Via Van (image: Sound Transit)

King County has piloted several on-demand services that connect people with transit hubs. The services address first/last mile access issues up to three miles around transit centers. Recent data indicates that Via continues to perform well in the Rainier Valley with growing ridership and progressively declining average costs. Meanwhile, the Ride2 services in West Seattle and Bellevue have seen stubbornly low ridership and higher per-rider costs.

“Via to Transit” launched in April as a partnership with on-demand transportation provider Via. The service mostly operates in the Rainier Valley connecting riders to light rail service at Mount Baker, Columbia City, Othello, and Rainier Beach. A more limited version is available in Tukwila. Early results were promising and have gotten appreciably better as the program reached the six-month mark. Via now serves almost four riders per hour at a cost per rider just over $10. That’s above the average of Metro services, but is declining as ridership scales. It rates well against the coverage routes that are the more immediate alternative.

Ride2 launched in October 2018 around the Eastgate P&R in partnership with Chariot, a Ford subsidiary that withdrew from the market earlier this year. Operations transitioned to Hopelink in February after Ford quit the business. It now serves just over two riders per hour at a high $35 cost per rider. On the bright side, it has displaced driving, with a majority of riders surveyed saying they used to drive to either the transit center or their final destination. Service levels are also high with an average wait time under five minutes vs a program target of ten minutes.

Ride2 was extended to West Seattle last December as a one-year pilot, serving the bus hub at Alaska Junction and the Water Taxi terminal at Seacrest Park. There has been little rider interest, and the service carries fewer than one rider per vehicle per hour at a cost of over $80 per rider. The pilot seems unlikely to be extended.

A March 2018 map of Metro Community Connections services (image: Metro, click to enlarge)

Metro’s Community Connections program encompasses an array of alternatives to fixed route service. Most are in communities that lack the land use or density for fixed route services to be productive. Because it is conceived as a program for innovative services where traditional transit routes are not economically viable, some failures are to be expected. While none of the services have large ridership (otherwise there would be a bus), some have done remarkably well at cost-effectively connecting riders to transit. The best of the Community Shuttle routes operate at costs per rider comparable to the busiest fixed routes.

Via and some of the community shuttle services are more than cost-competitive with fixed routes, making them excellent candidates for ongoing funding after the grants run out. The economics of such small vehicle transit is tied to their low operating cost, so it’s likely they will continue to be implemented via partnerships with private companies. It’s not always clear why some routes work out and others don’t. Ride2 in West Seattle probably aligns too closely with fixed route service to the ferries. But why Via in the Rainier Valley and not Ride2 in Eastgate? Expect continued experimentation as long as the budget allows and while Metro remains capacity constrained on fixed routes.

Performance data for Metro Community Connections services (image: Metro)

85 Replies to “Via shuttles performing well; Ride2 not so much”

  1. Literally reading this from inside a via van! The service has been great. It is reliable, quick and I’m surprised they have been able to wrangle the costs down to $10.

  2. For many of these community ride services (I’m thinking specifically of the Burien, Mercer Island-Seattle, Des Moines, Snoqualmie Valley Shuttle, and Snoqualmie Community Shuttle), what is the main difference between these and a DART route? These services seem on their face like exactly the same thing, with a fixed route plus a flexible service area (except Des Moines, which is purely fixed route). Is there something they gain by not calling it DART?

    1. Perhaps the DART moniker has run its course. “Dial A Ride Transit” seems very 1900s.

    2. Metro’s contract with the ATU (driver’s union) caps contracted service at 3% of all service hours. By not calling these shuttles DART routes and instead calling them “alternative service” or “community connections”, Metro can exclude them from the 3% calculation.

      1. Ah, sneaky. I wonder what TRU thinks of these services. I know they fight like heck whenever Metro talks about contracting out service.

    3. It doesn’t have a fixed route. DART runs hourly between a transit center and (A) a shopping center, (B) another transfer point, or (C) a residential loop. Thus it gets casual riders who the schedule works for, who are making spontaneous trips and maybe don’t use apps or don’t want to wait for an unscheduled pickup. I’m like that, I’m only going where the fixed route goes, and I like having a scheduled time rather than using an app, so I’m more likely to ride it if it has a fixed component. I rode the Snoqualmie Valley Shuttle that way. It has flexible service areas at both ends (Snoqualmie and Duvall), and I think you can order a deviation an hour ahead. Completely dial-a-ride routes require a reservation a day ahead, or at least they used to. Via and Ride 2 are like dial-a-ride with a shorter reservation wait (typically 10-45 minutes I guess), while semi-fixed routes like the Valley Shuttle come during their scheduled runs. I think semi-fixed routes are the best option, and would like to see them half-hourly or 15 minutes rather than hourly. That may not be feasible but it could be a long-term goal.

  3. My coworker has mostly given up on Via after being declined for rides several times, apparently because they are full. She walks to the bus instead. I wonder how it compares cost wise to adding a few more trips on those fixed routes. Since it on-peak, it may still be cheaper than Metro acquiring busses and hiring drivers. However, Via may only be cost effective because they are so full they are turning people away.

    1. It would take increased service on the 50 and new bus service on east-west arterials in Rainier Valley to do what Via does. Or allowing private jitneys.

      1. Private jitneys meaning share taxis, colectivos, matutu, whatever you want to call them. Private versions of DART: semi-fixed route, small vehicle service. In a sense this is what some implementations of private microtransit do, except that microtransit doesn’t really make the same kind of scale because it doesn’t go poaching riders from bus stops the way a jitney does.

        Generally speaking, we don’t have this kind of service on a large scale because of regulations and an already-good-enough public system that obviates critical mass of potential rider demand.

    2. Metro’s long-range plan addresses all of those. It just needs funding. The only reason Metro isn’t addressing them now — and why a Graham route isn’t planned until the 2030s — is it doesn’t have the money. The only reason Ballard Link isn’t opening until 2036 (or later) and not in the 2020s is ST has to wait until sufficient revenue accumulates (and because it prioritized West Seattle first). Metro’s plan may not be perfect but it’s a plan, and now is the time to evaluate whether it comprehensively fills the district’s transit holes. And whether the routes connect locations to the most common destinations.

      For instance, route 1997 (Beacon Hill-Judkins Park-23rd-Jackson-MLK to Madison), route 3033 (Fairview-Boren-Yesler-Leschi-Mt Baker), route 3034 (West Seattle-SODO-Columbian Way-Columbia City-38th Ave S (northward)-Mt Baker), route 1039 (Westwood Village-S Cloverdale St (South Park)-Georgetown-Swift Ave-Graham St). Do these routes comprehensively serve southeast Seattle between Yesler and Othello? Every house? Do they go where the people in between want to go? Can everybody get to Judkins Park station reasonably? I don’t know enough about the area to fully evaluate these, but somebody needs to. The answer will tell us whether these routes should be changed, whether additional routes are needed, and whether there’s a legitimate role for Via at that point. And if so, which trip patterns are Via’s core mission and justification. For instance, Via is not justified near the Othello corridor if the Othello route is sufficiently frequent — but what do “near the corridor” and “sufficiently frequent” mean? What’s a reasonable expectation for a person living a few blocks from Othello, who may be facing a 2- or 3- or 4-seat ride trip with additional waiting and walking in other parts of the trip?

  4. The most obvious reason why so many fewer people are using it in Eastgate than in the Rainier Valley is parking. Rainier Valley stations have very limited parking, creating a stick incentive for people to try alternatives. Eastgate, on the other hand, there’s no reason to bother with a service like Via as long as free parking exists in the garage. That limits users of the service to either around 9 AM, after the parking is at risk of being filled up, or for reverse commutes to various suburban office parks, or special situations, like someone needing a ride because their car is in the repair shop.

    1. Also, a significant difference, I would assume, is that Via (in Rainier Valley) operates about 20 hours per day, the same as Link. This allows for many more types of riders the opportunity to rely on it. That being said, I’m curious to know what its ridership is like during peak times vs off-peak.

  5. There might be a perception in quality when it comes to Rainier valley vs West Seattle. My friend, who lives near Graham takes it all the time. Loves being dropped off at the train. He would not take it if it dropped him off at a bus. There might not be any times savings or loss. But that does not matter. Getting dropped off at Link is better to him.
    My perception is different than his, since I take busses all the time. Most 1 mile trips would be more covenient for me even if it only goes to a bus. Even if the logic is flawed, the perception is there for my friend.

    1. The thing in Rainier Valley and nearby is that there are North-South bus routes on all the S Seattle arterials (including even Swift Ave, kind of). Unless you’re down by Seward park, you can walk from anywhere on Graham to one of the 107, 36, 106, or 7 pretty quickly.

      Imagine you’re at 42nd Ave S and Graham, for example. Waiting for a ride to get to the 7 bus is pointless; you can already walk there in 5 minutes or less. But if you can get a quick ride to the Link, you could turn a 40-50 minute trip downtown on the slow, oft-stuck 7 into a consistent 35 minute one by Via/Link.

      (Personally I would just walk the half mile to Othello Station, but a lot of people won’t walk that far to transit in my experience. And you can imagine that Via gets even better if you’re 1 mile from the Link instead of 0.5.)

      1. I’m not sure JJ’s point is that there should be Via service to buses in Southeast Seattle. Quite the contrary: no one would use it because they prefer train to bus.

        Anyway, there are no expresses in Rainier Valley that are comparable to the routes that Ride2 serves. (Not even the 9X, which Metro has been trying to cut for years.)

  6. Does Via keep stats of rides? If so, I’d like to know what the average ride looks like. What’s the average number of blocks it’s taken? And let me stop the anticipated comment section exaggerator who claim they have several disabilities and takes Via twenty blocks over a dozen hills … I said the AVERAGE trip. Does anyone have that stat?


    1. I would guess the average trip to be about 1 mile, replacing a 20 minute walk. 10 minutes wait time +10 minutes ride time takes the same 20 minutes. Density drops off quickly east of Rainier, so the number of people within the service area more than 2 miles from a Link station is quite small.

      I would also guess that most of the people are able bodied and more than capable of doing the walk, but are simply lazy.

      1. I would guess the average Via trip is around a 1/8 of a mile, which is how far I have to walk to get to my nearest bus stop.

      2. I thought I read somewhere that Via won’t even do trips less than 1/4 mile. So, definitely more than that.

      3. Not everyone walks a 20 minute mile, or is able to comfortably walk a mile at any speed. People may also choose to ride rather than walk for reasons other than physical ability. Assuming that “most of the people are able bodied and more than capable of doing the walk, but are simply lazy” is incredibly judgmental.

      4. My experience with Via is phenomenal.
        My choices:
        •18 min/1.2 mile walk west to LINK (half is a Hill)
        •8 min/.5 mile west to Metro 7 (hill)
        •8 min/.5 mile east to Metro 50 (level) to LINK (total 18 min)
        •8-9 min wait for Via, walk 1/2 block, ride 8-10 min to LINK

        On a decent day, I’m walking to LINK. If it’s raining and/or with parcels, I opt for 50, but 75% of the time it’s 20min away.

        With parcels or inclement weather, Via has been the reliable option when walking east (wrong way) to wait 20min for a 50 isn’t reasonable.

        I always look at what my last LINK out to reach Via before last pick up is and I’ve been fine.

      5. @David – that “incredibly judgmental” comment is unfortunately not uncommon on these boards. Thank you for calling it out.

        (said as someone who does a walking commute of 3 miles/day, rain or shine… not that it should matter in the least)

  7. Dan, the piece could mention the role of the Seattle TBD in providing the service subsidy for West Seattle Ride2 and Via. Measurement of success could include consideration of the opportunity cost, even in the face of bus constraints, and the net impact on network productivity. The STBD could have funded the off-peak adds of September 2019 earlier in March 2019 or transit capital. Via will have trouble scaling. First, the vehicle capacity is small. Second, intending riders will not like being taking through the deviations to pick up subsequent riders. This could cause them to catch later Link trips than they could have if they had just taken the frequent bus lines instead; routes 7 and 36 have 10-minute headway; routes 14, 106, and 107 have 15-minute headway. Via service is not cost-effective relative to SE Seattle local routes. there is no other local service in Snoqualmie. It may be better to use Via type service in areas without local service. Even the best line in the chart, Route 630 between MI and First Hill, just seems transfer adverse; strategically, why have a transitvan, attracting a passenger load less than 20 duplicate routes 550, 554, and 216? at the time of service reductions and today, would not improved Route 204 service be more important?

    1. The 630 provides direct service to First Hill that went away when the 205 was cut. So in that regard it is allowing people to avoid a transfer downtown, but the transfer penalty getting to First Hill from Mercer Island is big enough that direct service at this scale is not unreasonable. Transferring to the 550 is much faster for getting downtown.

      The 204 was changed from mainline Metro service using 30/35-foot coaches to DART service in Spring 2019, and used those savings along with deleting the 201 to add Saturday service and change peak frequency to 30 minutes (hourly midday). So the 204 has already been improved. A longer span in the evening would help; service currently ends before 7:00pm.

      1. yes, routes 203 and 205 were weak and were deleted in the reductions. yes, reaching First Hill takes a transfer. should all 39 cities have a direct one-seat ride to and from First Hill? the frequency and span of Route 204 remains inadequate.

      2. If people weren’t too lazy to just good it up the hill, they wouldn’t even need a transfer.

        I agree, singling out Mercer Island for special one-seat ride privileges to First Hill looks suspect. Especially when the 204 shuts down before 7 PM.

      3. Maybe we should take this as a warning that we urgently need to speed up transfers to First Hill – which we do.

      4. RapidRide G won’t have convenient transfers to Link until 2035 at the least.

        Meanwhile, only the infrequent Route 9X connects Judkins Park, Mt Baker and SE Seattle with First Hill even though the distance from Judkins Park is only just over a mile. This is one reason why rerouting Route 4 (trolley bus wires needed) just a few blocks to Judkins Park would be very beneficial.

      5. The G Line alignment has longish walking transfers with Link. it does not serve the Capitol Hill Station. will it get its FTA grant? it crosses 3rd Avenue in between Link stations.

      6. Mercer Island is paying for (most of) the 630, so if that’s what they want to spend their TBD money on, great. Of course if I-976 holds up, that funding disappears.

      7. The downtown Link stations are close enough together that we’re talking about about a block. Walking one block is not a hardship.

        If you’re trying to ride the bus, but can’t walk even one block, you’ve likely got bigger problems.

      8. Serving the Madison corridor is incompatible with serving Capitol Hill Station. If it goes up to Capitol Hill and back, it’s a time-consuming detour, and it would leave out Seattle University and Swedish First Hill, which only the G serves. The 2 will be rerouted to Pine-12th-Union to replace the 11 and 49, so it won’t be available on Seneca-Union.

      9. The ultimate problem is that when the DSTT was designed, University Street Station should have been a bit further south (Madison Station). Westlake Station should have had an additional entrance on the south side of 3rd & Pike to fill the gap, and also to give a closer entrance to the eastbound Pike Street buses and Pike Place Market. Or add a station for Madison. Madison Street is the primary corridor in central downtown/lower downtown/whatever you call the area between Pine and Yesler, so it should have a station. Not little University Street. That would also obviate the confusion with the University of Washington. (“Is this where the main university in the state is? It says university.”)

      10. eddiew is right: stopping 2 blocks away from light rail is not the same as serving light rail. For usability and accessibility, every bus should have a same-block stop as a Link station entrance somewhere on its route. It’s great for reliability that the G has a nice, straight route, but it’s not well-integrated with the rest of the transit system.

      11. Won’t the G bus have to go around the block to turn around, anyway? If so, perhaps a stop could be added a block closer to University St. Station at essentially no additional cost.

      12. The G’s 3rd & Spring stop is one block is one block from the 3rd & Seneca entrance to University Street Station. The G will live loop downtown so that everybody both eastbound and westbound can use that station if they don’t want to walk an additional block from 3rd & Madison. The turnaround on 1st Avenue in the streetcar/bus-exclusive lanes will supposedly be fast enough to avoid people cursing at the traffic.

      13. And, it’s better than the current Route 12 service where you have to walk further to Marion Street, and it can take half an hour to get from 3rd Avenue to 12th Avenue in the PM peak, and service drops to half-hourly evenings and Sundays. The comparison is not only to an ideal network but also to the current network. A better network will be better, even if it’s not perfect.

      14. I’m also amazed at how no wheelchair user will be able to transfer from either Madison or Seneca to or from Link near Third Avenue because the stop is too sloped. To me, RapidRide G is inherently flawed until 2035.

      15. The steepness of the hill is what it is, and it’s not reasonable to add unnecessary detours to the bus, just to move the stop one block over to avoid it. In any case, if the wheelchair has a battery-powered motor, you should be able to just power up the hill yourself, and not bother with the G-line at all.

        Worst case, you can always switch to the First Hill Streetcar at International District Station, instead of riding the G-line. The streetcar has many flaws, but the one thing it does do better than a bus is easy loading and unloading of wheelchairs.

        Even in 2035, I don’t see your issue really being addressed. The hill is not any less steep over at 5th Ave. than at 3rd. And, to reach the new tunnel, you’ll have to switch trains at International District Station for an additional transfer.

        I think the best hope for wheelchair users is simply cheaper prices and longer battery life for the motorized models. Downtown->First Hill is really not that far in distance, so simply motoring up the hill on battery power ought to be as good an option for last-mile travel as anything.

      16. “Won’t the G bus have to go around the block to turn around?”

        It turns around on 1st in the streetcar lanes, with a shared streetcar/G station.

    2. Measurement of success could include consideration of the opportunity cost, even in the face of bus constraints, and the net impact on network productivity. Via will have trouble scaling.

      Exactly. This is considered a stupendous success since one particular area gets almost four riders per hour in the city. That would be considered a horrible failure if a new route did that and yet that is the best part. Furthermore, there is no evidence that this is actually increasing overall ridership. It is quite possible that it is simply poaching riders from the 50. Either way, there is no reason to assume that this is actually a good value, since this doesn’t scale.

      In contrast, it is pretty easy to see how spending money on borderline routes helps the entire system. Let’s you run a truncated version of the 17, that goes from Sunset Hill to Ballard, makes a live loop and returns. Now riders in that area are connected to the system, and have a bus to the nearest destination (Ballard). This, in turn, increases ridership on the 40 and D, making it easier to justify increases there. This, in turn, leads to more riders on those buses as well as the truncated 17. This is the virtuous cycle of increased service.

      In contrast, these microtransit options don’t even offer that. Ironically, they only connect with Link, and the dent they make in ridership to Link (if there even is one) is minimal, and not even noticed. You aren’t going to run more big, expensive trains just because a few people got there by minibus. This is a bad value.

      1. Running 35-foot coaches on neighborhood circulator routes would be better than Via, but it’s hard to fund and staff them, and even harder to get people to ride them.

      2. People would ride them if they were more frequent. If a low-volume route comes every 15 minutes, it’s relatively easy to take it whenever you’re ready to travel. If it comes every 30 minutes, it’s harder to fit your schedule around it, and it makes the door-to-door length of a total trip longer, and means you can fit less activities into the day. If the bus is hourly, it makes it exponentially worse. All these levels are the tipping points for some users. So when you increase a 60-minute route to 30 minutes, you get more than twice the riders. And if you increase a 30-minute route to 15 minutes, you get more than twice the riders again. It may not happen immediately but it will happen over a few years as people start to realize the service is there, how it can benefit their specific trips, and as people make decisions about where to live, especially people who are willing to take transit if it’s sufficiently frequent. Those people preferentially chose locations with good transit, while people who drive everywhere don’t weigh that factor. That may not affect things in the short term but it will in the long term.

  8. If we’re asking why the Via program has outperformed the Ride2 pilot, I’d guess a lot of it comes down to similar kinds of “fundamentals” that drive performance of fixed routes: land use, layout, quality of transit connections, and demographics.

    The fundamentals of the Rainier Valley for transit performance are mostly really good! This isn’t a paper phenomenon — the Rainier Valley has been home to some of Seattle’s most popular mass transit for decades! It serves to reason that there is serious demand for transit connections to Link, our best and most popular transit route. The big problem for connecting routes, largely running east-west, is layout. Fixed transit routes perform well when they aggregate “transit demand” (i.e. demand for the kinds of trips transit serves well) along a corridor. The RV has the transit demand, it’s just hard to find good corridors. Something like Via aggregates the demand without being so particular about corridors.

    Eastlake and West Seattle just have worse fundamentals beyond layout. Even in West Seattle density drops off really quickly outside of the C Line’s existing corridor. The quality of the connections is alright, but not as good as Link. Alaska Junction itself is a pedestrian destination in its own right but Eastgate isn’t. There’s just less transit demand in these places and more people have access to cars and parking so they’ll solve their layout issues that way.

    So maybe the Rainier Valley is the exceptional area where paratransit isn’t just a least-we-can-do coverage solution. Even at that, according to accounts here, it runs up against capacity limits without spectacular efficiency, but it seems to be providing a real benefit to people that would be hard to provide in other ways.

    1. The RV has the transit demand, it’s just hard to find good corridors. Something like Via aggregates the demand without being so particular about corridors.

      Yet even in the Rainier Valley, it performs very poorly. At less than 4 riders per hour, it is well below the 50 (19 per hour), which is well below the 106 (31), which is well below the 7 (47). Those numbers are for off-peak; the numbers are of course higher for peak. In other words, if the best microtransit was a fixed route, it would be by far the worst microtransit in Seattle, several times less than routes in the area.

      It is quite likely that additional coverage routes — in various parts of the city — would perform much better (since literally every existing coverage route does). There is nothing really special about West Seattle, or this service. It is simply an area that could use an additional coverage routes — especially east-west — just as other parts of the city could. Providing such service comes at a cost — either you increase taxes, have less service elsewhere, or in this case, choose to pay the drivers less than a living wage.

  9. I think it’s really hard to evaluate or compare these services without time-of-day information, and without knowing what users have with them. Via is great when one returns from a plane trip and gets off Link at 10;30 pm with luggage, for example.

    I’ve known of a few transit systems that operate services differently — doing things like flex-routes or stopping anywhere (rather than at a bus stop) — after 8 or 9 pm. A more detailed breakdown would explain the niches that the services serve.

  10. I see a successor to Ride2 bring more successful when East Link opens in 2023, and the hub is South Bellevue or Mercer Island. The service seems ideal for situations where reliable (almost to the exact minute) and high-frequency transit can be used to a point where actual schedules are not a concern and the significant issue is just the last mile or two.

    1. Agree. But, the issue still remains that, as long as there is ample parking at South Bellevue, there is no reason to bother using it, except for reverse commuters, or people whose cars broke down.

      1. This is why I wonder what percentage of users are coming from the airport or Amtrak or Bolt. Overnight parking probably won’t be allowed.

        I could even see an argument that the Port should help pay for these services to reduce the horrible airport congestion that exists now at SeaTac.

  11. With students getting free ORCA passes, and some schools being next to transit stations, can students use Via as a free taxi service from their home to school and back?

    1. Yes, Your Trollitude. They’re not going to police whether you actually get on transit or walk somewhere. Is that a problem? Going to the train and going to the station area are the same thing transportation-wise. And it remains to be seen whether any students actually do this. It’s the kind of thing that may be more of a theoretical problem than an actual problem, and that’s where dwelling on it is trollish.

  12. It’s worth investigating how much of the difference is due to the location, how much to company-specific issues, and how much to publicization.

    Rainier Valley is a pre-automobile district, so it has a tradition of transit use. Some old families remember the streetcars, or their family had only one car, or they are more open to taking shuttles, or they moved there because of that culture, or many in the neighborhood are poor and can’t afford a car.

    Eastgate/Phantom Lake was developed in the 1960s and designed for cars. Most people moved there assuming they would drive everywhere, so the percent of residents who would ever consider taking a shuttle or bus is much lower. That plus the fact that Eastgate P&R is not a strong draw: off-peak buses are every 30-60 minutes, there are no shopping or jobs there except the college.

    West Seattle was historically a transit district but its current residents are more affluent like Eastgate so they can afford to drive everywhere and bought their house assuming they would do so, so it’s in between Rainier and Eastgate in terms of the number of expected riders. If it’s lagging behind that, then my first question is how many residents know about Ride 2? Maybe it was publicized less there. I don’t think it’s closeness to fixed routes, because a lot of West Seattle is near steep hills, and if there is a bus it’s infrequent, and it’s a 2- seat ride to get from one part of West Seattle to another. For instance, in north Admiral the 50 and 128 runs every 30 minutes but recently they were 60 minutes (and will probably be again if I-976 goes through or the TBD isn’t renewed), and neither route goes to a major destination. In both cases you can ride them for 30 or 60 minutes and end up in SODO, White Center, Seward Park, or TIB — all of which have little there or you have to transfer to something else to get anywhere. So I don’t think proximity to fixed routes is what’s crushing ridership; it’s probably lack of publicity, lack of desirable destinations on the service, or unwillingness to take public transit/shuttles of any kind. (By “public shuttles” I mean that they might still take Uber, which doesn’t require transferring, is private, has a nifty app, and is well known.)

  13. Interesting piece.

    “The economics of such small vehicle transit is tied to their low operating cost, so it’s likely they will continue to be implemented via partnerships with private companies.”

    Do you happen to have any comparative info on the fully loaded driver costs?

  14. If Via is so necessary, how come during the rollout of Central Link, how come none of you “transit experts” said, “Hold on a minute! We have a big problem here! It will be impossible for people in the Mount Baker, Columbia City, Othello, Rainier Beach, and Tukwila neighborhoods to get to or from the Link Stations!” You all jumped on the micro transit bandwagon after the fact. If it’s such a vital service, why didn’t you point out a micro transit necessity before Via?

    1. We tried to get more coverage service to the stations. Metro had a new route on Othello to Seward Park and it was going to be more frequent, but the community clamored to keep the 42 and a Beacon Hill station-Mt Baker station shuttle, and Metro took the service hours for them out of the Othello route. Since then the 50 has replaced the Othello route and become more frequent (30 minutes now). There may be other needs for coverage routes in the valley; Metro should clarify this. It does have some ideas in its long-range plan.

      Via came about because of external pressure and partners. It was a pilot and not necessarily as effective as coverage routes would be. But it’s what we have now. There’s an ongoing debate about the merits of microtransit vs coverage routes; most transit fans are on the side of coverage routes. Which you’d realize if you weren’t trolling.

      1. Care to say where micro transit will be necessary and vital in the future? Which neighborhoods on which future light rail lines?

      2. None. How’s that for an answer? There are neighborhoods I know well like the U-District and Capitol Hill where I know what most of the trip patterns are and the gaps in the transit network. For instance, from Bellevue & Denny to Swedish First Hill and Swedish Cherry Hill there are no good bus routes; there’s a transit hole between SLU and those hospital areas. If the 11 is deleted with RapidRide G there will be another transit hole around 12th-17th & Madison going west.

        Rainier Valley I don’t know that well; I stick to the commercial districts, except one friend in Rainier View near the 107. So I don’t know what specific transit holes exist; I just have a general wish that everybody can get to the stations without an unreasonable burden. I look for input from others on what exactly that should be, and I look first to coverage routes, and microtransit as a last resort.

        I used to be very pro-microtransit before it was called microtransit in areas like Issaquah, eastern Mt Baker, etc — areas that may be too low-density for comprehensive bus routes but should have some last-mile access. But then I saw Jarrett Walker’s research at about how even lightly-used coverage routes get five times more passengers per hour (10 passengers) than on-demand neighborhood taxis (2 passengers), so it’s a more effective way to spend resources. That makes it worth having coverage routes even more extensive than they are now, and bumping up the frequency to 30 minutes or even 15 minutes. So now I think the burden of proof is on microtransit. And I can’t answer your question without knowing what the coverage routes will be, should be, and what remaining trips are vital and not covered by them.

      3. Before you say, “But Mike, Via is better than that according to the article”, those are nationwide averages. It’s very interesting that Via is almost 4/hour. That’s worth watching to see if it grows, and it may turn out that Rainier Valley is an exception to the trend and that microtransit could be cost-effective there. (Maybe due to its specific size or density or trip patterns or willingness to use transit.) We also need to know what percent of Via trips have no reasonable transit option vs how many have a parallel bus the person is too lazy to use. Or whether the Via trips are because the bus’s frequency is unreasonable. All three cases are possible. If it turns out that most Via trips are near a parallel bus route one block away and the bus frequency is reasonably close to the time of the person’s trip, then we’re wasting money on Via. Conversely, if it turns out that most Via trips are filling legitimate transit holes, then it may be worthwhile.

      4. That’s worth watching to see if it grows, and it may turn out that Rainier Valley is an exception to the trend and that microtransit could be cost-effective there.

        The four riders an hour only looks impressive compared to other microtransit, which is worse. If the problem turns out to be the low frequency on the 50, then increasing frequency on the 50 (which gets about five times the ridership per hour) would be way more cost effective. If it turns out the issue is lack of coverage, then running buses on the major cross streets (Graham, Orcas, etc.) would be a better value. If there are people with mobility problems, then they should use other services. If it turns out that none of that is the case, then this is just paying for front door service for a handful, while hurting service for way more.

      5. There is one thing that Via does do that improving fixed-route service doesn’t, is that it provides data on where people accessing Link are actually coming from, which could be used to help guide future investment in fixed-route service. For example, if the Via users mostly live along the 50, boost frequency in the 50. If they live along Graham St., add another bus there. Also, what times of day are these people riding?

        As long as Via is jump a temporary pilot to collect the data, with the intent of transitioning to fixed route service as soon as practical, it can gather the data at less cost than adding a whole bunch of different fixed-route services for the pilot, which may or may not be used.

        The cost of the pilot can be bounded by limiting the number of vehicles. Even if the vehicles are full and people have to wait, that’s ok because you’re still collecting data, and it’s a temporary problem that, once the microtransit money gets re-invested into fixed-route service, will go away.

        Another interesting question is what types of trips people are actually riding this for. Do people use it to go to work? Or, do they mostly walk to Link to go to work and save Via for trips to the airport, loaded down with heavy luggage?

      6. Yes, I hope they analyze the information thoroughly and ask the right questions and analyze the results. Where are most people coming from? What’s the nearest bus alternative? Are they riding it because a bus alternative is nonexistent or several blocks away or over a hill, or they can’t be bothered to walk one block to it, or they can’t walk five blocks?

        Access accepts only the most severe disabilities, so there are many people who have difficulty walking five blocks to a bus stop but don’t qualify for Access. This is a transit gap that doesn’t get much attention. But those needs are everywhere in the city and county, not just in Rainier Valley. And we need more regular bus service, so it’s not necessarily right to address sub-5-block needs if it takes money away from addressing general underservice. There’s also the overlap issue: if you have open-access demand-response service to address those who can’t walk 5 blocks, some people who can walk 5 blocks will use it. The question is which kinds of riders are the majority. The Via results will hopefully answer this, at least in Rainier Valley and for a short snapshot in time.

      7. Yes, that makes sense (and I noted that back when they were planning this). This should give folks a lot of data that they can then use to improve the bus service. Whether they do that, or continue to hire non-union labor in the name of techno-wizardry is another matter.

  15. I’m curious on how the Via for Tukwila performed relative to Via in the Rainier Valley, especially since it’s commuter oriented (peak only, Monday – Friday only). Did it get near the $10 cost per rider like Via as a whole?

  16. Via now serves almost four riders per hour at a cost per rider just over $10 largely because of the lower labor costs. (For some reason you left that second part out).

    Basically, we still find ourselves on this flowchart: Some of these services are terrible values, while the only ones that are remotely close to being a good value are the result of increased economic inequality.

    1. ‘(For some reason you left that second part out)’

      Specifically addressed down the page. But really, what’s the goal here? If you want to saddle small vehicle transit with big bus union labor overhead, you’re not going to have these services at all. Willing riders won’t be served and willing drivers will have to go find a lower paid or otherwise less appealing job. Not my equity agenda.

      1. Well, why not go all the way then? If you are focused entirely on cost savings, then why not run fixed route service using non-union labor? You would have way more riders per hour than even the best of these (generally under-performing) routes.

        As I wrote below, four buses an hour is crap for inside the city. Heck, it is crap for outside it. There is nothing special here; this is just union busting masquerading as innovation. It is a waste of money, and we should just put it into regular service, where even with union wages, more riders will benefit per dollar spent.

      2. Oh, and as far as “Willing riders won’t be served” with regular transit, how do we know that? Let’s say I live in Seward Park, it is noon, and I just missed the 50. In the past, I would wait for the next 50, or start walking up the hill to catch the 7. But wait — now I can call VIA, and they will get me to the station well before the next 50 arrives, let alone gets me to the station.

        Is there any evidence that these are actually additional riders? The ridership levels are so low, that it would be hardly noticed if they are poaching riders from the 50. But it is quite possible that these riders are doing exactly that. Instead of waiting or walking, they are simply using their phone, and getting Cadillac service. Is it really that shocking that the area with probably the best overall transit service — but not to Link — also has the highest numbers? Run this thing in the Central Area and I’m sure you will see very high numbers. Of course you will. If I’m headed to the UW from, say, Madrona, I don’t want to ride the 2 or 3 and then slog on the 48. I want a cab to the Capitol Hill Station, when I can quickly take the train to my destination.

        That is all good and well for the riders, but it is hardly a good value, regardless of what you pay the drivers.

      3. “There is nothing special here; this is just union busting masquerading as innovation. It is a waste of money, and we should just put it into regular service, where even with union wages, more riders will benefit per dollar spent.”

        That was my takeaway here as well and essentially why I was asking the OP about fully-loaded driver costs in my comment above.

        I’d rather the agency put the money into DART and/or other fixed route service where it’s appropriate, using smaller shuttle-type vehicles when feasible.

      4. Great post, Dan. What’s the source for the Via data? Do you have anything more granular? Four riders per hour seems much lower than what I’m seeing every day living near Rainier Beach Station. Folks line up on Henderson during the evening rush, and every vehicle that pulls up seems not only to fill up immediately, but to leave more people waiting on the sidewalk. The app just denied me a ride at 2:30 in the afternoon because all seats were full.

        Would be very helpful, as others have said, to see station-by-station numbers, and to know if there’s any late night ridership at all.

      5. “I would wait for the next 50, or start walking up the hill to catch the 7”

        … OR DRIVE. And demand a P&R at Othello Station. There have been people all along demanding P&Rs at Rainer Beach, Othello, and Beacon Hill Stations. it hasn’t happened because Seattle has a law against new P&Rs within the city. (And there are private parking lots near some of the stations. They just charge more than a P&R or street parking would.)

    2. By the way, it is tough to find bus routes in Seattle that have less than four riders per hour. There are a handful in the suburbs that are that low, but none in Seattle. Of the “urban” routes, the lowest is the 192. I’m not sure how this got put in the “urban” category, since it is an express from Kent to downtown. Yet at its worst (off peak) it gets 7.4 riders per hour.

      It is worth noting that DART service — which is largely fixed route, despite its name — performs better than this service. DART is focused entirely on the suburbs. There is no DART service to Sunset Hill, for example, even though it would likely perform extremely well. Yet just about every DART route — in the booniest part of the boonies — performs better than these services. Enumclaw to Auburn gets 19.1 riders per hour off peak, or about five times the ridership of the best of these microtransit services. The average is about ten riders per hour, and the absolute worst DART route — the 931 — gets 3.5 riders per hour *off peak* or roughly the same as the *best* of these services.

      The big takeaway is that we should run more small buses (and maybe more big ones) in places like Rainier Valley, and we should pay the drivers better.

      1. The lowest-ridership routes were deleted in the 2014 cuts; e.g., the 25, 42, and the 32nd Ave NW shuttle. The only comparable routes now are maybe the 78 and the 71’s tail. And both of those are a stopgap until Northgate Link opens in 1 1/2 years.

        I’ve ridden the 906 (Southcenter-Fairwood) a few times from Southcenter to Ikea. It’s a van route, so just one step up from DART, and it’s hourly. The first times I rode it it was hourly, and there was hardly anybody else on it if I remember. But the last time I rode it, wow. It was a Saturday afternoon and I counted 8 people who got on between Southcenter and Ikea. If it got only 2 more people in the other 2/3 of the route, then it would reach the threshold of 10 riders/hour. The entire route end-to-end is 38 minutes, so with a driver’s break it’s effectively one hour. And the van had more rows of seats than I expected, enough for 20 people. Almost as many rows as a 30-foot bus. I rode another van that had that many rows too; it must have been Trailhead Direct.

        There’s a more general principle here too. The low-ridership segments we lament: the eastern 62, the northern 71, etc — the entire route end-to-end is an hour or less. So Metro can still meet the 10 riders/hour threshold even with those tails. That partly shows what a good coverage route should be: a short low-volume segment combined with a higher-volume segment that goes to a strategic endpoint (e.g., the U-District, downtown, a Link station). Then you get overlapping trips in both segments and across the middle, and the higher-volume segment saves the lower-volume segment from extinction. There’s also a place for longer-distance coverage routes, like the 50 crosstown. But what you don’t want is a route like the 42 or former 131/132 (Seattle-Des Moines) that take an hour to get downtown and parallel a more frequent route (7) or should have an all-day express option (131/132) — they just give bad service and suck hours away from a more frequent grid. The 106 is in between. It serves overlapping trip pairs between Renton and Link, Renton and the valley, and across the valley. It’s redundant as a one-seat ride from Renton to downtown. But because it serves all those other trips, that flaw is minor. It’s certainly better than the 42 that didn’t go to Renton but just terminated in a low-density area south of Rainier Beach.

      2. RossB, the routes are classified by orientation; urban routes are oriented to downtown Seattle or the U District; all other routes are suburban. the practice began in 2011 when the subarea financial rules were killed.

      3. Metro groups the routes as eddiew says for planning and performance analysis. That’s really lopsided because it puts core routes like the 3 together with long-distance peak expresses, which are a completely different thing and serve a completely different mission goal. Long-distance peak expresses should be their own group.

        Peak expresses in general serve three different functions: to give a fast one-seat ride to outer suburbanites (Federal Way routes), to compensate for car congestion slowing down the all-day routes (15), and to manage capacity peaks because you have to run an additional bus anyway so why not make it an express (218, probably some Seattle routes). They should thus be put into two or three groups. Only the capacity and congestion issues possibly justify grouping them together with local routes. That doesn’t apply to expresses from Federal Way, Auburn, and East Kent. It only applies to the 218 in the sense that if it weren’t there, the 554 would need more runs, Sound Transit can’t accommodate that, and if they weren’t 218s they’d be 216s or 219s and what’s the difference.

  17. the routes are classified by orientation; urban routes are oriented to downtown Seattle or the U District; all other routes are suburban.

    Ah, OK, that makes sense. Thanks. I could never quite figure out how some routes were in one category or the other, but “goes to the UW and/or downtown Seattle” is a simple way to separate them, even if the labels (“urban” if it does, “suburban” if it doesn’t) are a bit confusing.

    I agree Mike, you could categorize them further, and my guess is that when they actually restructure, cut or expand, they do. But even in the broad context there are some interesting routes. The 37, for example, is definitely in that “urban” category (goes to downtown) and is actually completely within the city. But it performs poorly compared to other routes. Why is that? It is a bit surprising, really, given that the 56 and 57 perform so much better. The 43 and 47 also underperform, and they are definitely in that “urban” category, by any definition.

    There are lots of other examples of underperformance though, that are different beasts. Buses like the 111, 113, 114, etc., really shouldn’t be lumped together with buses like the 3/4. They just can’t compete.

    1. “But it performs poorly compared to other routes. Why is that?”

      The waterfront homes it goes by are probably too rich to take the bus. The 56 and 57 are inland. And if you’re going from the Junction area to downtown it’s much longer than the C. I don’t know why it even goes to the Junction, unless it’s to take people from there to the water taxi.

      I would take it to Alki on the weekend but it doesn’t run on the weekend.

Comments are closed.