I’ve been thinking out loud about microtransit for a week now, and have processed lots of great comments. But if you comment, don’t just respond to this post. Go to the detailed posts that really lay out the argument: Is microtransit an actual idea? Does that matter? And most importantly: Is microtransit capable of being a sensible investment that can be justified from widely accepted goals of public transit?
To sum up, here’s what I think we know. As always, I’m open to enlightenment by anyone who reads these posts and wants to engage my argument. I’m not contrarian for its own sake. My job is simply to help transit agencies make clear decisions whose consequences they understand.
The context for this thinking is pre-automation, so labor cost is a dominant issue. The question is whether transit agencies should subsidize microtransit, which implies that microtransit is in direct competition for funds with other possible transit agency investments. Thus, the question is about the public interest and benefits to the taxpayer.
Microtransit May Be a Slogan, Not an Idea
As I explored here, microtransit seems to consist of:
- flexible “on-demand” routing, an idea that transit agencies have known about, and experimented with, for decades.
- subsidies of privately provided services by a transit agency, which has been happening for decades under a range of contracting arrangements.
- the use of apps for hailing, navigation, and payment.
Only the last of these is new, but there’s reason to doubt that the apps, by themselves, create a radically new business model. (Evidence that they do could include Uber being more profitable than non-cartel taxis were, controlling for labor costs.) In any case, the statement “transit agencies should consider microtransit” can be translated as: “Transit agencies should use apps to improve the efficiency and customer experience of their flex-route services.” Put that way, it’s uncontroversial and hardly justifies all the hype.
Watch the Ratio of Drivers to Passengers
In transit, before automation, operating cost is mostly labor. Even if you race to the bottom on labor costs, as Uber has done, you won’t save more than 50% off of transit’s big bus operating cost. You still need one driver for every vehicle. That’s why passenger transport services, unlike Amazon, don’t become much more efficient as they get bigger.
That means efficiency in transit is the ratio of passengers to drivers. Microtransit, by definition, is a low-capacity service, carrying small numbers of people at a time. This is, by definition, a way to serve very few people at very high cost, compared to fixed routes.
And as soon as we talk about transit agencies funding microtransit, we are saying that they should do this instead of adding fixed route services that are proven to attract vastly more riders and serve them vastly more efficiently (see table in this post.).
Do Not Confuse Customer Experience with Financial Viability
A common rookie investing mistake is to buy a company’s stock solely because you love its product. Successful ventures don’t just provide a good product or customer experience. They do it in a way that’s financially viable. In the private sector that’s measured in profit. In the public sector the equivalent idea is some kind of cost/benefit or “bang for buck” ratio. Microtransit’s performance on those measures is generally worse than terrible, just as the performance of flexible-route services has always been. Talking about microtransit’s superior “customer experience” doesn’t change that fact.
For example, a recent Eno Foundation report promoting microtransit cited two pilots that achieved less than 1 (one) passenger trip per vehicle service hour. A decent fixed route bus does 20-100 and most terrible fixed routes do at least 10. The most upbeat data Eno’s report could find was a microtransit pilot in Newark, California that achieves 3 passengers/hour, but this is down from 7 passengers/hour on the fixed route it replaced. The transit agency lost 20% of the old route’s ridership when it made this change. And that is the most hopeful data point that microtransit boosters can cite.
Microtransit’s Poor Performance is a Mathematical Fact, not Question of Technology, Social Science, or Marketing
Transit agencies not only have decades of experience with low-performing flexible route experiments. There’s also a purely geometric argument for why fixed routes perform so much better in almost all cases. It’s about the way the customer’s walk to a fixed route stop allows the bus to operate on a straighter path that’s more likely to be useful to more people. The correlation between fixed route performance and the straightness of the route is very strong. Microtransit is meandering by definition, as it has to roam a large area and pick up people who are not in any kind of linear path. Technology never changes geometry.
Microtransit is Not a Way to Increase Ridership Overall
Because of its low productivity, transit agency funding of microtransit arises from a coverage goal, which is the opposite of a ridership goal. Coverage means “predictably low ridership service run for a non-ridership reason,” typically access to places where the built environment makes high-ridership service impossible. The microtransit boosters assume that agencies must run lots of coverage service but this is actually an issue that should be debated; many agencies I’ve worked with have shifted their priorities the other way.
This also means it is incoherent to cite a desire for higher total ridership, or disappointment with declining ridership, as a reason to invest in microtransit. If you want higher ridership, you invest in services that are physically capable of carrying lots of riders and have a proven ability to attract them when run at sufficient quality, like big-bus fixed routes. Microtransit is about taking both funds and political attention away from the services that are actually relevant to ridership at a large scale.
I Cannot Reconcile Microtransit with Economic Equality or Environmental Justice Goals
On average, microtransit seems to trigger an upward redistribution of the benefits of public subsidy. This is a Very Bad Thing for the public sector. It is not hard to make some low-income or social service advocates like a microtransit idea from the customer perspective, and see it as liberating for their clients, but if it isn’t financially viable at a large scale, it won’t matter to the vast majority of disadvantaged people.
If a transit agency invests in a microtransit service hour for 3 people instead of a fixed route service hour for 30 people, solely to give those 3 people a better “customer experience,” we must ask “why are these 3 people so special?” Why shouldn’t they pay the full cost of their superior customer experience, rather than expecting the taxpayer to subsidize it? More on this line of thought here.
Unlike private businesses, U.S. transit agencies operate under intense scrutiny about equity outcomes. Civil rights and environmental justice tests are much more extensive for transit agencies than to the private sector, largely because they are conditions of Federal funding that these agencies rely on. We have not begun to see the blowback against microtransit from environmental justice and civil rights perspectives, but if fixed routes are neglected to fund microtransit investments, the math is potentially there to justify it.
The Popularity of Microtransit Has Explanations Unrelated to its Value
The microtransit movement, like so many fads that have blown over transit agencies during my 25-year career, appears to be an example of elite projection, the tendency of fortunate people to assume that whatever they personally like will be good for society as a whole. An urban elite has seen their lives transformed by ride-hailing services, and understandably wants to believe that this transformation can be brought to transit too. This helps to explain why so much talk of microtransit is so dreamy, so obviously stated in the tone of a sales pitch rather than an analysis. To think clearly in this context, you need to lean into the wind, being skeptical but not cynical about ideas that obviously serve someone’s commercial interest.
The Talk about Microtransit May Be Doing Harm
Fixed routes are spectacularly cost-effective investments compared to almost any flex-route option. They even do better in cases (as in Newark, California above) where the geography is very unfavorable to fixed route success. The reasons for this are geometric, as described above, so technology won’t change them.
Recent declines in bus ridership are triggering all kinds of triumphalist claims from people who want to sweep fixed routes away, or at least shift resources away from them as the microtransit movement proposes. Much of this chatter is intended to push transit experts out of the discussion by implying that their expertise is obsolete. The rigidity of fixed routes, the chatter suggests, arises from rigid minds.
But the neglect of fixed routes, encouraged at the highest levels, is the real source of transit’s declining relevance. My firm works in cities all over the US, and most of them have appallingly low levels of fixed route service compared to potential demand. In most American cities, the quantity of service is growing far slower than population, which means that on average, the availability and usefulness of transit is getting worse. Most cities, in short, are forcing low-income people to buy cars by making that the only way to have a life, even in places where fixed route service could succeed.
In this reality, should transit agencies really focus on ways to move tiny numbers of people more expensively, to deliver them a special “customer experience”, as the microtransit idea proposes? Clearly that’s not the path to ridership.
Meanwhile, cities that are forcefully recommitting to fixed routes are bucking the trend of falling ridership, and these show a clear path. Ridership is up in Seattle, despite all the countervailing trends, because of an unusually high commitment to quality service and to protecting fixed routes from congestion — a commitment shared by the transit agency and the City of Seattle. Houston continues to do far better than its Texas peers, partly due to the 2015 network redesign that expanded the bus network’s usefulness.
We know how to increase ridership. It’s by offering useful, civilized, and cost-effective mobility to large numbers of people, not obsessing about the customer experience of a few. And while ridership is not the only goal of transit, it’s hard to get to microtransit from any of transit’s other common goals either.
If You’re Going to Comment …
I would love to see comments engaging my argument. I’m not especially interested in comments that ignore my argument, or change the subject, or give me dreamy visions of a future where laws of math have been repealed, or lecture me about how I represent tired old thinking. If you’re going to challenge me on a point where I’ve linked to another article, follow that link and read that article, because the meat of that argument may be there.
Thanks for your help making me smarter about this. I advise a lot of transit agencies, and I want to best for them and the cities they serve.
Here’s one suggestion – stop thinking about microtransit as just improving “customer experience.” Think of it as removing very real, non-elite, and important barriers to transit use.
App-based microtransit can have some real advantages. It can reduce lost time for some customers: by shortening walking distances, eliminating the need for transfers between modes, and providing more direct routes for travel between trip ends that don’t correspond to a single fixed route. It can provide for greater reliability and predictability – instead of a moderate walk to an uncertain pick-up time, it gives you a shorter walk to a known pick-up time, so you can time your departure more certainly.
These would not be “elite” benefits. As you know, the longer commute times and uncertainty of transit systems are real obstacles to transit use for *all* users.
These systems may have lower ridership – but they can carry some lower costs as well. You could service a microtransit area with an off-the-lot Pacifica for a tenth the cost of a municipal bus, and better fuel economy besides. And because the product is better, you could even charge a modestly higher fare. You might end up acquiring riders with a lower per-rider subsidy than many other coverage routes.
As you lament, we’re not providing fixed-route service in lots of places. That may be because the customer-borne inefficiencies of fixed-route service (lost time to indirect travel routes, transfers, and unreliability of pick-up schedules) are blocking *customers,* and thereby blocking electeds from expanding service. Perhaps microtransit can eliminate some of those very non-elite barriers…..by focusing on the customers, not just the needs and goals of the agencies.
“Higher fares” are the only idea that makes sense here from an agency perspective. Some agencies (notably in Livermore-Dublin, CA) are experimenting with subsidizing TNCs to serve areas where fixed routes have been removed, but all these subsidies buy are small discounts off the naturally high TNC fare. Cheap vehicles don’t matter much because operating costs are mostly labor, and cheap vehicles have shorter lives.
Again, the provider’s perspective has to make sense, especially when spending taxpayer dollars.
Please see my response to your post on February 20 “Is Microtransit a Sensible Transit Investment?” coincidentally trying to play out hypothetical scenarios for higher fares. Please let me know what you think.
Waymo is now licensed to operate as a TNC in Arizona. They expect to offer commercial service in 2018. Their level 4 autonomy removes the labor costs. The less expensive vehicles can actually last as long, or longer than a bus. Electric AV vehicles have fewer moving parts. AV operations can optimize regular maintenance. For example, it improve longevity of brake pads, and battery function. Batteries are highest replacement cost and even that life cycle is extending every year. Some companies are even looking at swapping low batteries for charged ones during operations, rather than forcing a vehicle to come out of service to complete a full charge cycle, This reduces the number of vehicles needed to maintain a minimal level of service.
If your thinking the cost of making that Pacifica Autonomous will kill the costs benefit, Tesla model 3 sells for $35K and comes already equipped with autonomous hardware. The software is expected to be $10-12K to purchase. so for $45-50K you could have a 5 passenger AV TNC vehicle in operation. In a 20 year life cycle that is roughly $200/mo. Companies like Easy Mile and Local Motors have already looked at subscription models to cover purchase, operations and maintenance expense, and continuing to look for ways to lower vehicle manufacturing and operations costs.
Even if the buses don’t go driverless, they will go electric, as the technology matures, and the old diesel buses reach their end-of-life and need to be replaced.
If we have electric AV cars, that means we also have electric AV buses, which also have the same labor cost savings. That basically keeps the advantage of fixed route traditional buses over microtransit.
Bang for the buck. That’s the deal. And “customer experience” may be crappy in public transit(it does not have to be, here in Montpellier, Hérault, it’s excellent, good money has been invested, trams are full), it’s still the best transport option in terms of finance.
I live 4.5km from my workplace. That’s 2.1/2 miles in US terms. If I take the car that’s 9km, and the cost of 9km would be around 3.5€ (tyhe standard here is 0.4€ per km). Taking a 10-travels ticket costs 10€. That’s 2€ per day(I don’t count the yearly ticket I’m using, which does not make sense for most people over there who don’t have a fixed schedule all year long). I’m neither counting park costs at home or at work(which can be non-negligible).
now what about the transport time? Maps tells me 9 minutes by car, but it’s misleading. I need 6 minutes to reach my car from my flat, exit the underground car park, and I need anoth 4 minutes to properly park at work and reach the building’s entrance. That’s 19 minutes, not counting the frequent traffic jams.
Tram, compared to that, is a 4 minutes walk from flat door to the station, average wait 7 minutes(there’s one every 14 minutes, my job is on a low-service branch), 6 minutes of tram drive, and 3 minutes walk to the office’s door. 20 minutes. One more than using my personal car, for nearly half the cost.
Microtransit would be far worse. I’d need only 1 minute to exit my flat & go outside, but then an unforecastable amount of time for the microtransit, that would then do a non-straight transport. I’d be probably the last one to be taken, considering I’m at the edge of inhabited areas, but other users would have done meanders to get me in the car, while I’m not on their way. Then the car would probably do meanders in the Boirargues commercial area to drop a few retail workers, then another one to drop me at ecopole, then go to the airport for a last person. And I don’t see how all that could cost less than 3.5€.
The only competitive option, honestly, is not microtransit. It’s bike(’cause there is an excellent bike path all along the way).
Now, there are other scenarios.
Someone living at the other side of the town? Tram is 10/15 minutes longer than car, but does not cost more. And micro-transit would do even more meanders, being probably as long as tram, if not worse(traffic jams are not a legend, over there).
Someone living in the countryside? Well, by car, he’s at work directly, for a global cost that is probably not much higher than microtransit, and as he lives in the middle of nowhere, his parking costs are moot.
Someone living in the dense centertown? well, cars can’t go there, so his options are basically walk 10 minutes to take the tram, or bike(both are biking, actually). If they would walk 10 minutes, then they’d take a direct tram – far more secure than a microtransit that might go here, and there, at unpredictalbe schedules.
Said otherwise, I can’t see any scenario where microtransit is not severely beaten by at least one of those 3 options : bike, tram, personal car. I’ts overall as costly than car, you have to face promiscuity as well as in tram, and given the meanders and random wait, it’s at least as long as bike.
Yes, the tram – and other fixed-route transit – can be much more convenient for some riders for some trips. If you’re close to a transit stop, and your destination is close to a transit stop *on that route*, fixed-route transit (especially traffic-separated transit) can be very efficient.
But what if you’re not? What if your flat is far from a transit stop, and you’re not going to a destination that’s near the route that transit stop is on? What if you’re not going to the center of Montpellier, but trying to go from Saint-Jean-de-Vedas to somewhere near Juvignac – with a ten-minute walk on either end (going between spoke ends in a hub-and-spoke system)? Even with a few minor detours, it’s easy to see that mictrotransit could offer an improvement over the travel time of a tram ride downtown, a transfer to another line back out to the edge of town, and a ten-minute walk on both ends.
That condition may not happen very much in dense, compact Montpellier (4.5K people/km2). That condition happens a *lot* in U.S. metro areas. For trips that don’t originate near a transit stop, and don’t go downtown. Mictrotransit offers a lot of advantages.
>>Even with a few minor detours, it’s easy to see that mictrotransit could offer an improvement over the travel time of a tram ride downtown, a transfer to another line back out to the edge of town, and a ten-minute walk on both ends.
Of course it would — at great cost! That is the point. If you are talking about an area that can’t justify good fixed route service, then it is:
1) Expensive to serve (no matter how you serve it).
2) An area with low demand.
3) An area that the agency involved doesn’t think is worth serving.
All these go together. Of course there are places that will be served this way, just as there are places that are served this way right now (as Walker mentioned). But a place with even moderate demand is much cheaper to serve with fixed route service.
In other words, if a place is served poorly with fixed route service, what makes you think an agency will serve it well with service that is more expensive?
“Of course it would — at great cost! ”
Yes, the cost per rider becomes very high. I am not sold on microtransit, but was still shocked to see the low productivity and high costs per passenger in some of the cited studies.
>> It can reduce lost time for some customers: by shortening walking distances, eliminating the need for transfers between modes, and providing more direct routes for travel between trip ends that don’t correspond to a single fixed route. It can provide for greater reliability and predictability – instead of a moderate walk to an uncertain pick-up time, it gives you a shorter walk to a known pick-up time, so you can time your departure more certainly.
These would not be “elite” benefits.
Yes, they are. What you are talking about is simply a lot more expensive to operate. In that sense, they are certainly elite. A fixed route, fixed time system, operating on a grid and involving transfers is much cheaper to operate. The only reason that on-demand routing provides a better level of service is because it is more expensive.
I’ve made this analogy before. Consider an airport shuttle. An agency is tasked with getting people from the airport to their home. The best type of service involves taxi-cabs. Everyone gets a straight shot to the airport. Obviously this is extremely expensive. It is the cheapest option if you have very low ridership, but it still isn’t cheap.
So people decide to share rides. You buy vans instead of cars. But how many riders do the vans carry? You could wait until the van is full, but that is not good at all for the riders. A van might have to wait an extra half hour just so that they can get that last passenger. Obviously it makes more sense to run the vans at set intervals — basically as often as you can afford to. If demand is high you could leave every five minutes, which is both great for the customer (at most a five minute wait) and good for the company (the vans are mostly full). Already you can see how the system has evolved to be fixed time (not on-demand) because on-demand is either too expensive (very few people in the van) or not good at all for the customer (too long of a wait in the van).
So now you have to decide where to send the vans. It doesn’t make sense to send them both north and south. It makes sense to send a van to one part of town. So know you have different vans going to different parts of town at a fixed time. Since this is an airport shuttle, you drop people off at their house. This is a great service for folks close to the airport. But those at the end of the line have to wait while the vans go through various cul-de-sacs and other detours. There is another problem. The vans take a lot longer to drop everyone off. This means you can’t run the vans as often. For the greater good, it becomes clear that you want your route to be fairly straight, so those at the end of the line get a faster ride home and the vans can run more often. That means some of your riders have to walk a couple blocks.
But now you run into a different problem. You’ve found that delivering some of your riders involves either major detours, or excessive walking. It is one thing to ask someone to walk a couple of blocks, it is another to ask them to walk a mile or two. While mulling over this tough problem, you are just notified that someone wants to expand your system. They want to use it to get across town. Great! You can kill two birds with one stone. You run one van away from the airport, and other vans perpendicular to that van. You provide cross town service (using the same approach) and now you don’t have to worry about going out of your way to serve those riders who live off the beaten path.
Alright, things are humming now. But it still isn’t that cheap. You are spending a lot on the drivers, who operate all of these vans. You want to increase service, but can’t afford to. So you replace the vans with buses. This means fewer trips and more stops — which means a degradation in service — but you, the company, make more money. It also means you can charge less (or if you are a public agency, your subsidy for each rider is less). Buses on a fixed route, fixed time schedule just makes the most sense — even if the service isn’t what you had hoped initially. Riders have to wait a while at the airport for their bus. They may have to make a transfer. They may have to walk a few blocks. But overall, it is the cheapest way to provide the service.
I’ve just described how a fixed route system evolves into a standard bus system, with only moderate use. Of course if you only have a dozen or so riders a day, then this system doesn’t make sense. With ridership that low, a taxi-cab makes more sense. But that is (obviously) the most expensive to operate. An on-demand, sharing system operates in a very small window between the fixed route systems — that are cost effective with even moderate demand — and the taxi cab system (which is only the cheapest option when ridership is extremely low). Such a system will never be large. Once it gets large, it is far more cost effective (and better for the customer) to switch over to a standard bus layout.
“It can provide for greater reliability and predictability – instead of a moderate walk to an uncertain pick-up time, it gives you a shorter walk to a known pick-up time, so you can time your departure more certainly.”
Maybe I’m missing something, but I don’t see how microtransit is necessarily more reliable and predictable. A microtransit is as likely to be stuck in traffic as a bus (and less so if your city has decent bus lanes). And if your transit agency has an app with real time data (like TransLink in Brisbane), you have a very good idea of when your bus is going to arrive.
“My firm works in cities all over the US, and most of them have appallingly low levels of fixed route service compared to potential demand.”
What is the potential demand in average American cities and suburbs? For instance in south King County and Snohomish County (suburbs around Seattle), or in the Bay Area outside San Francisco? If these areas had a complete frequent grid network, would non-car travel grow to levels approaching the primary cities, or what percent of households might be willing to go carless? I’m thinking more of local transit rather than trains/express buses to the city, because that’s where the biggest neglect is. It seems like we and the media talk around this but never say what is the ultimate goal or maximum potential, and how much do our lower-density neighborhoods and suburbs limit us compared to Canada and Europe? This would be worth an article because I’d like to hear how transit experts see it. I’ve heard that Canadian and European cities have at least twice as much transit per person, and not just because their neighborhoods are higher density; they also have half-hourly or hourly service to areas Americans consider too low-density for transit (the outskirts and rural towns). How much of America’s problem is the lack of fixed-route transit, and how much is it that our lower density and single-use zoning make it so difficult to get to things that people will drive even if transit is there. It must be something in the middle of those.
The comparison to Canada is a good one, because Canadian metro areas look pretty much like American ones in terms of the range of land uses. At a macro level it’s clearly the service level, not the land use.
For another reference, Australian capital cities (or largest) have transit ridership of 10-15%. As our cities are all less than 250 years old and experienced most of their development in a post-war car dominated environment, they have a similar density and land use to north America.
This may be somewhat of an aside, but related to — I think — some of the “problems” that Microtransit supposedly solves: I think American transit agencies need to be more focused on improving overall non-car transportation in the cities they serve (working with the local gov and community), by pushing for better walk and bike infrastructure. Transit agencies should not see walking / biking as competition, but rather as an extension of the transit network.
Is any agency doing this? Tri Met is perhaps the only one that I think /would/, but I don’t know if they actually do.
Because if we’re talking about Microtransit serving the “last mile” then that seems to be a colossal waste of money for such a non-actual-problem. The last “mile” should be walking or biking. Personally, my experience using most U.S. cities’ substandard transit service (compared to EU/Japan) with 15 minute frequency at best, if I’m only going a mile or so, it’s faster to walk (not to mention I can start walking immediately vs waiting for the bus). And if I’m going 5 miles or less, it’s often faster to ride a bike.
I think many transit agencies would better serve their communities by focusing on frequency of service and the “Time in transit”. Buses in the U.S. are generally too slow — not because they are stuck in traffic, but because they don’t come often enough, and the bus stops are spaced too close together so it’s always stop-and-go and takes an obscene amount of time just to go a couple miles.
Somehow, U.S. transit agencies don’t seem to view this as a problem, even though it seems every other country has figured out that infrequent service and closely spaced stops are NOT the right answer. Will we ever learn?
Integrating transit/walking/biking alternatives is the job of city planners, and it’s usually beyond the authority of transit agencies. Transit agencies should partner in it, but somebody else has to take the lead, and if your city/county isn’t doing it then you should urge your elected officials to get it done.
My agency, King County Metro, has recently started widening stop spacing from 0.12 miles to 0.25-0.5 miles and making core corridors frequent, but it’s a long process of a route or two a year due to limited funding. The state does not fund transit and strictly limits how much counties can raise for transit, and voters in some counties are more enthusiastic than in others. Jarret’s point about the neglect of fixed routes it the issue: not enough politicians or voters see it as a priority, so they don’ enact it.
The main problem we have here in France with bikes is theft. Securing the bike parks next to the bus/tram stations would help a lot.
Ben G., at my transit system we routinely work with local municipalities in our service area, as well as the MPOs, to advocate for and encourage upgrades to pedestrian infrastructure. Over my career, I’ve worked at four transit systems in four different states, and they all advocated for this to the extent possible. In select circumstances, we are even able to assist with funding it. However, as Mike said, it is usually beyond the authority of transit agencies. We typically do not own the right of ways that our services are on, or that people would use to get to our services.
Current trends show us the testing of microtransit is going to occur with, or without transit agencies. Experiments in both microtransit and Transportation Network companies are well funded and expanding. Transit agencies must participate in pilot projects, simply to be a part of the conversation. Why?
1) We represent the core values of, and have long relationships with the communities we serve. Those perspectives must be weighed against private interests.
2) Technology has given us the ability to organize vast amounts of complex data in ways never available to us before. Transit solutions are not only about geometric laws, they are also about temporal ones. Nobody is arguing that technology changes geometric or temporal laws (though Einstein suggested the later is possible). Technology is changing the spatial efficiencies within urban geometry, and how that efficiency is spread across time. It’s also changing the speed in which real time information is gathered, analyzed, and disseminated. These changes alone can alter the results of calculating geometry and time in transit planning.
3) Younger generations and recent grads are showing us new ways to organize and think about this data and it’s application. Their entire educational career since kindergarten, has had access to the internet. Smart phones have been available to them since adolescence. Their ability to shape technology to improve approaches to problem solving, and finding better solutions to old problems is unprecedented, and why tech companies like Amazon look for high concentrations of millennial populations to site their new headquarters. If given the chance, they could make transit systems better. Under private company influence, they’ve already helped reshaped taxi service (maybe for the worse), and everything from rocket science, to business models, and more recently even politics (maybe for the better).
4) The ENO report doesn’t show the Microtransit pilots reviewed improved geometric efficiency, but does attribute failures to the inability to make quick adjustments within existing bureaucracy, and the role that marketing makes in determining success. Citymapper in England states it greatest limitation to building a responsive transit network is the archaic bureauracy in London that limits it to using 8 passenger minivans, instead of larger shuttles. Limiting to 8 passengers to avoid strict outdated regulations is not efficient use of geometric space. To understand microtransit and it’s role as a piece of the mobility spectrum, and whether can be an effective cog within a networked transit system, we need more pilot projects that address these deficiencies.
5) To not participate, could allow transit agencies to become privatized over time, or simply go they way of the dinosaur, which may be to the detriment of our cities….or maybe not? Could a well funded demand responsive private network company simply following fixed route public transit routes using automation, bundle tech services, and real time data to low costs, diversify revenues, offer lower fares and better, more equitable service? It’s not elitist to at least want to explore the concept to see if there is benefit to a public-private network.
6) Fare income will not be the only economic driver determining the success of autonomous microtransit. (If you’re wondering why anyone would fund a private service described above in an attempt to replace publicly owned transit, just ask the Koch brothers). Advancing micro transit pilots isn’t just about data on past failures, it is also about current politics
Microtransit is not a replacement to fixed route transit, but COULD be a supplement to it. Microtransit can be another option on the mobility spectrum. That could bring new riders from private cars, onto transit. We have to dig deeper into this potential, and to not do so at this time, seems irresponsible.
JD. Would you please unpack this sentence? “Technology is changing the spatial efficiencies within urban geometry, and how that efficiency is spread across time.” I don’t know what this means.
I am unclear why any of the laudable goals you mention requiring throwing taxpayer money at something that doesn’t make mathematical sense, just because cool people want us to do it.
Eno’s claim that microtransit could be saved by marketing makes no sense to me. The problem isn’t microtransit’s ability to attract passengers. It’s the low capacity intrinsic to the whole idea. As with paratransit, it’s not in a transit agency’s interest to grow a market that can only be served at very high cost/rider.
People won’t use a service if they don’t know it exists. It’s fair to say that to truly understand a new idea’s performance, you need to remove other variables. Making sure users know about the new micro transit service an agency is trying to pilot, removes a variable. If a marketing campaign had filled those services daily in those pilots, the performance analysis and cost per rider tells a better story.
I agree, even full it is still a high cost today, but technology is driving down the costs of shuttle maintenance, operations (both in Labor and Fuel, fewer moving parts), capital costs to make the vehicles and purchase them, and reducing costs of the secondary impacts on infrastructure (Large buses deteriorate roads faster than lightweight transit shuttles designed to date). Whole system and life cycle costs of that system must be considered. If in 5-10 years, you could manufacture, maintain, and operate a 10-passenger electric shuttle for the same costs as a high end dutch e-assist cargo bike that holds 3 passengers, wouldn’t the cost per rider be more amenable to the use of tax payer $ to pay for the service? Not saying cost will get that low, but the point is still there, costs are now, and will continue to decrease, before leveling off. Besides electric fuel and autonomy, another example of how costs will reduce; many companies are looking at carbon fiber bodies that can be 3D printed.
Do you know how low microtransit service would need to be to get to a cost effective cost per rider with an 8-12 person capacity?
To be clear, I do not see microtransit replacing fixed routes for the same geometric reasons you do. I do see them as a potential tool in helping those fixed routes capture new riders that usually drive, and to compete with, and reduce (fight?) the impacts of privately owned single occupancy autonomous vehicles coming in the near future.
Walk with me for a second (pun intended), Imagine the potential transit users who choose to drive instead of take transit; What is their reasoning for not taking transit. One of the biggest reasons can be tied to “the walk”. “I don’t have time to walk, I’m simply to lazy to walk, it’s too far to walk, I’m a parent with kids too young to walk. It’s too cold or snowy outside 9 months out of the year, but I walk the other 3. It gets dark early and as a women I don’t like to walk alone at night. I am simply physically unable to walk.”
Microtransit that takes people from their front door to a fixed route service can address all of the issues with the walk, and feed riders onto existing fixed routes. More riders on high capacity fixed routes equals better cost/rider and better use of subsidies. If it feeds to certain high density/high ridership stations, perhaps it can consolidate stops to make the high capacity route even faster, which could equate to even more riders (more money savings to be considered). So if microtransit costs get lower, and then serve to increase overall system ridership, aren’t we winning?
Other examples why people don’t take transit, “I drive to a P&R then take transit but when I get there it’s full, so I drive”. Demand responsive and networked Microtransit could help this. “Fixed route service is inflexible and it’s too hard to trip plan for multiple locations throughout the day.” Algorithms can use real time data to make instance decisions, that allow people to make the best choice available to them, at any given time, based on the facts occurring elsewhere. App interfaces are already making trip planning seamless and effortless, even with multiple stops. We have every reason to believe it will one day be capable of planning our whole day for us, without us thinking about it all.
If you’re continuing to think the subsidy for cost per rider is still too high to validate this type of front door to transit service throughout our cities, then let’s not forget one significant reason this issue is being revisited in the first place; autonomous vehicles. Microtransit isn’t about catering to a few elites, it’s about using all tools to make a more robust transit/new mobility toolbox to fight against a future with privately owned, Single occupancy autonomous vehicles that offer users convenience and more free time. That AV scenario definitely caters to the elite few, and can exacerbate the current TNC trends that increased VMT and shifted transit ridership growth, to TNC growth. Unchecked, it could be the next catalysts of urban sprawl. I hate to say it, but we taxpayers are already helping to subsidize AV development.
Well said JD! I totally agree with you. I believe Microtransit has a potential for success and complement the current transit system for the same reasons you mentioned.
>> If you’re continuing to think the subsidy for cost per rider is still too high to validate this type of front door to transit service throughout our cities,
Yes, yes I am. Of course it is. You are talking about either:
1) Taxi-cab service. This is extremely expensive, as you only have one rider per vehicle.
2) Shared taxi-cab service. This is either expensive or inconvenient for the riders. Someone at the end of the line has to wait for the inconvenient detour. It is still expensive, since you are doubling up routes (no transfers). These aren’t likely buses, which means that you carry very few riders per driver, and that is extremely expensive.
(Sorry, I interrupted, you were saying)
>> If you’re continuing to think the subsidy for cost per rider is still too high to validate this type of front door to transit service throughout our cities, then let’s not forget one significant reason this issue is being revisited in the first place; autonomous vehicles.
Ah, OK. Let’s try that first one then, shall we (since it is simpler). Let’s compare an automated taxi-cab system with an automated bus system. The automated bus system works the way any efficient, ideal bus system would work. You run buses frequently on every corridor a few blocks apart in a grid. This means that riders have to walk a bit to their bus, but not too far. It also means that transfers are common. But it also means that the buses run fairly fast, and as Walker has written about many times, this is by far the cheapest way to deliver anywhere to anywhere service within the city.
You’ve eliminated labor costs, so now the most expensive cost is buying and maintaining the vehicles. But how many vehicles do you buy? On a lot of the corridors, vans will do just fine, but on others, you need buses. You buy enough vehicles to handle peak demand with good frequency. Each vehicle — at least at some point in the line — should be approaching capacity. There is a moment, during the day — probably around 5:30 PM — when you are at peak ridership. You have a lot of vehicles, and they are all carrying a lot of people.
So how many cabs do you need? Well, at that same moment, you will need one for every person that rides. That is a lot more vehicles. An enormous amount, really. Obviously that system is just a lot more expensive.
But wait, you still have that second option. Have people share the ride, but have the fancy algorithm figure out the most efficient way to get people around. OK, but at a minimum, that means the same set of vehicles, all operating at close to peak capacity (remember, that is what the traditional bus system was doing). If you aren’t operating those buses and vans in a similar manner, you are buying more vehicles. So you are claiming (without any evidence) that you can find a more efficient way to run those buses and vans; something other than straight down the road — no turns — with stop spacing designed to balance walking distance and speed? Call me skeptical. No one has yet — to my knowledge — come up with a mathematical proof of that. My guess is because it isn’t the case. You can run all the computer power in the world trying to figure out the ideal shape from an area to perimeter standpoint, and you still come up with a circle. The same is true for a grid and transit.
Which means the question is whether it is actually worth it. Is a dynamic, shared taxi-cab system so superior to a fixed route grid system to be worth the extra money. I seriously doubt it. I am talking about a system that is every transit expert’s fantasy. Buses (or vans) running every five minutes on every corridor. Not just the big ones — the smaller ones too. Yes, riders have to make a transfer — but the wait time is minimal. You want something that is more expensive, because it *might* provide better service for some. Yeah, OK.
Besides what JArret says, there is a contradiction, also. You say that it would work better with bigger shuttles. But the caveat is :
(1) The bigger the shuttle, the more passengers it takes
(2) The more passenger it has, the more complex the route(’cause everyone goes from a different place to a different place)
(3) The more complex the route, the more time is lost in meanders.
I mean, the strong point of the taxi is that it takes the shortest available route. “straight fixed routes” Jarret praises a lot try to mimic this strong point. They are not as good, but they are the goal of any transit planner for a very good reason.
In my long answer to Albaby, I assumed much smaller vehicles(up to 6 places) for this very reason. Of course, there are plenty of optimizations a good algorithm can make on such a tortuous route, and this in real time. Still, if not everyone is on the way, the total travel time will be muuuuuch longer than in my fucking tramway. So of course you don’t wait 10 minutes for the transfer if you’re not on the straight line. But for this benefit, you pay : (1) a level of comfort even lower (2) a much higher fare (3) a longer travel through areas of the cities you’d better never have visited.
And I’m by no means a transit expert, but I’d like also to know the cost in time of each micro stop. In a properly set up fixed line, the stop is very quick thanks to optimizations. The bus/tram stops in front of you becauses you’re at the stop and it knows the stop. With your prepaid ticket, you enter, validate, and seat(or try to find a place to stand if it’s overcrowded). And you exit the other way., so that it goes quick.
In a microtransit, you have 2 problems at each stop. First one, the driver does not know you, and loses time stopping right in front of you. Augmented reality could maybe improve that(highlighting you on a screen), but the ease of parking in a random place would not be as good as the ease to park in a reserved, pre-defined place. Second one, if we assume vehicles are smaller(they’d better be, 60-places buses are not easy to park everywhere), ensuring a proper flow of passengers is much harder, so the boarding/unboarding time is also increased a lot.
Those are additional problems that modern computer technology can solve only partially(finding the right place) or not at all(parking at the right place).
I think about whether microtransit has a shot in the first mile/last mile market to/from suburban commuter rail stations for customers who currently drive and park themselves.
Thinking about the geometry issues you raised – these should be reduced in a first mile/last mile connection context because everyone has at least a collective destination (in the a.m.) and origin (in the p.m.)
Thinking about economic equality, environment goals and subsidies – many suburban rail stations have free or at least heavily subsidized parking (particularly those with large parking structures). So if the subsidy transfer is from building/operation parking at these stations to running microtransit…maybe, on balance, that’s a good thing?
The series this week has been a fantastic read Jarrett. Always appreciate your insights.
I don’t see how microtransit solves the issue at suburban rail terminals. My home is 10 minutes by car, 20 by fixed-route bus from the nearest metro rail line. The reason I typically drive and park at the rail station is not the extra 10 minute ride or 5 minute walk from the bus stop to home, but the abysmal evening frequency of the bus which can entail a 30+ minute wait from train to bus. Can microtransit solve the frequency problem at a reasonable cost? If money can be found from selling the parking garages, it appears to me that increasing the evening frequency of the fixed bus routes is more cost-efficient than a fleet of subsidized taxis.
“Can microtransit solve the frequency problem at reasonable cost?” – I think it might be able to. Particularly if you don’t have to provide a $30,000 parking space in a parking structure with no revenue stream for that customer.
Hamish. Thanks for the kind words. If you do things that build up the market enough to support microtransit, you’ve probably done enough to support a fixed route. My point is that there seems to be very little range where microtransit is better than fixed routes at the same labor cost. There’s a small range of special cases only.
“There seems to be a very little range where microtransit is better than fixed routes at the same labo(u)r cost.” Absolutely agree with this.
Again, only from the perspective of the transit agency. Microtransit is generally better for the customer. Fixed-route transit places enormous burdens on many of its users, especially outside of the densest urban areas, which microtransit can alleviate. Customers have to bear those costs to make it easier for fixed-route transit to work; microtransit takes on those costs so *customers’* lives are easier.
In the near term, customers have to bear the travel time to the transit stop to start their trip. They may have to bear wait times for transfers. They may have a longer trip due to an indirect route. If they have to travel to a new location, or have to get to a location at a time they are not used to, they may have to spend time educating themselves on trip schedules in order to plan their route.
In the longer term, customers have really adjust their lives to the needs of the transit system, rather than the other way around. To efficiently use fixed route transit, they have to locate their home and work – and other frequently-visited destinations – around where the routes run. Hence, they may have to pay a rent premium to live near a convenient transit stop, or forego a job opportunity because it’s not proximate to a route that runs by where they live.
So now look at mictrotransit. Now, *every* dwelling unit in your metro area is within a block or two of a transit stop. That transit stop offers a near-direct route to *every* destination, with no transfers required. You never have to research transit schedules, even when visiting a new destination. It’s worlds better for the riders.
More people would support that type of system, because it’s *better for the riders* than fixed route. You could charge higher fares. You could get the electeds and taxpayers to support more public funds – because the system is more convenient, so they could see themselves using it.
That’s why the electeds – and the riders – are very interested in microtransit. Not because cool tech people are advocating it, but because everyone has seen how much more convenient a TNC pool service is than using fixed route transit in many areas of the metro, especially the coverage route areas. If you keep focusing on the needs of the agency, rather than the needs of the taxpayers that use the system, you’re overlooking the most important aspect of this.
Remember, a TNC offering a pool service doesn’t need to be *profitable* to pose a threat to the transit agencies. It just needs to have a better farebox recovery than the transit system. At around 30% (ex NYC MTA), that’s not a high bar. They’re all subsidized now by equity investment….but their business plan could change to trying to capture the public subsidy that now goes to fixed-route buses in most metros. The pitch is simple – better customer service for lower public cost. The goals of the agency aren’t going to matter if TNC’s can serve the customers better.
“a TNC offering a pool service doesn’t need to be *profitable* to pose a threat to the transit agencies. It just needs to have a better farebox recovery than the transit system. At around 30% (ex NYC MTA), that’s not a high bar”
Farebox recovery is entirely a question of policy: what level does the agency want to charge riders? This intersects with market forces: passengers won’t pay a fare they consider unreasonable or uncompetitive with other alternatives, including driving or or not making the trip. So agencies choose an arbitrary rate for political or budget reasons, but there’s an implicit ceiling in terms of what passengers are willing to pay or other policy priorities (i.e., serving poor/disabled riders and coverage routes).
I think your real question is, “Can microtransit have lower expenses for comparable service?” Smaller vehicles cost less, but on the other hand it requires more vehicles and drivers. And while one fixed-route run can serve several overlapping trips, the equivalent multiple car runs may serve different portions of the route or different turnoffs. Depending on the cars’ operating patterns, that could lead to lower frequency for any particular origin/destination pair. And the microtransit will probably charge higher fares, possibly two or three times as much as the bus given existing Uber/Lyft practice, leaving out a subset of passengers. And the city or agency may deny that it is leaving them out, so they not only lose access to transit but their plight is ignored.
If we’re looking at higher fares for microtransit, then the same farebox recovery percentage is still more expensive than fixed route (assuming 30% farebox recovery, a $2 fare means that the subsidy is $4.67 for a total ride cost of $6.67, where a $4 fare means the subsidy is $9.33 for a $13.33 total ride cost.) This also ignores the fact that almost all agencies get less than full fare from each origin – destination pair at least some of the time (senior/disabled/child discounted fares, daily/weekly/monthly unlimited fare passes, possibly workplace/college/school discounted fare pass products if the difference isn’t made up by the school or workplace, and people making multiple destination stops on a single fare.) While I don’t have exact numbers, I’d imagine after considering all those factors we’d be looking at closer to $5 for each origin-destination trip currently, with probably $3.50 of that being subsidized.
I think it’s hard to see microtransit becoming truly competitive even with a $3.50 per ride subsidy, especially if we want to ensure that people in poverty are able to afford fares for their daily commute and we want to have some sort of living take-home wage for drivers.
This was one of the intended uses of dial-a-bus in the 1970s, at least in the Toronto area. Somewhere in my collection I have separate reports on this concept, one for a commuter rail station (Pickering GO) and another for a recently built suburban subway station (York Mills TTC). The idea was that on the home leg of the trip, you would dial a central dispatch number that would generate an on-the-fly route for all customers wanting to get to the station, and on the return trip, the driver would generate the route depending on the passengers boarding the bus. (Remember that this was before cell phones of any kind, although pay phones were widespread.) Obviously this never really took off and instead those stations ended up being served by fixed-route networks.
The optimist might say “it never took off because the technology wasn’t there to make it work, but now smart phones and GIS etc.”.
The pessimist might say “it never took off because it became too much of a pain in the ass for both the transit agency and for the customer.” (Even the most optimized routing is going to go so far out of the way that it only makes sense for the closest couple of pick-ups/drop-offs to the station.)
I am broadly agreed with you on this, but there is some subtlety that I think you are missing out on. I really am concerned in this discussion that we are looking at typical cases and typical applications but there are many ‘edge’ situations with special requirements.
There is also something a little circular in your criteria. There are ‘successful’ micro transit -esque schemes out there. I am thinking of Tel Aviv minibuses and Belfast shared taxis. You could argue that these are basically small fixed bus routes rather than micro transit and that because they make money that they don’t need subsidy. This is certainly true. But I can imagine there might be a city somewhere where there is no public transport, and that there is not sufficient demand or economic resources for a fixed bus, but a slightly subsidised minibus or shared taxi (subsidised through marketing or ticketing arrangements maybe) might be worthwhile.
Secondly, there might be a niche for services for people with disabilities. For example: http://www.disabilityaction.org/services-and-projects/transport-services/ . Of course these are not mass transit as you talk about it, but they form part of the transport ‘mix’.
If microtransit remains completely demand response, it’s probably doomed to remain below 3 passengers per hour. But if microtransit is operating in a place with demand spikes (for example, as a first- or last-mile connection to a transit station in a suburban office district), it could be implemented as partially demand-response and partially fixed-route behind the scenes, though the “user interface” for the customer would be demand-response-like either way.
For example, if you know that the 8:25 AM train will always have a handful of riders heading to Company K, you schedule a trip from the station to Company K at 8:30 AM. Anyone who puts in a ride request from the station to Company K (or somewhere on the way to Company K) near that time gets assigned to the fixed trip, while people heading to other companies get assigned on a more demand-response basis. And if someone shows up at 10:25 AM going to Company K, they are also mixed in with other trips on a demand-response basis.
The fact that customers are always going through the demand-response user interface gives service planners and dispatchers more flexibility to adjust which trips are fixed and which aren’t, and thus how many vehicles are out on the street. For example, you could use fewer vehicles on Fridays. You could schedule fixed trips during the summer when Company Z hires dozens of interns, and delete them the rest of the year. And on the days between Christmas and New Year’s, you could delete all fixed trips, put fewer vehicles on the road, and just give the rest of the operators holiday pay instead of incurring full operating costs.
Those sorts of changes are difficult to make on fixed-route service because you have to update published materials and conduct public outreach on each change, and customers would have a dizzying array of schedules to confirm based on their day of the week, their exact travel time, and whether Company Q just had a bunch of layoffs. The demand-response component of the service provides a “backstop” that ensures customers are taken care of even if one of these changes removes their fixed trip.
This will never be efficient as a fixed route in an urban setting, of course. But suburban office districts tend to be politically important despite the emptiness of the buses serving them, and a hybrid system could serve more passengers per revenue hour than fixed routes by knocking away some of those empty buses.
Matthew. As I said, the transit industry knows all about hybrids between flexible and fixed routes. Every conceivable kind has existed for decades, and the correlation between fixity and productivity is pretty strong.
I think we underestimate the degree to which the attraction of fixed route service is its predictability. All technology-driven ways to make it less predictable may be, on balance, a net negative for customer utility.
I agree with you about the predictability of fixed-route service being a huge advantage. But customers may be willing to trade off “available with predictable departure times, but only from 6:00am-9:00am and 3:30pm-6:30pm” for “available from 6:00am-6:30pm, but departure times may vary up to 15 minutes.”
What I’m not sure I agree with you about is “the transit industry knows all about hybrids between flexible and fixed routes.” Everything I know about transit planning beyond on-the-job training and personal bus-riding experience comes from your book, various other blogs (Pedestrian Observations, Seattle Transit Blog…), and the occasional useful TCRP report. (TCRP managed to make their site significantly less useful with a recent update though…) The agencies in my area have been running 100% fixed routes, backed up with 100% demand-response ADA paratransit, for as long as I can remember, and never the twain did meet. Institutional knowledge about “hybrid” services, if it ever existed, has been lost.
Maybe if the research about experiences with “hybrid” services was more easily available (perhaps you could write a blog post about how some of the hybrid services you designed worked?) it would be easier for transit planners and tech people to look back at those and see whether app-based automated dispatching really would help with efficiency. Then everyone might not feel the need to try it out again.
St. Joseph, MO merged their fixed-route and paratransit service into a deviated-fixed-route system years ago. For reference, schedules and a map are here: http://www.stjoemo.info/index.aspx?NID=312
and the NTD profile is here: https://www.transit.dot.gov/sites/fta.dot.gov/files/transit_agency_profile_doc/2016/70032.pdf
1. At 6.5 pax/hr, service productivity is less for fixed route service than comparable small car-oriented cities, but higher than paratransit service.
2. Route cycle times are very long and more circuitous than in comparable cities. As an example, if the routes were fixed Route 16 could probably make it between Downtown and the SE Walmart with an hourly cycle, or about 25-27 minutes one-way. All routes are instead scheduled on a bi-hourly cycle, meaning that this trip takes 50 minutes one-way. I suspect that some of the circuity could be removed if the agency were willing to tolerate split-directional timed transfers, (ex: outbound at :45 and inbound at :20) plus other scheduling tricks.
I commented three months ago on this subject but here again. My example might not (yet) be practical for a north American city but yet again, it is not an elite project, it could improved with technology (apps and driverless vehicles), it has been around for decades and is sucessfull.
Kiev has a train, metro, tram, trolleybus, bus, marshrutka, suburban bus, funicular and water taxi service. I found some numbers for the Kiev Marshrutka network, it has 1.1 million passengers per day or 24% of all public transportation journey, second to the metro network. Modal split for Kiev is 57% public transportation.
Using the Newark case study as an example, I’d like to see the work on how one gets from 7 pph to 3 pph. The AC Transit board memo clearly stated that some of the costs were allocated for technology and that many of the riders were redistributed onto major routes, and that there was a significant cost savings when the technology was factored out.
The route it replaced in Newark, and the similar route in Castro Valley which was replaced with Flex service, were circular routes connecting homes to a BART station and shopping/work points. Jarrett Walker has said many times no one wants to go in circles, yet this is what the routes were. Because of the roadway network, a standard grid bus service operating at the ideal 15 minute headway would not work well. Now the argument could be, from an equity standpoint, do these people “deserve” bus service even though they are high propensity, highly engaged voters. Alon Levy has used the commuter rail argument by saying that the upper middle class support does not transfer to regular transit, but given ballot measures in California require 2/3 vote you cannot afford to alienate any constituents. Microtransit can bring the perception that there is service to everyone, that no one lost out on service, while allowing agencies to concentrate on the routes that get 20 or more passengers an hour, as was the case in Newark, where ridership shifted to the major corridors with buses every 30 minutes or greater.
Many people recognize that, like senior dial a rides, it could be oversubscribed, but at least they can wait in a queue and get door to door service if they want or need it, while walking a mile to a bus stop if they need the certainty. The same could go for the situation previously mentioned on this blog of the Bellingham airport. Rather than sending a bus to make the deviation, you can pay a driver in a minivan a lot less to provide the critical connectivity to the transit system. Microtransit could be used as a backstop to canceling unproductive service in outer regions and hilly/suburban areas not suitable for line haul, “on the way” transit, thus speeding up or making more frequent line haul routes while not unserving anyone and engendering political backlash that prevents the larger goals Jarrett argues from being implemented.
Taxi/taxi-like dispatch just has to match two inputs; the number of riders and the number of cars available. When more cars are waiting than riders, the nearest ride request is matched with the nearest car. Even when demand is higher than supply, ride requests can be predictably queued in a first-in-first-served manner.
At the other end of the scale, fixed routes are also reasonably reliable. Even ‘bad’ fixed-routes that only come every 30-60 minutes or so will arrive within 5 minutes of their scheduled time 85-95% of the time.
Shared ride services have to not only match up ride requests and cars, but match requests in an efficient order. This makes shared ride service wait times particularly volatile.
My university runs an app-based shared-ride service overnight. On the handful of times I’ve used it, I almost always get an “estimated wait time” of 15 minutes or less. This seems great, but often the actual wait time is more like 30-60 minutes. The app could over-weight wait time, but that creates delays if the shuttle shows up before I’ve gathered up my belongings and I am ready to leave.
There are rumors that airlines must run some routes at a lost to get contracts with large businesses which lead to an overall profit.
Similarly, it may be that microtransit could serve as a backstop that gives customers confidence to eliminate car usage -if they have the assurance that the service is always available, they may be an incremental customer that drives demand to increase service quality elsewhere.
There also seems to be a belief that microtransit will reveal untapped demand for full fledged routes, spurring growth onward. However, analysis of other data sources probably reveals this with less cost. As a conditiin for TNCs to operate, couldn’t aggregate pick up and drop off locations be required to be reported?
Flex routes often grow into fixed routes, and we often used pickup/dropoff patterns to tell us what the fixed route should do.
But I’m afraid that permanence is a poor thing to count on for any low-ridership service, since most transit agencies see periodic culls of low-performing services.
Naturally, I’m pretty familiar with the Newark example, since I work for AC Transit. (My involvement with the Flex was marginal.) I think that, because it overlapped so much with other routes in the area, it can’t really be taken as indicative of either the success or failure of microtransit generally. At least 2/3 of it shared stops with other routes — so it’s no wonder that, as the report said, 50% of the ridership from the 275 went to those other routes instead. That’s not going to be typical of microtransit introductions, and it’s not clear what would have happened if those other routes didn’t exist either.
What I don’t understand is the comment in the other post about how since microtransit gets maybe 3 riders per hour, and “coverage-justified suburban bus” gets 10 riders per hour, that microtransit isn’t useful. I mean, if the goal is coverage, should it be judged by ridership?
I think there’s something to be said for saying “we don’t expect a lot of people to use this, but it’s there when it’s needed” — like insurance. Maybe the criterion for flexible service should be hectares served per service hour, rather than riders per service hour. Or “potential origins/destinations” per service hour, or something like that.
The Newark Flex route not only serves the stops that were on line 275, but it serves a lot of other stops in the area that once had fixed-route service on other lines, but that don’t anymore. There are a series of business parks on Newark Flex that have had fixed-route service added, and removed, several times since I started at AC in 2000. Maybe the flexible service can continue to cover those areas in the relatively rare times when they’ll be used, and along with that, cover other areas, at less cost than fixed-route would be for all that area.
Or maybe not; I don’t know. But I think it’s a valid question.
I do agree that “microtransit” is fundamentally the same as old-style dial-a-ride service. But there are advantages now. I wasn’t around in the 1970s when dial-a-ride was big at AC Transit, but photos I’ve seen show a centralized facility designed to take customers’ phone calls and then guide the bus drivers to the right places in the right order. Apps can’t change the fundamental geography, but they can remove some of that kind of overhead, and I do think that’s significant. Not order-of-magnitude different, but different nonetheless.
The app based routing also helps mitigate the traveling salesman problem and can provide real time info as to when a bus is coming, rather than for the driver to wait aimlessly for minutes or for a rider to miss their bus because they weren’t ready. Having fixed pick up points also helps in not having deviations into neighborhoods which reduce ridership. I would like to see Castro Valley run longer, and to see if perhaps this could be expanded in low ridership evening hours through a larger area.
“What I don’t understand is the comment in the other post about how since microtransit gets maybe 3 riders per hour, and “coverage-justified suburban bus” gets 10 riders per hour, that microtransit isn’t useful. I mean, if the goal is coverage, should it be judged by ridership?”
The bus’s performance matters only to the extent that both are giving similar coverage, or the bus can give more coverage per cost. But the starting point is how much coverage the city thinks it needs. Does the fixed-route transit reach most of it? Are houses so isolated that many people are a mile or more from a bus stop? The microtransit may add just enough to fulfill all the city’s coverage goals. Or conversely, it may suck up resources preventing they city from offering fuller coverage that fixed-route buses would provide.
There’s also the issue of total vs individual convenience as Jarrett mentioned. Fixed-route coverage is typically every 30-60 minutes and doesn’t go off arterials. Some flexible-route service also runs hourly, and may require a 1-hour or 24-hour reservation to deviate, while demand-response is “first come, first served, possibly in minutes. So that may give better service to fewer people, and that may or may not be acceptable to the city’s residents as a whole. The residents may not understand the issue and not realize how many people are being left out, and thus latch onto microtransit more eagerly than they should.
If the microtransit budget were unconstrained that would be an issue, but many agencies are looking at it to replace 3-6 pph service at lower costs (due to smaller vehicles and reduced O&M). Many senior dial a rides are oversubscribed and have to accommodate people at times that they don’t want to travel, or prioritize certain individuals over others (i.e. dialysis patients and people with medical appointments over shoppers).
Transit agencies are not the only entities–public or private–with an interest in “micro-transit.” In most major metro areas, there are many municipalities and public service agencies that run (or would like to run) services for their citizens/clients, independent of the local transit agencies. So far, there has not been a land-rush business from those agencies to establish tests/partnerships.
While technology is great, the most appealing mobility options involve no smart phones or apps. Just get up and go. No calling for a ride, no checking on schedules, just go. This is possible only for high density fixed route and for private autos. Also, as part of this, transfers are death to alternatives. If I have to get out of one vehicle to transfer to another, the odds of my using that option decrease dramatically.
James. It’s funny you mention transfers because they’re intrinsic to all the uses of microtransit as transit, as well as to effective rapid transit networks. Anything as tiny as microtransit will get its market from connections to higher-volume transit (rail and bus) to which you’ll have to connect. Transfers are also intrinsit to all highly efficient transit networks, and we see the result in ridership. https://humantransit.org/2009/04/why-transferring-is-good-for-you-and-good-for-your-city.html
There is a technological assumption made repeatedly by micro-transit proponents: A very good algorithm, given enough computing power, will be able to provide better of providing transit.
This sounds reasonable, but it is still an assumption. In contrast, the idea of a grid, as well as appropriate stop spacing has plenty of evidence to support it. We know that for a fixed route system, this is by far the most efficient way to deliver anywhere to anywhere transit service.
It is possible that a flexible on-demand system will deliver better outcomes at a cheaper price. But there is no evidence to support that hypothesis. My guess is because it simply isn’t true.
It seems counter intuitive. An extremely “smart” system that knows where everyone is headed should be able to improve upon a fixed route, fixed time system. This is obviously the case when load is very small (hence the appropriate term “micro-transit”). But it isn’t clear at all that an on-demand system is cheaper under even moderate load. Buses all going straight, never making a turn, stopping at a distance that allows for a good balance of speed and walk distance seems very efficient. Fixed time system running in that manner seems like a very good balance of cost and service. Your maximum wait time is not that big, yet you have one vehicle serving many trips. I think it is quite likely that a “smart’ on-demand system only makes sense when a “dumb” on demand system makes sense (under very low load).
Again, there may be a study out there supporting this idea. But almost all of the arguments for micro-transit ignore this idea, and almost immediately go into the benefits (e. g. front door service) without any of the costs (delays caused by front door service). They don’t do modeling comparing the various systems. My guess is the cross over point occurs very quickly.
One of the arguments often used to support on-demand service is that it will be much cheaper when cars are automated. But they rarely, if ever, consider the effect that automated fixed route transit vehicles will have. If anything, I would assume that fixed route service becomes appropriate in more situations, not less. A fixed route, fixed time system is always more efficient, but if buses run almost empty than it is expensive. If they run every hour, it is inconvenient. Automated vehicles would therefore increase the number of situations where fixed route systems are both convenient and cost effective. In that sense they would operate like a huge grant — allowing for the virtuous cycle that comes with increased headways to flow to areas that would otherwise be very expensive to serve.
Furthermore, when the cost of driving is eliminated, the biggest cost becomes the vehicles. Specifically, the number of vehicles needed to carry peak ridership. Again, it seems like the best system is one in which those vehicles are close to full, and aren’t spending much time stopped or making detours, so they can drop off passengers.* A fixed route, fixed time grid with appropriate stop spacing seems like the right choice. It may be inconvenient to ask people to walk a couple blocks, or take two buses to get where they want to go, but it is likely the best value for the money in most cases.
JW: “As I explored here, microtransit seems to consist of:
flexible “on-demand” routing, an idea that transit agencies have known about, and experimented with, for decades.
subsidies of privately provided services by a transit agency, which has been happening for decades under a range of contracting arrangements.
the use of apps for hailing, navigation, and payment.”
No discussion of benefits of choosing vans over buses. Right. Because very little benefits exist from one design over the other. The only way to create “micro” design advantages is to drive narrow cars instead of standard size cars. Proof: If bus and van width don’t matter, why not make buses and vans twice as wide as they are now to carry more passengers? Further, why not provide side-seated bikes instead of single-width bikes for bike share? The answer is, obviously, because they would cause unnecessary congestion. Width is key design element not discussed.
As Transportist PhD Transportation Professor David Levinson said recently in a SideWalk talk: “Moving towards the one-passenger vehicle has huge benefits, and that’s the biggest challenge we’re not recognizing.” https://transportist.org/2018/02/12/sidewalk-talk/
JW: “Fixed routes are spectacularly cost-effective investments compared to almost any flex-route option” However, they are not as popular. Bike and car share would be less expensive if bikes were only allowed in bike lanes or on dilineated tracks, but that takes away the flexibility.
JW: “Recent declines in bus ridership are triggering all kinds of triumphalist claims from people who want to sweep fixed routes away, or at least shift resources away from them as the microtransit movement proposes.”
A lot of the decline in popularity is due to static bus design. Clearly, bus design is far too long and wide to efficiently work in cities. “Micro” transit – really “van” transit takes less road length, but doesn’t help relieve congestion.
My conclusion: on micro-transit: it’s good to transition away from buses to vans due to vans more flexible desin, but the real change in our lives, like David Levinson suggests, is transitioning from side-seated auto/van/bus design to single-width.
On vehicle width, that’s part of the reason why heavy rail has more passenger capacity then light rail or streetcars, because the trains are wider.
On the supposed advantages of vans and small vehicles over buses- no, that’s not true, for reasons RossB details above in other responses. Space efficiency is scaled like this, from most efficient to least efficient: Trains>buses>cars.
Maybe the solution would be to increase the frequency of fixed route services by replacing large buses with minibuses. This was common in the UK in the 1980s when buses were deregulated ad lost their subsidies. Bus companies reacted to protect their new commercial urban networks by increasing frequencies, hoping to discourage new competitors from setting up.
The pioneer city for this trend was the city of Exeter. The following article describes what happened. Note that the increase in labour costs was mitigated by paying a lower hourly rate to minibus drivers, and recruiting the extra staff from outside the industry, mainly customer focussed retail staff.
Extract – Not carrying empty seats around:
Harry explained to Mark Burton of the Coach & Bus Manager Magazine in March 1990 “On full current cost accounting – providing for depreciation for replacement of the asset and meeting financial charges we are still talking 70p per mile costs for a Minibus in Exeter as opposed to £2.00 per mile costs in similar situations for a Double Decker – the great thing is you are not carrying empty seats around – there’s no money in empty seats.” When pushed Harry admitted that the real saving in Staff costs following the switch from Double Deckers to Minibuses was initially ‘exactly 38%’ Harry reflected back and, to quote again from the Coach & Bus Manager article stated “If I remember rightly the initial figures were around a 250% increase in passengers, a 200% plus increase in revenue but only a 180% increase in costs.”
(arriving here from a recent post) Is the fixed-route versus microtransit not a false dichotomy? Rather, as you indicate, the fallback when fixed route buses fail is to personal cars. The progression, then is maybe something like:
Personal car < car share < taxi/uber < microtransit < fixed-route transit < light rail < rapid transit
…, since your argument that microtransit seems sound insofar as the costs compared to fixed transit are too high, but here I am, living in a low-service area in London Zone 2, often choosing car share and Uber rather than buses which are sufficiently infrequent that they're unreliable, and either take fixed routes that are often extremely congested or circuitous routes that suffer the worst of the microtransit frustrations.
Surely, reliable fixed transit is preferable, but in the meantime I'm often paying 2-5x public transit rates for the convenience. And frankly, I'm one of the progressive ones. Most of my neighbours own their own cars, or are trapped spending a huge chunk of their day on sub-par transit (which given the demographics of the neighborhood amounts to a manifestation of systemic classism and racism).
An awful lot of conjecture and back of the napkin analysis here in the comments section. What Mr. Walker says makes a lot of sense. I wonder if there have been any serious studies or trials on the use of Microtransit to supplement traditional transit. Geometrically, it may not make sense, but as with other technologies, there may be unexpected outcomes we can’t necessarily predict based on the traditional applications we are accustomed to. I agree the “app” component is probably the most novel part (which is to say, not particularly novel at this point). A virtual rickshaw, if you will. In some countries where transit is fairly developed you see other non-traditional modes of conveyance (the moped in the Caribbean, for example or the rickshaw, or the guys in the back of the pickup truck) which takes people directly to their destination affordably. At any rate, I think it warrants studying further before we reject it entirely.
As Jarrett Walker details in the links in the article above, there have been some trails, and they turn out to not work that well in most cases.