Bloomfield Avenue case study: Market segmentation for buses is bad

Market segmentation, in its broadest definition, is dividing the market into subsets of consumers who can be provided with individually tailored services.  This works for clothes.  This works for restaurants. But it does not work for buses.

I’ll illustrate the problem. Take Bloomfield Avenue in Essex County-  it’s one of the area’s main roads, nearly a straight line between Caldwell, Montclair, Bloomfield, and Newark.

Currently, the Caldwell-Bloomfield stretch of Bloomfield Avenue is shared between the NJ Transit 29 and the Decamp 33.  NJ Transit takes local passengers, and Decamp takes only passengers to New York.

Caldwell is considering starting a commuter shuttle service that would take Bloomfield Avenue between Caldwell and Bay Street Station in Montclair. That means there would be 3 separate service on Bloomfield Avenue for 3 separate market segments:

1. NJ Transit buses for local trips between Caldwell,  Newark, and everywhere in between
2. Decamp buses for passengers heading to the Port Authority
3. Caldwell shuttle for train commuters to New York

This is a problem because frequency is one of the most important elements of a good transit service.  3 separate buses along Bloomfield Avenue will mean that riders can only use a third of the buses to get from, say, Verona to Montclair. The local rider loses out in this situation.

In a classic market segmentation situation, the market segments each get a product that suits their needs.  But for transit, part of quality is quantity.  Having local riders to fewer buses means that the local bus service between points on Bloomfield Avenue is qualitatively and quantitatively worse.

Riders going to New York on Decamp have the most to gain from this arrangement. Since the Decamp Bus isn’t picking up and dropping off local riders, the speed the New York is marginally faster.  The local bus isn’t much use to them anyway.

The train commuters from Caldwell will still have a broader range of options than the other two market segments.  They still have the choice between the local NJ Transit bus and the Caldwell shuttle to take up Bloomfield Avenue.  If one is late, they can take the other.

The real losers in this situation are the parties who are paying for bus service.  By allowing local riders onto other bus services, there could be more transit options, but provided by fewer buses. Essentially, more money is being spend on bus service than needed. If Decamp dropped off at Bay Street Station, the Township of Caldwell might not need to pay for a municipal shuttle to the train station.

Now, let’s acknowledge that there’s are reasons things are this way. NJ Transit local bus service is not that good.  Even at rush hours, you can wait 15 minutes or more for a local bus down Bloomfield Avenue. They can be slow and not always on time.  Clearly, for whatever confluence of reasons, the 29 is an inadequate means for Caldwell commuters to reach Bay Street station.

Decamp also has its reasons for not taking local passengers. Decamp might not be interested in the low fares that come from local rides,  but the primary reason is that the company is actually prohibited from doing so. Its franchise restricts its buses from competing with NJ Transit. But this is an outdated way of thinking, a relic of the time when local suburban transit service were a profitable business. But if NJ Transit dropped this prohibition, it might mean better service for its passengers. Alternatively, local service on Decamp might mean that NJT could scale back its services with no ill effect.

In times of tight budgets and little expansion to the transit system, we should focus on making the best use out of the transit we already have. This calls for a little creativity from transit providers which these days, is sadly lacking.

Dwell times- why your train is slow

In theory, commuter trains can go up to 90mph, but most of the time they stop every few minutes and never reach this speed. The amount of time a train spends at each stop isn’t much, but it adds up.

This is called “dwell time,” the total time that the train stops, or “dwells” in at a station.

Here’s some data sent to me by a friend. He recorded how long his train spent at each stop, from Upper Montclair to Newark.  The timer starts when the train comes to a full stop, and ends when the train begins moving again.

According to the Montclair Boonton-Line schedule, The trip from Upper Montclair to Newark Broad Street takes anywhere from 24 to 27 minutes. If we subtract the dwell time for Upper Montclair and Newark, The train is not moving at all for 9 minutes 33 seconds inbound and 8 minutes 45 seconds outbound. That’s anywhere from 32% to 37% of the scheduled trip time from Upper Montclair to Newark.

Now granted, the data here is from rush hour trains and doesn’t include the section of the line between New York and Newark with fewer stops, so it’s not entirely representative.

Here’s the same information in graphical format:

You’ll notice that the stops that took the least time were Upper Montclair, Bay Street, and Watsessing. Watsessing and Upper Montclair took less time because they have lower ridership.  In Fall 2013 (the most recent available stats), Watsessing has 223 daily boardings and Upper Montclair had 519.

Bay Street, on the other hand, had 1,166 boardings, making it the busiest stop on the line.  This is interesting.  Bay Street is the busiest stop on the line yet it has one of the lowest dwell times.

Where will the dual-modes go next?

NJ Transit owns 35 dual-mode locomotives that can operate with or without overhead electrical power.  Most importantly, these locomotives can provide a one-seat ride from non-electrified lines through the tunnels to Penn Station. Originally, they were ordered to provide service to the now-cancelled ARC tunnel.

The first dual-mode revenue train was operated in May 2012, but since then, the roll-out of new one-seat-ride services with the dual-modes has been slow.  In March 2014, the first Raritan Valley Line trains to New York Penn Station began running at off-peak hours.  It took almost two years for NJ Transit to use the dual-modes for their intended purpose – to eliminate transfers and provide direct service to Penn Station.

In May 2015, NJT took another step forward with the dual-modes, inaugurating 3 direct New York to Bay Head trains in each direction, eliminating a transfer at Long Branch.  This was the first time NJT operated regular through trains from New York to the lower Coast Line (not counting summer beach express trains that ran the year before).

Except for these two particular instances, the dual-mode locomotives are being wasted.  They cost about $10 million each. For comparison, a diesel locomotive without electric capabilities costs $5 million.  Yet many of them are circulating through the train equipment pool as regular, diesel-only locomotives.  The total locomotive order was for some $408 million, yet their incredibly expensive dual-mode abilities are only being used by 23 trains per day.

Now, granted, there is another advantage to operating dual-mode locomotives other than providing new one-seat-ride service. Some of these locomotives are operating from Hoboken on the Morris & Essex and Montclair-Boonton Lines.  They can run most of the way in electric mode, and then change to diesel mode after the wires end. For instance, a train to Lake Hopatcong can operate under electric power to Dover, then run the last few miles using diesel power.  Electric traction is quieter, generates less pollution and provides faster acceleration.  A dual-mode locomotive can run a train like this faster than a standard diesel locomotive could.  But unless NJ Transit commits to consistently operating those trains with dual-modes, schedules can’t be updated to reflect faster running times.  Right now, these trains are holding for time at intermediate stations to keep on schedule.

There are opportunities to use the dual-modes to their full potential elsewhere. In addition to running direct Raritan Valley and Coast Line trains at more times of day, NJT could use the dual-modes to expand Midtown Direct service to the outer Montclair-Boonton Line. Currently, Midtown Direct service is only offered east of Montclair State University, and at Denville and Dover via the Morristown Line.  It would be possible to operate limited rush hour service from Lake Hopatcong simply by extending a Dover-New York train to Lake Hopatcong.  A Lake Hopatcong-Hoboken train could be moved to start at Dover, so the change could essentially be done at no cost.  A direct train over the outer Montclair-Boonton Line to New York could be similarly done by swapping its slot with a Hoboken train.  The main question is whether enough seating capacity exists on the trains in question to accommodate passengers from the new stations.

It’s said that the second mouse gets the cheese.  The same principle, to a less gruesome extent, applies to commissioning new rail equipment. 

Articulated buses- use with caution

You might have seen them.  NJ Transit operates articulated buses on only a few routes. In the Newark Area these are the 13, 39, and 70. Articulated buses are also used on the 154, 156, 158 and 159 in Hudson County.  These buses are 60 feet long and have a bend in the middle.  For comparison, a standard transit bus is 40 feet long.  The main advantage of these buses is their additional seating capacity, which is important on busy routes.

The important question to ask-  how busy is busy enough to justify using articulated buses?  Their main advantage is that they allow NJ Transit to carry more passengers while not running any more buses.  Frequencies don’t have to increase.  But as we are well aware, increased frequency makes transit more convenient and increases ridership.

NJT’s new 40-foot buses have 39 seats, and the articulated buses have 59 seats. It stands to reason that if these routes were operated with shorter buses, the frequency would have to be about 50% higher to achieve the same number of seat-miles per hour.

The 13 runs a 12 minute headway or less from Monday – Saturday and a 15 minute headway on Sundays. Hypothetically, this bus would run every 8 minutes with standard-length buses during the week and every 10 minutes on Sunday – definitely an upgrade.

The case against articulated buses becomes even more apparent when you realize that the 13 has branches on either end.   At the southern end in Irvington, buses on each branch run every 24 minutes during the week, and every 30 minutes on Sunday.  With regular buses, this would be reduced to headways on 16 and 20 minutes, respectively.

The use of articulated buses is artificially reducing transit service. And since there’s a direct relationship between service increases and ridership, it’s safe to assume that articulated buses are artificially reducing ridership too.

I’m not saying that there’s no place for articulated buses.  While they aren’t appropriate for the 13 on Sundays or late nights, they definitely help accommodate crowds on that line during the rush hour.

Here are a few situations where there is a genuine need for articulated buses:

Missed opportunities with summer busing on the Gladstone Branch

Summer is here! Jersey tomatoes are ripening in the fields, and beach traffic is at its full height. Every summer for the past few years, the summer has been the time when NJ Transit substitutes buses for trains on the Gladstone Branch.  Buses replace trains on weekends in order to accommodate work on the line’s overhead catenary system, the wires that power the trains.

Wooden catenary poles destroyed during hurricane Sandy (WNYC)

This is a significant inconvenience for Gladstone Line riders.  The buses are scheduled to take an hour and 10 minutes from Summit to Gladstone, versus about 50 minutes for a normal train.  The work includes replacing the line’s outdated wooden catenary poles with sturdier, steel poles.

The line is almost entirely single-tracked, with only a few passing sidings, so it’s not possible to close one track and operate all trains on the other,  which is what NJ Transit does to accommodate work on other lines.

The single track is also the reason why weekday peak service is only offered in one direction.  There is a gap between inbound trains between the 4:52 pm from Gladstone, and the 8:08 pm from Peapack. In the morning, the first outbound train arrives at Bernardsville at 8:31am, and the first train that runs all the way to Gladstone arrives at 9:36 am.

To fill in the gap in reverse-peak service, NJ Transit advertises a bus connection on the timetable. This bus, the 986, only offers service to New Providence and Murray Hill stations.  The rest of the line has no bus substitute service.

In addition to offering substitute service on the weekend during closures, why not use buses to provide reverse peak service on the Gladstone line?  The simple answer, like for so many things at NJ Transit, is that it has never been done before.

During the morning, there is no public transportation service at all to Gladstone Branch points, nor any service from the Gladstone Branch to rest of the state from 5pm-8:30pm.The only exception is a few short-turn trains at Murray Hill and New Providence.

This makes reverse commuting on NJ Transit practically impossible. There isn’t much in the way of large employment centers on the Gladstone Branch, other than a few corporate office parks near Murray Hill and Berkeley Heights stations. But there are, of course, schools, small businesses, hospitals and the like.  There probably isn’t enough reverse commuting traffic to fill a train, but it could definitely be the right amount of people to fill a bus.

Speculation on what a partial tunnel closure would look like

In addition to further reliance on ferries, the PATH, and Lincoln and Holland tunnels, closing one of the North River tubes to Penn Station would mean tough decisions on how to use the remaining capacity.  From NJ.com:

The next question is who gets the little commuter rail capacity that’s left?

“The first question is who get those 6 trains?” Zuppan said, who assumed Amtrak’s two hourly trains would continue running. “NJ Transit has four instead of 20 (trains). The big question is what lines get them?

Would that mean two trains an hour each get allocated for the Northeast Corridor, and Morris & Essex lines, he questioned.

“There will be economic ramifications,” said Amtrak’s Schulz. “Regardless of  how the numbers shape up, there will be a reduction and people still need to go to work. How we do it, remains to be seen.”

NJ Transit’s multilevel cars have about 140 seats.  If we assume a heavy load of 50 standees per car, a 12-car train with two locomotives could hold (140+50)*12=2,280 passengers. Multiply that by 4 trains, and we get 9,120 passengers per hour

2013 hub-bound travel data shows that about 48,000 passengers enter Manhattan on NJ Transit trains between 7am and 10am. 22,000 of these people arrive between 8am and 9am. That leaves NJ Transit with a capacity deficit of about 13,000 people between 8 and 9.

One of the tricks that NJ Transit could pull out of its sleeve could be to remove seats from its trains to create more standing room.

The above assumptions are that a single track would allow 6 trains in an 6 trains out per hour.  This could change to 8 trains in and 4 trains out in the morning, which could be done during the peak, but would require changes to equipment staging. During the morning, outbound commuter numbers are much smaller than the number of inbound commuters.  In the PM rush, the number of inbound passengers is higher, so it would be more difficult to change the balance between inbound and outbound slots.

To create alternatives, NJ Transit would have to make crossing the Hudson via ferry or PATH more attractive.  We could see an increasing number of Coast Line and Northeast Corridor Trains terminating at Newark Penn Station, and increased service into

Manual door operation on NJ Transit

This conductor opens the door just as the train is pulling into the station

Most commuters don’t like NJ Transit’s electric multiple unit cars.  These are the stainless steel sided rail cars that have built-in electric motors, rather than being pulled by a locomotive.  If you’re a regular rider, you might know these cars by their dated 1970s interiors with brown faux wood paneling and brown faux leather seats.

These cars also have a unique feature:  manual door operation.  Using a key, the conductor can open the end doors of the car ahead of time, often 10-15 seconds before the train comes to a complete stop.

At low-level platform stations on the Morris & Essex and Montclair-Boonton Lines, this is a standard practice. The conductor opens the door a few seconds early, and walks down to the bottom step as the train is coming to a stop.  The conductor jumps off the train just as the train stops.  Passengers can board or alight as soon as the train stops.

All of the other cars in NJT’s fleet have interlocking doors.  A circuit prevents doors from opening until after the train has come to a complete stop. Only a second or two after this happens, the doors begin opening.  It takes about 1-2 seconds for the doors to fully open and the conductor to begin stepping off the train.  Then, after the conductor has gotten out of the way, boarding/alighting activity begins.

All in all, opening the train doors while the train is pulling into the station saves about five seconds.  This is a mere trifle in time, but stop after stop, it adds up. A local from Hoboken to Gladstone makes 23 stops.  23×5=115 seconds, or nearly two minutes.

Other ways of cutting two minutes off a train’s running time can include engineering work to upgrade the tracks and switches, at the cost of millions of dollars, or by straightening curves on the right of way, which could cost even more.

Manual door operation saves time, but it’s a dying practice.