LAS...

Jim

ps - don't worry. Between my wife and the f/a's, I'll never get a swelled head...

 

[WorldTraveler]

Very nice work, BoeingBoy.

Two other ingredients that are necessary for a depeaked hub to work is for the hub to be in a large metro area with lots of connecting traffic and for the hub to be very large. When hubs are depeaked, connectivity is decreased. The way to offset the loss of connectivity is to take more local passengers - which incidentally are more profitable for an airline in most cases. A hub also has to be very large or you lose more connectivity than you gain in operational efficiency. that's why depeaking ORD and DFW were no-brainers. PHL certainly is a good candidate but there are not a whole lot of hubs in the US that are good candidates for depeaking. Given that CLT has a fairly small local population relative to the amount of service US provides, that hub needs to operate very efficiently and with good connections in order to compete against hubs like ATL that offer far more connecting possibilities. If you lose connectivity, then the reason for serving a number of the small cities that US and its regional partners serve is lessened. If US is going to serve those small cities, they need one well-structured hub in order to compete for connecting passengers through that hub as long as any other carrier serving that city also has a highly structured (vs. depeaked) hub.
Given the ATC contraints and the large local population, PHL makes sense to depeak but CLT does not.

 

[mweiss]

WorldTraveler,

I agree that conventional wisdom would dictate that PHL gets depeaked and CLT does not. However, the calculations are more complex than you suggest.

If one can provide the service at a lower cost, as would result from depeaking, the O&D prices could profitably be lowered, which would stimulate demand, offsetting the loss in demand associated with the reduced connectivity of the hub.

One thing to keep in mind is that a connecting passenger costs much more than a nonstop passenger. US would be much better off flying one passenger from BOS to CLT and another passenger from CLT to MIA than flying the single passenger from BOS to MIA via CLT. US would get close to twice the revenue, and move precisely the same number of ASMs.

 

[BoeingBoy]

And here's PDX....

Jim

 

[BoeingBoy]

Finally, WN at BWI...

Jim

 

[BoeingBoy]

WorldTraveler,

I'm with you to a certain extent, but mweiss has valid points also. Somewhere between the most efficient depeaked hub (an absolutely even flow of flights throughout the day) and a traditional hub (see CLT) there is a point where both efficiency and connectiveness are maximized to the extent possible.

It's easy for me to sit here and "talk" about the broad-brush aspects of depeaking a hub - that's not rocket science. However, actually planning and implementing that process is very much akin to rocket science with a strong dose of voodoo thrown in.

In a city like CLT, it is very likely that the 3 flights a day from/to East Stumpwater would have to arrive early in a "bank" and leave late in that "bank", meaning a less efficient operation. That doesn't mean you can't depeak the flights from/to Metropolis with 6-8 frequencies a day and gain that efficiency

Jim.

 

[ELP_WN_Psgr]

Thanks a lot, B.B.

A lot of folks (including from their GO at Love Field) will try to convince people that WN is not a hub and spoke carrier.

I tend to think of them more along the lines of Huey Long. Huey Long was Governor of Louisiana back in the 30s and 40s, as I recall. He often campaigned on a platform of/with a slogan promoting "every man a king."

WN tends to operate more on the principle of every city a hub. If you consider LAS and BWI as significant hubs, then how in the world do you have, in addition to a plethora of nonstops and one-stops, a published connection in Albuquerque (when flying from BWI to LAS)? BTW that one has, if I remember correctly, a layover of exactly 35 minutes-which is not bad graound time under any circumstances.

Depeaking the hubs is long overdue and is absolutely necessary for survival.

Once upon a time SABRE etc etc sorted flights by total elapsed times. Folks didn't look at price.

Now, everyone looks at price. After price, practically everyone looks at schedule, which FF program they can get credits on, and is the layover at a particularly onerous intermediate point. (Hint: If I was routed to hell via DL, I am sure I would have to change at ATL...and would much prefer just taking a nonstop straight to hell).

In the end, efficiency conquers all. Big gaps and peaks and valleys do not make the best and highest use of the personnel and fixed facility resources the airline has.

 

[BoeingBoy]

ELP_WN_Psgr,

I've seen some figures for WN's connecting traffic systemwide (something like 20% IIRC), but none for the cities with more numerous flights. Do you have any info along those lines? Just looking at the graphs for LAS, PHX, and BWI it seems they look more like a depeaked hub than a focus city.

Jim

 

[ELP_WN_Psgr]

I can probably get some....I know a few semi legitimate sources.

The 20% figure is pretty close to everything I have seen, though, and holds true for most of the "big" stations.

 

[WorldTraveler]

MWeiss and BoeingBoy,
Cost is not the only factor. You have to consider the competitive environment for each of those O&Ds you are carrying over CLT and have to be capable of replacing every lost connecting passenger with a local passenger. CLT is a $1.1B annual local market according to DOT data, of which US has about 60% and has commanding positions in most of the top markets EXCEPT to/from other carrier’s hubs; in nearly every O&D that is to/from another carrier’s hub, the other carrier has higher share than US. That says that CLT is a much weaker hub than those of its competitors and cannot afford to lose its connectivity. It is also a fairly small local market – on the order of cities like RDU and CLE, smaller than STL, and a bit bigger than CVG and IND. CVG is the only city of that size that supports a hub as large as US’ CLT operation and it is a highly structured hub. Like CLT, CVG has good weather, relatively few ATC delays, and is an inexpensive place to operate. DL has about 80% of the CVG market.
In contrast, DFW is a $3.5B local market of which AA has 60% while ORD is a $4.5B local market of which AA has 35%. ATL is a $3.7B local market, of which DL controls about 70%.
If CLT is depeaked it will certainly result in the loss of traffic to connecting Delta which has highly structured hubs surrounding CLT and its catchment areas.
My recommendation: use the new RJs to grow the total US network esp. at CLT. The only markets US does not control are hubs to other markets so the domestic growth potential is limited. The Caribbean strategy makes sense but that is dependent on a strong connecting hub, not the local market. US’ Caribbean advantage is that it uniquely serves the smaller Caribbean markets through a true hub. If US depeaks CLT, it comes at the risk of being able to compete in many smaller O&D’s, around which US has built it route system. That may be okay but there will be a lot of cities which will no longer support US service. International connecting O&Ds, including the Caribbean are thin, if only because the traffic usually only flows each way during a couple hour window each day. I believe it would be a mistake for US to mess with the connectivity of the CLT hub.
It is also apparent looking at the hub charts that activity peaks about every two hours. Efficiency could be increased by keeping a hub structure but restructuring the existing schedule and adding another bank or two during the day so that banks occur about every 75-90 minutes, minimizing the dead time between banks.

 

[mweiss]

You have to consider the competitive environment for each of those O&Ds you are carrying over CLT and have to be capable of replacing every lost connecting passenger with a local passenger.

But that's just it. You don't have to replace them all. You only have to replace a little more than half of them.