Speeding Up Muni By Letting All Aboard, Through Any Door

IMG_1664.jpgThe back door of one of Muni’s new low-floor buses. All-door boarding and more low-floor buses could speed up Muni by reducing boarding time. Photo: Michael Rhodes

Some of the most important changes for Muni are also the simplest
ones and there are a number of relatively basic adjustments the MTA could make that would improve service and speed up the city’s buses.

Much of the debate on improving overall Muni speed revolves around how to move buses through intersections faster or give them priority lanes. It’s almost as important, however, to consider how long buses spend not moving: Between 15 and 30 percent of travel time on Muni lines throughout the city is taken up by customers getting on and off the bus.

That’s why speeding up boarding is one of the most important things Muni can do to improve service, which we detail at length in our new Streetfilm, the first in a series of five.

Right now, boarding the bus is a slow process, where every passenger must board through the same door. If one person is fumbling for change, it can back up a long line of riders who already have a Fast Pass or other payment ready to go.

Without any major technological upgrades, Muni could switch to a much faster system of boarding. It’s called proof-of-payment (also known as POP), and it allows riders to enter through any door on the bus as long as they have a Fast Pass or transfer, or if they tag their TransLink card at the rear card readers.

Switching to proof-of-payment and all-door boarding would require more fare inspectors for the buses, and some relearning of old routines for passengers, but it’s hardly a radical idea. In fact, it’s already in place on Muni’s light-rail vehicles when they’re running on the street, where they operate almost identically to buses. Passengers can board through any door — no need to flash the driver any proof you’ve paid — and MTA inspectors make random inspections to stop fare evasion. You still have to pay for a pass at the front of the vehicle, but frequent riders can skip the line and board through other doors.

It works pretty well, too: fare evasion rates are 50 percent lower on Muni’s light rail vehicles than they are on the buses.

Muni’s longest buses are similar to the light rail vehicles in that they have more than two doors. Riders seem to already grasp the similarity: many board through rear doors even now, though it’s still illegal. While that has inspired a lot of frustration from riders who don’t appreciate freeloaders, the reality is that many back-door boarders — about 45 percent, according to an MTA study [PDF] — do have valid fares, and if Muni were to crack down on all back-door boarding, it would likely seriously impact boarding speeds. An "immediate shift to exclusive front-door boarding could result in longer times at stops and slower travel times," that same study notes.

At present, Muni has the highest rate of bus boardings per hour in the country, at 70 per hour, beating out New York by a nose. By switching to all-door boarding, that rate could go even higher, at significant benefit to Muni’s speed.

Proof-of-payment and all-door boarding are common in Europe, but the MTA can also look to a closer-to-home example in Ottawa, Canada. Like Muni’s LRVs,
customers board in front if they still need to pay their fare, and
everyone must be able to present valid proof-of-payment to fare
inspectors at any time. Several U.S. bus rapid transit lines include proof-of-payment already as well, including Cleveland’s HealthLine and Los Angeles’ Metro Orange Line.

But all-door boarding isn’t the only solution. Ideally, major stops would have ticket vending machines, allowing riders who don’t have a monthly Fast Pass to pay their fare without holding up the bus. An example of this already exists for the light rail vehicles at the 4th and King Caltrain station, but it has yet to roll out for any bus stops.

"We’re currently researching different options for our ticket vending machines," said Julie Kirschbaum, Project Manager for the MTA’s Transit Effectiveness Project. "The equipment that we have now is very expensive to maintain, which is why we’re also researching things like the TransLink Add Payment machines, because those can be easier to maintain. We’re currently working on getting more of them in at light rail stations and eventually on the surface as well."

In the long term, the MTA also hopes to switch over its entire Muni fleet to low-floor buses. With bus floors that are nearly flush with the curb, they’re much quicker and easier for all passengers to board, but especially those who are mobility-impaired. Muni has about 80 of these vehicles now, out of a total bus fleet of over 900.

Converting the whole bus fleet to low-floor vehicles will take a while. Switching to proof-of-payment and all-door boarding could happen much sooner, but not happen overnight — the MTA needs to retrain operators and
riders, and hire enough fare inspectors to cover its buses as well. The agency hasn’t offered a time frame for switching yet, perhaps because staff members are tied-up dealing with the current budget crisis and service cuts, but it could be done relatively quickly, with great benefit to its riders and maybe even its budget.

  • Fran Taylor

    In 2006, I noticed that conductors at the Caltrain station were speeding Peninsula commuters onto buses by lining them up and selling or checking tickets to the #10 Townsend, Presidio express, and other buses. When asked if this wonderful idea could be extended to other stops that served local riders, then-MTA spokeswoman Maggie Lynch basically said San Francisco riders couldn’t be trusted.

    I don’t know if this program is still in place at 4th and Townsend. Below is link to Mission Dispatch column that describes it and quotes Lynch. Relevant portion appears toward the end, paragraph starting “Not all riders are regarded as criminals.”


  • The program still is in practice at the Caltrain station. Those are the guys I was talking to this morning about the accident.

    I don’t understand what this has to do with trusting riders or not. Either they have a transfer when they get on the bus or they don’t – regardless if they buy one on the bus or from a MUNI working at the bus stop.

    Question about low floor buses (asking before reading this whole article), how will they handle the hills? I was always under the impression we had higher floor buses so they wouldn’t bottom out, but I could be totally wrong.

  • Seriously, we need to do this. I ride the bus a lot, and people already enter the back door. Seems like those people MOSTLY don’t have fares, so de-stigmatizing back-door boarding would encourage everyone to board that way and really speed up boarding.

    However, I really want to see Muni putting many, many more fare inspectors on buses if this happens. Scratch that — they should do it anyway, but doing this should give them another reason.

  • Andy Chow

    I am not against POP per se, but I can see that if POP were to introduce systemwide then it would create security problems for lower ridership routes in certain neighborhoods. Security is already bad as it is now.

    POP could be experimented on routes that mostly use articulated buses (including routes on Stockton).

    I don’t think ticket machines help. Partly because ticket machines require Muni to maintain, and that it would create unnecessary confusion for riders.

  • david vartanoff

    Rear door loaders were common Muni practice in previous decades.

    @ Andy Chow you write “security problems for lower ridership routes in certain neighborhoods.” I don’t understand. Any door loading is not an issue on low ridership buses. In my experience, security issues are more common on high ridership lines.

    from the I told you so dept. Rescue Muni sent MTA a long list of cheap/free speed ups for the LRV lines a decade ago. They responded with excuses why each idea could not be implemented.

  • MG

    Honestly, it’s been said over and over again before, but eliminating some stops would also do wonders for the bus lines. The difference between the 38L and the regular 38 is astounding. Now I understand the need for more “local” routes, but is it really necessary to stop at what sometimes feels like every single block? I can understand why they do that on hilly streets, but why do it on a perfectly level road?

  • E G

    I have three words for you, “Steveson and 3rd”. There is more than modifications to the service that need to be done, how about cultural education about how to properly make a single file line to board a bus.
    Otherwise no matter how you put we are still going to have the royal cluster F**k that are lines 45 and 30.

  • Andy Chow

    Backdoor loading is not an issue for low/moderate ridership routes because backdoor boarding is still illegal. If you board through the front, you have to show or pay fare.

    If POP is in place, it is possible for someone to insist the driver to open the back door even if the front is perfectly available. Would it be possible for the passenger not to show or pay fare at the front door because supposedly some fare checkers are going to check on the 39-Coit?

    Security is indeed an issue on high ridership routes. Backdoor boarding is one of the causes because some people can ride without paying and hanging out in the back of the bus, therefore people up in the front do not move back as they’re supposed to. Muni would be better off if these folks don’t ride.

  • NBP

    @Andy Chow Onboard security would be less of an issue if SFPD actually rode Muni. Then there would be less tagging, fights, etc. On busier routes, the driver should be able to focus on driving instead of waiting for several dozen people to pay their fare. It is much easier for drivers to exercise discretion on lower ridership routes, where it is easier for the driver to ask for proof of payment, deny boarding, or escort someone off the bus. On busy routes, the driver should not have to delay 60 people, just because 5 hoodlums are misbehaving. That’s the job of the police, not the bus driver.

    People should still be able to pay cash fare on lower-use routes, but it should be actively discouraged.

  • david vartanoff

    @ MG, You have answered your own plea, ride the 38L! Stop consolidation is a wrong answer to a misunderstood problem. If the route in question has MANY longer segment riders like the 38/L then the correct strategy is to run more Ls fewer locals. If you ride the 14, you may see a dozen or more riders board/alight every block, many clutching shopping bags/herding children. They NEED the closest possible stops. OTOH, running more 14L/14X will improve the ride for those going further. The object is to tailor the service to many different rider patterns, NOT to make all riders tailor their trips to a one size fits all mistake.
    As for the other classic mgmt whine “no one uses that stop” if true, then NO ONE will either be there intending to board or signal to alight–no foul.

  • david vartanoff

    A further word about security. Wash DC policy for many years has been drivers are instructed NOT to confront fare evaders. There is a button on the farebox “No Fare Paid”. Security personnel check the buses at end of run in the barn to see which routes they should blitz w/plainclothes officers. The point is to protect the driver from potential violence and leave thug suppression to those hired and trained for the task.

  • Nick

    These ideas have been out there for awhile. Why hasn’t the MTA made marginal changes to improve service over the years? It’s almost like management didn’t care.

    Here’s more backwoods service that mangagement would have noticed if they rode the busses:

    -Bus stops which are in talking distance of each other. Ride the 23 or the 48 line and you’ll see what I mean.

    -A nonsensical ban on folding bikes.

  • @david, I like that idea with DC, but then you’d have to get the cops to do their job – best of luck with that.

    However, SOME stop consolidation is needed. Yes, walking an extra block is going to be a pain for some people, but there are more then several examples of stops on the same block. There really is no excuse for that.

  • Sasha

    @mikesonn: Regarding low-floor buses and hills, until recently I lived right on the 35 and rode it often. All the buses on this route were replaced a year or two ago with low-floor buses, and they do fine going over the hill from Eureka Valley to Noe Valley, and then up the hill to Diamond Heights. The only problems I’ve ever experienced with hills and buses in SF have been on the 24 heading up Castro St. between 24th and 20th — sometimes the bike rack on the bus scrapes the road just past an intersection, which I’m guessing is at least in part due to the fact that those buses are so incredibly overloaded with people during commute hours.

  • Thanks Sasha! Then we have no excuse not to have them.

  • Alex

    @Andy I expect that ticket machines would speed things up quite a bit. Unfortunately the ticket machines that MUNI has tend to be out of service more often than not. In fact, I’d argue that the most confusing thing about the ticket machines is determining whether or not they’re actually working.

    At SFSU, I think it would make sense for there to be ticket machines at the bus stops as well as the rail platform. The 28 and 29 certainly see enough folks board @ SFSU to justify some small improvements.

    @mike Perhaps. But we’ve got no excuse for buying the ones that we did. The ventilation is atrocious, the rear doors are apparently still confusing

    @David As has been pointed out, some of these stops are simply way too close together. Sometimes for good reasons, sometimes not. On the L, off the top of my head:

    15th Ave. There’s a non revenue stop inbound on Taraval, then a revenue stop on 15th at Taraval, and then another revenue stop on 15th at Ulloa. Outbound it’s the same deal (stop on Ulloa @ 15th, then 15th @ Ulloa, then 15th @ Taraval). With a bit of willpower two of those stops could easily be rendered unnecessary.

    22nd/23rd/24th Ave. Yuck. Outbound there’s the stop at 22nd, then at 23rd, then the non-revenue stop at 24th. At 23rd there’s a wheelchair platform, and a small island that only extends as far as the 1st car. Because the island is there, the signage on the street indicates that vehicles are allowed to pass the LRVs… which makes offboarding the second car precarious at best. Inbound is about as bad with a stop at 24th, non-revenue stop at 23rd, and a wheelchair platform about halfway towards 24th that puts two car trains smack in the middle of the intersection. Perhaps the stops at 23rd should be removed, parking b/t 22nd and 23rd eliminated (or moved to 23-24), and both islands moved to the 22-23rd block.

    Yeah, it’s tricky to come up with a better solution, but designs like these absolutely slow down the system and present safety problems with minimal benefit. I agree that if you regularly have 10-12 people at a given bus stop, it’s being put to good use. However, there are a number of L stops west of 19th that rarely see more than five people waiting even during rush hour. Or how about the 28? Stopping every block in the Sunset typically results in one or two people getting off at most stops, and a whole bunch at the major transfer points.

    Blindly removing stops won’t solve a damn thing, granted. But some sort of methodical approach will help. Especially considering the budget constraints, there are plenty of compromises that ought to be at least considered (file that under the “yes the riders need to compromise a bit too” category).

  • @alex. I rarely ride the new ones on the routes I take, but I rememer reading about the back door situation. The LRVs are a joke too but don’t need to add another component to this discussion.

    Agreed on stops though, I’d rather have less stops then less buses.

  • Andy Chow

    The problem with the ticket machines are that they have different models and are not placed consistently. Translink so far is a better solution for those who ride on a somewhat regular but infrequent basis. For visitors, a day pass (can be as simple as an all-day transfer) that can be purchased on the bus will do.

    As for consolidating stops, I think it can be argued that you can remove some stops and enlarge other stops. On Van Ness Avenue, there are many stops less than 500 feet apart (BART/Caltrain platform length is 700 feet), but those stops are less than 80 feet in length. That means two buses can’t serve the stop at the same time. Since Van Ness has two local lines (which I think one of them should be limited), the transit service is actually quite bad. I think it is unfortunate since Van Ness Avenue have lots of destinations but have less transit than Market, Mission and Geary.

    I’ve been on buses on Van Ness that didn’t cross the street on a green light because there was another bus fully occupying the bus stop, and the bus I was on had to wait one more light cycle to cross the street and make a stop. The situation could be worse if that stop were occupied by a Golden Gate Transit bus which people don’t board through the back door and have to put more dollar bills into the farebox. (I’ve seen that in San Jose with buses going to Santa Cruz).

    By cutting some stops, other stops can expand the length without having to cut overall on-street parking.

  • david vartanoff

    @ mikesonn, the DC plan uses Metro’s own security personnel–kinda necessary when operating in DC, MD, VA.

    @ Alex, about Stop signs and revenue stops. The Rescue Muni Metro Committee ((including yours truly) surveyed ALL of the then extant LRV lines a decade or so back cataloging Stop signs and other delay problems. Muni did nothing to fix the issues. While I certainly agree in the cited L @ 15th Ulloa etc example, I am NOT on board with the wholesale stop elimination mantra.

  • Alex

    @David I remember reading about the assessment of stops. The big question is then, why is RM (or someone else) not working to keep this at the foreground of the discourse? Is your assessment of the stops available publicly somewhere? Turnover and attention spans at MUNI/SFMTA is such that ten years ago Nat Ford was embezzling money from MARTA, not SFMTA.

    I read some of Peter Ehrlich’s posts that are about a decade old indicating why the switches / interlocks at SFSU are really poorly designed. This is, of course, salient to the MTA’s latest harebrained scheme to turn the M back at SFSU regularly. These items were valid ten years ago, and they’re valid now.

    I was tempted to attend today’s town hall meeting to raise these issues, but decided against it. Perhaps for the next board meeting…

  • Alex

    @Mike The problems with the LRV doors are more of a maintenance problem created by a design compromise. The problems with the Orion doors is that their use is a bit ambiguous to most. FFS the MTA ditched the last Orion buses (the odd looking short ones) they had because they were simply unsafe (the frames were cracking) to operate any longer. I’m sure there are plenty of decent low or semi-low floor vehicles out there… but neither Orion nor Breda know how to make them.

    @Andy Ticket machines, TVMs, or whatever term is desired is something of a misnomer here. I agree that with this push to TransLink, the MTA ought to put TransLink machines out at heavily used stops. In a perfect setting TransLink can handle about 30 riders per minute (vs about 90/min for EZ-Rider). Having TL machines for riders to use is imperative if any sense of efficiency will be maintained. I’d argue this is more the case on routes that use 40ft buses (like the 28 and 29) regularly than routes that use LRVs or articulated buses. The longer vehicles have more transponders. This, of course, would absolutely go hand in hand with a proper POP system, and drivers that are fully relieved from any fare collecting duties.

    Quite frankly the diversity of TVMs is no reason to avoid deploying another one. The TVMs at SF State are plenty easy to use with very explicit instructions. They’re simply not very reliable. IMO, the real problem with the diversity is in the metro where it’s nearly impossible to pay a cash fare depending on the station you’re at.

    When I rode the bus in Baden Baden, the nature of their POP system didn’t sink in until after I left. They use paper tickets. You can buy a ticket from the driver or from someone else (a machine, or the Avis counter at the airport). When you board you validate the ticket yourself (there was a machine about in the middle of the bus). The driver drives the bus, the fare inspectors inspect your fare media, and so-on. In fact the driver didn’t care one whit whether or not someone showed him fare media (or even whether or not they validated it). The TVMs played a crucial role. Without them, the dwell time at each stop would have been as bad as GGT (the last few times I’ve taken GGT they’ve averaged about a two minute dwell time at each stop in the city).

  • patrick

    @david vartanoff, To my knowledge nobody is talking about indiscriminate stop removal. Most people are talking about the targeted removal of the least used stops. While a stop that sees absolutely sees no use doesn’t slow the bus down, the stops that regularly have 1-2, or even 4-5 people getting on or off are probably not necessary. The one in front of my house would be a perfect example, it ranges between 0 and 5 people using the stop, but there’s another stop on the same block, and another 1 block down.

    If the stop were removed the worst case would be needing to walk an extra block, and many of the people would really only have an extra 20-30 feet to walk. On the other hand, the benefit of improved trip time would benefit every rider and probably result in more riders overall.

  • Agreed with patrick. No one is calling for pulling every other stop without any concern for topography and usage. However, there are many stops that should be examined. As I’ve said before, I’d rather walk a block and have a bus show up, then have a stop right there but no bus.

  • david vartanoff

    @ Patrick, mikesonn et al, While in general two stops on a single short block are excessive, I would point out that Stop signs at corners w/out revenue stops are far worse. Secondarily, IMHO if five persons board or alight in a single instance at a single bus stop, the stop is valid. Third point, @ 65 yrs my knees occasionally give me problems, so though I generally walk a lot, sometimes it is an issue.

    As to “indiscriminate stop removal” an example from AC Transit is useful. They recently planned to discontinue a stop on their 1 Rapid which is a transfer point–WTF??? The puny time saving for the Rapid might be nice, the brokenm connection will decrease boardings.

  • patrick

    @david, if it’s 5 people every time, sure, but if it’s only 1 out of 10 times or less, even during peak hour, then I don’t think it is.

    I’m not saying that nobody will be affected by removing stops, but that the improvements, which benefit everybody (even people who never ride muni through better system efficiency), are well worth the detrimental side of losing a small portion of stops.

    Take the most extreme example of indiscriminately removing half of all stops. At worst, 50% of riders would be inconvenienced, while 100% of riders would benefit. Of course I’m not advocating anything so extreme. I think eliminating 5-10% of the least used stops would have major benefits for all riders, while only impacting 2-5% of riders negatively.

  • david vartanoff

    @ Patrick. In a recent AC Transit analysis of their 51 route (the second busiest route in the system) they estimated savings from eliminating stops at between 20 and 35 seconds each. So I ask if that savings for the bus is worth the inconvenience to the riders. The same or more can be squeezed out by queue jump lanes, signal preempts, bus bulbs. Extending the 5 person example, if five board at a time and then none for say eight passing buses, the latter eight don’t stop. For instance, in evening hours, many stops along Telegraph (AC’s busiest route) get no one on or off on some runs. What this translates to is “local” runs are nearly as fast as “Rapids” yet whoever does need/want service to/from these local stops is accommodated. The same is true of some stops on Muni’s 38 local.
    Again, I am not adamant that no stops be eliminated, just less willing for some.

  • James Figone

    Optimally, a bus would not stop at all between revenue stops. Bus lanes and replacement of stop signs with signal prioritized traffic lights would be exceedingly beneficial. Revenue stops should also be eliminated where it makes sense but it may be that the bigger win in terms of speed is in eliminating non-revenue stops.

  • Alex

    David and James, yes in an ideal world a bus wouldn’t need to stop between revenue stops. However, MUNI doesn’t operate in an ideal world. It operates mostly on streets that are shared with pedestrians, automobiles, motorcycles, skateboards, and the like. It operates on major arteries, and on streets perilously close to them.

    Often times the extra stop signs are put in place as traffic calming measures (or perhaps for other good reasons). A good example of how stop sign removal would be bad is Taraval between 40th and Sunset. The LRVs regularly hit 35-40mph on that stretch (well in excess of the posted speed limit of 25mph). This has obvious implications for pedestrians and automobiles trying to cross the street.

    And, David, out in the sticks… lots of the stops don’t see five passengers at a time on a regular basis.

    IMO, the MTA ought to look at the number of revenue stops first.

  • tea

    It should be easy to tell if this would help. Just look at the performance of lines like the 9x or 14, where backdoor boarding is already a de-facto, regular occurrence.


SFMTA Moves Closer to a System Wide All-Door Boarding Policy

SFMTA Director of Transportation Ed Reiskin told the agency’s board of directors yesterday that the earliest Muni could implement a system wide all-door boarding system would be sometime in the first half of next year. For years, transit advocates have been encouraging the agency to extend its all-door boarding policy on the city’s light-rail fleet […]

To Reduce Delay and Fare Evasion, Muni Considers All-Door Boarding

Source: National Transit Database. There are plenty of eye-popping statistics in the MTA’s new proof-of-payment study [PDF]: 9.5 percent of Muni riders don’t have valid proof-of-payment, costing the agency $19 million in missed revenue annually. The fare-evasion rate is even higher among riders who illegally board buses through the back door: 55 percent don’t have […]