Capacity Management/Revenue Enhancement

[CHANGEATJAMAICA]

Long before there was an Amtrak, back in the dark ages of the '60's airlines had difficulties with flights going out with many empty seats one day and with a half dozen or more oversales standing at the gate watching to plane taxi away the next. Reservations were figuratively index cards in a shoe box.
When airlines started to employ computers for their reservations, one stood head and shoulders above the rest: American's Sabre System. It was the first small step in using computers not only to keep track of reservations but to maximize revenue while at the same time reducing the incidence of oversales, or to use the buzzword of the era: Compensatory Bookings. My employer started out with "George" but wound up with PARS. George would have been superior to Sabre IF it had worked. It didn't and PARS wasn't.
The airlines have come a long way since the original Sabre and Amtrak has come right along with them. From my vantage point in front of the CRT Amtrak has one of the best computer reservations programs or more to the point Capacity Management and Revenue Enhancement systems. One need only read through numerous postings on this and other fora complaining how expensive it is to travel from point N to point C, while at the same time there is a bargain to be had by traveling from point B to point C, roughly the equivalent time and mileage. Why the difference? The fiscal montra: Supply and Demand.
Amtrak programmers know what legs are sure sells and keep the price high. They also know which segments are dogs and try the carrot approach of reduced fares to firm up sagging bookings.
We can all carp about the cost, but the name of the game is to maximize revenue while keeping as many seats full of warm blook pumping bodies as they can. Three Cheers!
Best regards,
Rodger

[wigwagfan]

I could be completely mistaken but isn't ARROW related to SAABRE?

I know that IBM has a contract with Amtrak and they likely manage ARROW from the software side of things. So "hats off!" to the good folks in Armonk, New York.

[Tadman]

Please correct me if my assumptions are off, but didn't a congressional mandate about 8 years ago prevent Amtrak from true demand pricing? If I recall, Amtrak is no longer allowed to sell space to passengers for less than half price (half of what??? though) on LD trains, in an effort to curb LD losses. Rather short sighted, considering it's marginal income at that point, and the seat will go empty otherwise...

[Noel Weaver]

I believe on the LD and non state supported trains there are up to five different price ranges and in this case there could be
a wide varience in the fares between two points.
For example right here in Florida it is generally considered off season and I suspect the fares are much lower than you
would find for the same trip in the winter months when more people are traveling.
Even before the all reserved trains in the corridor and in New York State services, they could sell tickets at one price for
travel on certain days and on days of heavier travel sell tickets at a higher price and I think that was the case on both lines.
Even back in the days when the railroads operated the passenger service, some of them sold tickets good on certain dates or
on certain trains, it was good business in those days and it is good business today too.
I would still prefer to see the corridor trains in the north east be non reserved with ticket prices simply governed by the time
and day of travel, it would make day trips much more convenient especially for business travelers. Doing this could result in
less manpower needed in some of the ticket offices too especially in New York.
Noel Weaver

[taoyue]

I could be completely mistaken but isn't ARROW related to SAABRE?

I know that IBM has a contract with Amtrak and they likely manage ARROW from the software side of things. So "hats off!" to the good folks in Armonk, New York.

We discussed ARROW and SABRE a while ago at this thread: http://www.railroad.net/forums/viewtopi ... 46&t=39404

Sabre has been moving away from its legacy TPF-based mainframe system. They now have a fresh new codebase that dates back only to 2002, and uses modern technologies (SQL databases, distributed systems, multi-tier architecture).

[jp1822]

To me this topic should also be covering the key question that I don't think ARROW or anyone at Amtrak has full control on:

1) At what point does one need to look at adding extra capacity (i.e. coach, sleeper etc.) to a train due to high demand and quick sellouts of said train?

For example if the train is at 80% capacity (coach train) 2 weeks prior to its departure, is there enough "history" or precedence or some other means to justify adding another coach to that train? If that were the case, one would think that the Ethan Allen Express would be operating with additional coaches nearly every Friday evening - due to its convenient "escape" from the city departure time of 5:45 p.m. on Friday and return late return on Sunday - affording a visit to the "house in the country" on weekends. Many New Yorkers have houses up in the Catskills and utilize Amtrak for travel to get there.

We all know that sleeper capacity is likely fixed on overnight long distance trains, but it would be nice to get to the point where VIA Rail Canada is at for its Canadian - where it can add cars when it feels it is nearing capacity and could fill another sleeper.

2) In terms of perhaps sleeper space, should Amtrak keep a waiting list, so as to advise travellers of space availability if cancellations occur?

Amtrak is pretty consistent - I'll give it that - in that it is not about to operate more than 39 Viewliner sleepers in revenue service, when 50 sleepers are on the roster. However, maybe there are times when the third sleeper line on a Florida train is not needed (due to weather related issues?) and instead an additional sleeper is needed for either the Crescent, Cardinal, or Lake Shore Limited. But I have yet to see Amtrak switch the sleepers around like this. And this is likely due to the "crew base" of sleeper attendants or the fact that one attendant is needed to look after one Viewliner sleeper. However, I've stated before that I think there needs to be some "flexibility" with the sleeping car attendants and the fact that they are assigned only one sleeper. VIA Rail Canada is able to isolate three sleepers to one attendant (that one attendant can go to a control panel and see what "car" has pushed the "attendant button"). It would be interesting if for three Viewliners, two attendants could handle the job for a single overnight. But let's not get ahead of ourselves. First Amtrak has to come up with more equipment to allow such flexibility - or at least be willing to operate more than just 39 out of the 50 Viewliner sleepers.

[David Benton]

Its not just about revenue either . To my mind , amtrak is there to provide rail transport to as many people as possible . If the seats are empty , and theres people wanting to travel , why not sell them the seat cheap . you see the buslines with thier ultra low promontial fares , why not amtrak ???
Tad is correct , there was legislation preventing deep discounting for amtrak due to pressure from bus operators , wether it is still in effect i dont know .

[jp1822]

Its not just about revenue either . To my mind , amtrak is there to provide rail transport to as many people as possible . If the seats are empty , and theres people wanting to travel , why not sell them the seat cheap . you see the buslines with thier ultra low promontial fares , why not amtrak ???
Tad is correct , there was legislation preventing deep discounting for amtrak due to pressure from bus operators , wether it is still in effect i dont know .

Amtrak has to start with a base line though with its ticket price and then build from there as necessary. If a train is routinely selling out with a three car passenger train at price "x" three weeks in advance, then one could almost summarize that 1) the train could use an extra coach and 2) if they did a price promotion on this train to reduce fares for say a week, it could lure even more passengers. But then you would have to figure out if the price promotion (discount) would yield enough passengers for yet another coach.

Overall, in the past few years, I think Amtrak took the NEC prices to an extreme - prices between Washington and NYC (and intermediate points), or Washington and Boston (and intermediate points) etc. Occasionally it would offer a price promotion (discount) for tains north of NYC to Boston - likely because Amtrak couldn't fill the seats! Amtrak has always struggled with pricing and filling the train north of NYC to Boston.

Could Amtrak have lowered the Northeast Regional coach fare slightly, added a coach, and still fill the train consist - especially for trains operting between Washington DC and NYC? This may have won over some new passengers. Likewise, Acela Express prices were getting out of control on the south-end of the corridor. One could ride from Washington to Vermont or NYC to Pittsburgh cheaper than even a Northeast Regional ticket between Washington DC and NYC at one point. Yes, it was all about economics - supply and demand. But again, I think Amtrak took it a "little too far" in terms of the pricing, and now they find themselves offering some pretty descent discounts in these economic times. So the revenue was at a peak and then sort of fell off on a cliff, if one were to look at revenue and likely ridership on the Northeast corridor over the past five years. Coming from someone "who's been there and done that" that's rough to explain sometimes in a board meeting, but I am not too sure how "in tune" Amtrak's board is to various issues.

[jobtraklite]

A neat website http://www.chuljin.net/amtrakpricematrix/ gives the coach price between two city pairs for each day up to 14 days in a row. Looking at QCY to CHI for Aug 1 to 14, it shows prices of 34, 46, 68, and 84 for the IZ, and 26, 34, and 46 for the CS (not Coast Starlight), depending on the day ($26 is low bucket).

I can see why Amtrak would charge more when the Cubs are at home, especially when it's Cubs vs. Cardinals, and the IZ is often sold out. But I wonder how many people seeing the $84 price would say "No way!" or accept it just once; and not knowing about the "bucket system" would never give Amtrak a look again.

Furthermore, Amtrak has an advantage here that it doesn't enjoy on very many routes. The running time QCY - CHI is 4 hours 25 minutes; whereas Google maps has driving time at over 5 hours, not to mention the hassles and cost of driving.

[gokeefe]

Amtrak's consistent record setting ridership, which appears to have only "paused" for a moment of interruption by natural disasters over the past three months, brings up some interesting questions.

1. Are they now in a position to further raise fares as spare capacity in their system is soaked up?

2. On a much broader note is Amtrak so heavily used now that they are going to be able to charge fares that provide for 100%+ farebox recovery on some off-NEC trains?

[mtuandrew]

Amtrak's consistent record setting ridership, which appears to have only "paused" for a moment of interruption by natural disasters over the past three months, brings up some interesting questions.

1. Are they now in a position to further raise fares as spare capacity in their system is soaked up?

2. On a much broader note is Amtrak so heavily used now that they are going to be able to charge fares that provide for 100%+ farebox recovery on some off-NEC trains?

I can't answer either, but I would like to expand on your second question:

3. Is there a point where Amtrak can recover 100% of its operating costs on its LD trains, within the 25 car limit per train? Presumably the expenses don't keep rising faster than the revenues, but it might take a while to catch them.

[jstolberg]

I can't answer either, but I would like to expand on your second question:

3. Is there a point where Amtrak can recover 100% of its operating costs on its LD trains, within the 25 car limit per train? Presumably the expenses don't keep rising faster than the revenues, but it might take a while to catch them.

That's a good question, but personally I'd like to see a limit of about 10 revenue cars on a permanent basis (OK to go longer on weekends or seasonally). I think once a train is ready to add an 11th revenue car, the train should be split into two trains with 7 revenue cars. With the additional frequency you reduce your fixed costs per train and attract more riders.

Of course, I don't have access to the real cost figures and don't know if that makes sense from an economic point of view.

[gokeefe]

John,

Why 10 cars? Is there some kind of data-based reason for this "break" point?

As far as question #3 goes I think at least part of the answer may lie in the sleeper/coach mix makeup of the consist (or for that matter the accommodations distribution for ridership).

[jstolberg]

Mr. O'Keefe,

I don't have access to the kind of data necessary to back up my break point.

As the train grows longer there are economies of scale, but there are also diseconomies of scale. One noticeable diseconomy is station dwell time. A 15-car train will have to stop twice at most stations, which can really slow things down. But my break point is probably tied to the biggest revenue loser of the consist, the dining car. At some length, a full diner can't feed all the passengers willing to pay for a meal, so a second diner would need to be added. I think that at that point, it's better to add a second train.

If the host railroad is agreeable, (and that's a big IF), having two trips per day is much more convenient than one. Ridership tends to increase by almost 50% based on that second frequency alone. Money doesn't have to be spent increasing the platform lengths, and the cost of maintaining the stations only increases slightly, but is now divided among two trains (4 stops per day).

Of course, the second train may not need to travel the whole length of the route, or necessarily follow the exact same route. A second train on the Southwest Chief might divert to Amarillo on the Transcon between Kansas City and Albuquerque. A second train on the Crescent might only run from New York to Atlanta. The second train needs to follow the same route for the peak leg, but may diverge thereafter. That type of thing already happens with the Silver Meteor, Silver Star and Palmetto.

Keep in mind that on a typical long distance train the average coach passenger is on the train for about 10 hours. If the train only runs once a day and he's ready to leave right now, his average wait time is 12 hours. That means it will be an average of 22 hours before he arrives at his destination. If the train runs twice a day, his effective travel time is 10 hours on the train plus an average wait time of 6 hours. By giving the passenger choices, he arrives an average of 6 hours earlier.

[gokeefe]

Mr. O'Keefe,

I don't have access to the kind of data necessary to back up my break point.

As the train grows longer there are economies of scale, but there are also diseconomies of scale. One noticeable diseconomy is station dwell time. A 15-car train will have to stop twice at most stations, which can really slow things down. But my break point is probably tied to the biggest revenue loser of the consist, the dining car. At some length, a full diner can't feed all the passengers willing to pay for a meal, so a second diner would need to be added. I think that at that point, it's better to add a second train.

I forgot to even think about platform length. Here's something else to keep in mind: discussions regarding diner/foodservice indicate that on many Long Distance trains Amtrak is just fine with having the diner maxed out and leaving the overflow to the cafe car. Also worth keeping in mind that on the Lakeshore Limited the split is almost even between diners from coach and diners from the sleeper cars.

I think its worth remembering that there really is a bottom line here. Amtrak only has so much capacity available to sell. In some markets the service is in fact underutilized but in many others it's not. Unlike the airlines Amtrak can't just bring train cars back into service from long term storage in the desert. All of this new ridership is in fact absorbing finite capacity which will eventually start to run out on some routes. Assuming Amtrak continues to adjust fares in accordance with demand I can't imagine we wouldn't see 100%+ farebox recovery sooner or later on some off-NEC routes.