Owl map, SDOT. Note that Routes 81 & 85 were changed to D & C Line trips in 2012.
Owl map, SDOT. Note that Routes 81 & 85 were changed to D & C Line trips in 2012.

Martin’s post this morning described the trade-offs around Mayor Murray’s decision to redirect some SDOT funds to prevent cuts to Seattle’s Night Owl service. In this post, I want to sidestep questions of whether the city should spend money on more very-late-night (post-1am) service, and where the city could find that money. Instead, I’d like to discuss the structure of the service which the Mayor proposes to buy.

Specifically, I’d like to call out the fact that in the context of 2014 Seattle, the service map above, which comprises a dismembered-flower-petal arrangement of one-way loops, designed to serve the city 60 (or more) years ago, is bonkers. If we’re going to spend city money on saving very-late-night service, we owe it to city taxpayers and transit riders to spend the money effectively, rather than perpetuating a horribly outdated route structure through sheer inertia and loss aversion.

Let’s state two basic transit planning precepts that apply here:

  • An effective transit service is a simple, comprehensible service. Ideally, transit should be like driving on a road: you show up and sit down, and it takes you in the same direction, the same way, every time. Riders shouldn’t have to memorize shifting, elaborate patterns of service for different times of day; variations, where unavoidable, should be minimized. While the petal-loops arrangement above is a great way to get a bus within half a mile of most of the people who lived in Seattle during the 1950s, its uniqueness, complexity and indirection detracts from the good work Metro has done over the last two decades, of focusing Seattle’s transit network down to a core of simple, direct, understandable routes.
  • An effective transit service connects the areas of highest likely demand. In the era of relatively cheap cab alternatives (UberX, Sidecar, etc.) and convenient by-the-minute car rentals (car2go), the people using very-late-night transit are likely low-income people going home from the city center, mostly after working swing shift, or maybe from a night out, so it makes sense to have the network radiate out from downtown; this much our current night owl service gets right. But those people aren’t then going home to Queen Anne, Broadmor or Madison Park, and they haven’t been for a generation. People working these jobs were some time ago pushed to less-tony areas of Seattle ($) — Northgate, Lake City, Delridge, White Center, and the southern Rainier Valley — if not out of the city altogether. Night owl routes need to reflect this not-really-new demographic reality.

Given the above, what would a sane night-owl network look like? I would start with the Seattle RapidRide network (Lines C, D, E), and then add the primary bus routes from each of the remaining points of the compass that remain unserved: Northgate, Lake City — 41; U-District, Northgate — 73*; Capitol Hill, U-District — 49; Central District — 3; Rainier Valley — 7; Beacon Hill — 36;  Delridge — 120; and, if I had any money left, Georgetown — 124. By comparison, under the current 15.6% Metro cuts scenario, Seattle would be left with very-late-night trips on RapidRide lines C, D, E, along with Routes 49 and 120; Routes 7, 36 and 124 would lose their very-late-night trips; and Routes 3, 41 and 73 have never had them (as far as I know).

Here’s the bottom line: The smart way for Seattle to spend any potential very-late-night transit money is not to preserve the existing mess of obsolete Night Owl loops, but to add (or preserve) very-late-night trips on the following services, roughly in order of priority: Routes 7, 3, 73, 41, 36 and 124. This restructure will benefit more riders, make better use of taxpayer money, and be more socially and geographically equitable than preserving the current Routes 82, 83, and 84. If we’re going to spend city money on service at this time of night, we should bring the route structure into the 21st century. As part of any “Save the Night Owls!” effort, the Mayor and City Council should seek to have Metro implement a restructure of night owl service along the lines I’ve described here.

* Note that I’m referring to the post-June 2015 restructured Route 73. Before the June restructure, it would probably be best just to run a truncated 73 to Ravenna Boulevard. For the purposes of this post, the alignments of the other core routes on this page don’t change significantly through the cuts process.

58 Replies to “Metro’s Night Owl Alignments are Insane”

  1. Just my experience from the defunct 85 but the people on the night owls weren’t coming from somewhere but rather going somewhere, especially on the second trip. I used it to be able to get to work before 6am as did others. It also served as a feeder for early morning flights (when the 560 went to the WS Junction in the mornings). I think the later Owl trips are more important than the earlier ones.

    (No, the owl C trips don’t help as I was the one person getting on along 35th)

    1. I had a similar experience when I used to occasionally take the old 81 at 8th and 85th around 4:00 AM to meet the first SB Link train of the morning at Stadium station to catch an early flight (This was easily the most time-efficient version of a Greenwood–SEA-TAC transit journey I have ever had the pleasure of experiencing). The bulk of riders appeared to be downtown office workers going to work early. Typical passenger load was 2-4 rolling shelter riders, 2-4 airporters, and 7-10 downtown workers. I don’t know how many people it let off on the way north.

      1. My only experiences with Night Owl buses in the last 20 years have been in connection with early morning trips to the airport and, yes, I was not alone on the bus.

  2. Bruce, I agree absolutely 100% with the thesis of your post but would suggest slightly different routes to prioritize in Northeast Seattle.

    I agree on the post-restructure 73 but think if we can’t have that I’d rather have the 43 and the 16. The 43 covers the U-District (and, believe it or not, does so pretty quickly at 3 in the morning) and part of east Capitol Hill now served by the 84, and the 16 does the same for Northgate while adding both Wallingford and Roosevelt. Little is gained by running a freeway express (41) to Northgate at this time of day. If the 43 and 16 together are too many hours, then let’s do the 66 rather than a truncated pre-restructure 73.

    I agree on the rest of your choices. Late-night service on the 3S, in particular, is the most no-brainer thing imaginable.

      1. I didn’t want to protract the body of this post with hypotheticals or excessive turd-polishing, but yes, if there was a case where it made sense to have a modified service pattern for night service, it would be northeast Seattle. I could get behind an “Route 86” NE Seattle Night Owl that would combine pre-restructure 66-41 or post-restructure 70-73-41.

        We’re all on the same page in any case. Less crazy loops, more straight lines; add trips to existing core corridors rather than dream up new ones.

      2. Another modified night service I could get behind: Have Seattle, King County, and Renton partner to extend the 7 trips to a 7-106 Rainier Valley-Renton Night Owl. This would be way more useful than the current 280.

    1. When I take night owls I run into a fairly large number of people who are trying to get to either Lake City or Northgate.

      I like the idea of a post-restructure 73 or a 66 a night owl route.

      Alternately late trips on the 72 could replace the 83 and provide service to Lake City.

      Another thought for Lake City service would be to get Sound Transit to run late trips on the 522. Bothell, Kenmore, and Woodinville have a fair number of people without access to a vehicle. There are also industrial jobs in Bothell and Woodinville with a very early start.

      Over the years here is what I’ve observed on the remaining night owl routes:

      82 – Most riders got on/off in Queen Anne or Fremont. The few remaining riders seemed to be going to Aurora. RR E replaces most of the 82 north of the ship canal.

      83 – Most riders get off in the U District. Some ridership to Roosevelt and Maple Leaf.

      84 – Almost all riders get off by 12th. Very few riders. I’ve been one of only 2 people on the bus leaving downtown other than the driver

      124 – a surprising number of riders, provides a vital link to the RR A and airport after Link shuts down for the night.

      280 – Really no riders other than in Renton.

      1. I spent a fun shakeup on the back end of the night Extra Board at Central Base, where I had the privilege of driving a bunch of Owls.

        The 83 is really just a 73 with a funny name. I found there was no point in driving it beyond the U-District. I probably could have turned into the 70 layover, parked for 20 minutes, and resumed my trip and no one would have known the difference.

        The 7 and 124 (then 174) are crowded!

        The 280 is a homeless shelter. Transients like the freeway ride. I drove it maybe 10 times, can remember two legitimate passengers, both headed for Bellevue. (It was also impossible to keep on schedule at that point unless you drove in a thoroughly un-Service-Quality-approved manner.)

        The 82 and 84 are ghost towns. The few passengers on the 82 are headed for near Aurora, but they all probably take RR E now. The 84 might have a person or two headed up Pine, and then a drunk straggler getting on somewhere on Jefferson or Cherry.

        I never drew the 81 or 85 for some reason.

      2. Why is central Fremont off the map with the Night Owls? It’s kind of a long walk from Aurora to Fremont.

      3. If the 124 is so popular, why does Metro give it so many fewer runs than the A or the 131/132? I thought it was mostly empty.

      4. The 124 is a good, but not great, route in the grand scheme of things. But it and its predecessor 174 have always had high night ridership (and recent Metro data which I happened to get for another purpose confirm that is still the case). It “serves” (by which I mean “gets sort of near, but nearer than any other bus”) several of the poorest transit-accessible neighborhoods in and near Seattle, so I think it probably has a lot of shift-worker ridership.

      5. Another glaring whole the the late-night network is that the 124 stops at Tukwila International Blvd. Station, forcing a transfer to the less-than-once-an-hour A-line to continue onto the airport during the hours that Link isn’t running.

        During the hours that Link isn’t running, the 124 should be extended to the airport to maintain the one-seat ride that used to exist before the A-line restructure. Lots of airport workers need to arrive before 5 AM to begin their shifts. By contrast, late at night, TIBS is a virtual dead zone – the number of people taking the bus at 4 AM to pick up their car in a deserted parking lot is negligible.

      6. @asdf: I suspect that a lot of people who would be benefited by that are not so much air passengers as SeaTac workers, and they deserve a convenient trip to work at that hour.

  3. I agree with this post’s content, but I would say that “inertia” and “loss aversion” do not amount to “bonkers” or “insane”, and the title is unwarranted hyperbole that distracts from the main point.

      1. Well, I suppose yes, you are right. I just think it takes a bit away from Bruce’s point, which is 100% spot on.

    1. The 82, 83, 84 routing is bonkers. One reason I don’t live in Lake City or Northgate is it’s a 1-3 mile walk from the night owls. I had enough of that growing up in Bellevue where I lived an hour from the 280 stop.

      And my roommate would take a night owl to his Kent job if there was one, but the 280 and A also stop an hour’s walk away, so he either takes the last 150 or first 150 and sometimes waits 45 minutes for his shift to start, or can’t take shifts because there’s no transit at that time.

  4. I’m hardly the night owl type, but once in the pre-Lyft/Uber days I took the 84 from downtown to Madrona. Half a dozen people went from downtown to Capitol Hill, and 3 of us stayed on through at least Madrona. Nobody boarded or deboarded in Madison Park, and the driver said “I hope you all enjoyed your pre-dawn tour!” Since the 49 already has Night Owl service and the 11 doesn’t need it, adding two trips to the 2 or (more likely) 3 is much more elegant solution than saving the 84.

    I do hope that post U-Link we can push for 22 or 24-hour Link service, or at least an earlier start to the service day in order to catch those pesky 6am flights.

    1. At some point, Link will need 24 hour service. When it takes 2 hours to go end to end, there’s never a time trains won’t be on the tracks.

      1. Except that it has to be shut down for maintenance at some time and there’s no separate local/express tracks like Chicago and NYC (the only 24-hour rail systems in the USA) have.

      2. But Chicago runs the Blue Line 24 hours, and I don’t remember its having express tracks. That seems to mean we don’t need maintenance every night.

      3. Chicago Blue Line maintenance is… a pain in the neck. It keeps accumulating slow zones because they aren’t really keeping up with the maintenance.

        It’s possible to do maintenance on 24-hour lines, it just takes a lot more worker-hours to do the same amount of maintenance (since they have to stop for 10 minutes every time a train goes by) so it costs twice as much.

      4. London has only recently announced night tube service. (And even that is only on a limited subset of lines and only on Friday and Saturday nights.)

        Running the Link Shuttle on a scheduled basis at night would be far more economical. The current routing is a bit tortured to hit all the stations but could be modified to avoid servicing any stops not likely to attract night ridership, thus speeding up the trip. That late at night, there likely wouldn’t be much traffic to get in the bus’s way.

      5. @VeloBusDriver — That makes a lot more sense. A special “late night” bus route that covers the corridor of Link is very easy to understand. That is in big contrast to the current schedule, which is very confusing.

      6. The only awakward stations currently are SODO and Beacon Hill. A bus serving SODO has to either backtrack to Holgate Street or take Columbian Way and backtrack on 15th Ave S to reach Beacon Hill station. On the other hand, it takes only two or three minutes to go down to SODO and back at night when there’s no traffic, so it’s an insignificant cost and travel time compared to the overall night owl network.

        University Link will present the biggest challenge, however. The only way to serve both Capitol Hill and UW stations is like the 43 does.

        Also, there’s the closure of the DSTT at night. It would have to remain open, or the bus would have to use surface streets downtown. Which means people would have to know where the night bus stops are.

      7. I don’t think it has to match the daytime route exactly, just something similar. In some cases (as you suggest) it might not match it at all. But a route that would allow someone at a station to walk a few blocks in the middle of the night and still pick up a bus would be a good thing. For example, a rider could arrive at the Capitol Hill station, read the sign, and realize the best thing to do is catch a 49 (either downtown or to the U-District).

      1. PATH just uses buses on the very early morning runs and weekends when heavy maintenance is required, but otherwise is 24 / 7.

      2. How about a compromise solution of running Link 24 hours on just Friday and Saturday nights?

  5. This all sounds about right, and most of it probably is. But how might it come to pass? If the council approves Night Owl rescue, it will inevitably do so “as is” given the September deadline and a (justifiable) wish not disrupt lives with a hiatus, even if the routes might be changed later.
    The best outcome is for the city to realize – for its future bus service funding – it should plan systematically and buy service, not specific routes. And then contract with Metro (and possibly others) to deliver the service. And of course this should include re-configuration of Night Owl service.

  6. I like this though I might quibble a bit with the particular routes and their priority.

    I’ll point out the 124 provides a vital link to the late runs on the RR A and to the Airport. There are a fair number of people using both.

    Here are some alternative pairings for a restructured 73 and 41:

    Owl runs on 16 and 72
    Owl runs on 66 and 522

    1. The 124 runs a long way beyond Seattle limits. I feel like service on it should be county-funded, or jointly funded by Seattle and Tukwila.

      1. Ah, I was thinking about the relative utility of various night owl routes and not who should be funding them,

        Even though there is demand on the 124 it should at most be 50% funded by the city.

  7. Another piece of illegibility quirk: there is one pair of stops only served by Route 82, on Queen Anne Drive/2nd Avenue N. And even more, the southbound stop is only served on the 2:51am trip, but is bypassed in favor of Aurora on the 4:01am trip. So you have one route trying all at once to be the 3, 4, 5, and 16, but with a separate routing number just for night owl service AND an express service variant.

  8. Here in Minneapolis, our 2/3 (depending on how it’s looked at) owl routes (5/19, Green Line) have timed connections downtown from 12:10 am to 5:00 am. It seems that a good model to use would be an hourly service that meets up, similar to the current situation.

    Suggested routes:
    – The new 3
    – Combined 5/E, up Dexter, Phinney, Greenwood, 80th, and Aurora to the County Line.
    – 7 with a possible extension to Renton TC.
    – 43/44
    – 70/73/41 (perhaps with a Green Lake detour)
    – Variant of the D Line providing better coverage in Ballard.
    – 124/A interline
    – 120 (to the Airport)
    – C/21 loop route
    – 180 potentially
    – I-5 South all stops (DT-SODO-Airport-FW-Tacoma-Lakewood)

    What should also be focused on is standardizing the operating hours of the daytime routes. A good rule of thumb would be daytime routes having the first arrival/departure from Downtown at 5 am, and the last trip arriving/departing around 1 am. Trips at 2, 3, and 4 am would be on the night bus network.

  9. Thanks for calling this to our attention. The route structure is so outdated that I wouldn’t know where to begin. Madison Park gets service, but not Beacon Hill or 23rd/Jackson? Everyone has to go Downtown to transfer? There is no crosstown bus from Ballard to U-District to Cap Hill to Mt Baker? Does Metro just want us to contact Uber?

  10. I would add 44 and Link (or equivalent shadow) to the list. Whatever we do should use existing day-time routing, to make it as accessible as possible to people, and to allow folks to plan their lives around a single, integrated transit system.

    1. I agree with euthanizing the remaining 80-series routes, but I would make night-owl shadow buses for Link a top priority. The new 73 and 49 are obvious winners. But also have a route connecting from the airport to downtown via MLK.

      That said, I hope the City asks Metro to offer a proposal for the best use of night owl service hours (which would most assuredly NOT mean keeping any of the 80-series routes). Let Metro do the planning, and make the best use of scarce night-owl resources.

      1. I thought about the idea of a Link shadow, but considering how scare night-owl service-hours are, it does seem a bit redundant with high-priority routes. The 36 would already cover Beacon Hill, the 7, the Ranier Valley. The mandatory transfer to the A-line at TIBS to get to the airport does need go, but it would be much cheaper to accomplish that with a small extension of the 124 than to introduce a whole new route going from Ranier Valley to the airport.

        A shadow bus would also be significantly slower than Link, due to having to wait at stoplights and make awkward turns to reach all the stops. A shadow route serving all the Link stations would probably take longer to get to the airport from downtown than the 124, plus a small extension on the south end to eliminate the transfer.

      2. While I agree on the importance of night owl service on the 7, it is no substitute for shadow service along MLK in the middle of the night.

        Nor would such a shadow bus be much slower than Link, especially if it has signal priority. The 8 does pretty well at keeping up with Link along that stretch during the day.

        From Rainier Beach Station to TIBS is two freeways. But, if a combined A/124 could take care of TIBS, the shadow bus could go straight to the airport.

        The shadow bus shouldn’t have to compete with other all-day bus routes for funding. Metro (perhaps now with the City of Seattle’s help) pays for the all-day bus routes. Sound Transit can afford to pay for overnight night owl service. This will become even more important when U-Link opens.

  11. I’m not really qualified to dispute anything in the article, but I think that the premise that no-one in typically wealthy neighborhoods wants to ride late-service transit is at least anecdotally untrue. If ridership numbers don’t support my notion, I think it’s because these people can afford to choose, and transit in it’s current state doesn’t get them where they need to go in a reasonable amount of time (day or night).

    1. Agreed. I’m a well-off tech worker who used to ride the last outbound 18 or the first run of the 81 to get home from weekend excursions to Capitol Hill. When Metro decimated Ballard’s evening service, I switched to Uber.

      I would take transit if it were convenient. I do happen to think that the only way to make it reasonably competitive with low-cost services like UberX is now going to be grade-separated rail.

    1. Also, how does the cost of a service hour at 3 in the morning compare to a service hour at 3 in the afternoon? On the on hand, traffic is light and buses can actually move reliably through routes that, during the daytime, would be hopelessly congested. On the hand, I would not be at all surprised if union rules demand that bus drivers get paid significantly more for working the hours that most people would rather spend home in bed.

  12. >> the people using very-late-night transit are likely low-income people going home …

    My thoughts exactly. You also have the people who have to start work early. I discussed this with my wife, who is in health care. I used to be a security guard, and we both came to the same conclusion: A lot of people who work late at night have shift changes around midnight or 11 P. M. (and 8 hours later). This means that “night owl” service of this type isn’t needed for a lot of jobs. It is more important to just continue the regular route until around 1:00 A. M. (which I assume to be the case).

    I think the biggest group of workers who need night owl service are the folks who work in the restaurants and bars. These locations are spread out around the city, but there are certain places that have a lot more activity. From what I can tell, the routes do a pretty good job of at least going to those locations (Capitol Hill, Pioneer Square, Belltown, Ballard and the U-District). One obvious exception is Fremont. The Aurora bus comes close, but the speed savings for avoiding Fremont are minimal at that hour.

    For late night service, a system that requires a downtown transfer isn’t the end of the world, but it better have really good transfers. If you have to take a bus from Ballard to downtown, then transfer to a bus heading to the U-District, that other bus better be waiting for you, and it better not leave early. Security is also an issue. I think it is the only time of day where a downtown Seattle transit center (with security guards) would make sense. Since that probably won’t happen, it makes sense to add some cross town buses. Just as traveling through downtown isn’t bad late at night, neither is traveling on 45th, or Denny, or 23rd. Since traffic is pretty much non-existent at this hour, it could allow you to synchronize the transfers. Now someone from Ballard could take the 44 to the U-District, then transfer to a northbound bus.

    The other group of workers who want night owl service are the folks who need to arrive early in the morning for work. I think this includes people who work at bakeries, donut shops, and sometimes coffee stands. These places are more spread out, and thus more difficult to categorize. I think the same general approach should apply, though. Easy to understand routes, good transfers and a few more cross town lines.

  13. Great post, and the changes are very overdue.

    As usual, you get what you measure. In this case, I think it’s important to be clear about exactly what goals we have for the owl network. Here are my ordered priorities:

    1. Coverage: The owl network should maximize the number of destinations (homes and late-night/early-morning businesses) within a 1-mile walk of an owl stop.

    2. Consistency: Each owl bus should have a single routing for the entire night, and ideally the same routing in both directions (modulo one-way streets).

    3. Familiarity: Owl routes should resemble daytime bus routes to the greatest extent possible.

    4. Direct: Owl routes should follow logical, direct routings.

    These goals are somewhat in conflict. For example, Route 41 scores well on priorities 2, 3, and 4, but does not help much for priority 1, due to its long non-stop segment. Route 82, meanwhile, scores poorly on all four points (#1 due to the lack of coverage north of 85th).

    The biggest point I want to make with this priority list is that coverage is more important than familiarity. For example, the more I think about it, the more I think that RapidRide E is not a very good choice for an owl route. North of Winona — basically, where it becomes a local bus — it’s great, and probably the most important route in central north Seattle. But between Denny and Winona, it’s basically an express bus. It makes only a handful of stops, and it skips many important nearby destinations (e.g. Dexter, Fremont). An alternative route might take Dexter/Fremont between downtown and 50th; Phinney/Greenwood between 50th and 85th; and Aurora between 85th and the E terminus. This is a somewhat slower route for people heading to Aurora Village, but it’s enormously better for people heading to Dexter, Fremont, or Phinney Ridge.

    As another example, I think that the most important CD bus is the 3S, but the most important Queen Anne bus is the 13. These two buses would make a logical pairing at night. (Actually, I think they’d make a logical pairing during the day, but that’s neither here nor there.) And if we keep running the E unmodified, it could also make sense to extend this bus a little further north, crossing the bridge into Fremont (or even just stopping at its southern terminus).

    I’m not saying that what I proposed is definitely the way to go. I just think that we should be aware of what our priorities are. The two biggest problems with the legacy night network are that it stops at 85th, and that the routes run in different directions over the course of the night. Having night routes that are different from day routes is not ideal, but if coverage is our #1 goal, the alternative might be even worse.

    1. Aleks, I think you’re assuming too much willingness on the part of the general public to learn details. I think, realistically, the owl loop experience shows that people will ride late-night buses if and only if they are trips on “their” bus (or a nearby bus they know.) If the last trip on the 73 is listed as leaving at 1:15 a.m., they’re not going to go searching for an 83 schedule — they’re going to assume the last bus is 1:15 a.m. And showing trips on a substantially different route on the same schedule causes all sorts of problems.

      So I prefer more legibility even at the cost of some coverage.

      You might be able to get away with a slash number combining two routes, and listing the trip on both schedules, if *all* or almost all of the stops of both routes were served. A 66/41 combination, signed “41/66 Lake City Northgate Roosevelt” could work like this (and would actually be an excellent owl route). But if you start cutting off substantial portions of routes, you’ll get people lost. I don’t think your 26/5/E Line, as well as it would provide coverage, would be any more navigable than the 82 for most riders.

      As a side note, the 3S and 13 are already through-routed in the evenings after the 2S drops back to half-hourly.

      1. I think the biggest problem with the night owl run is that it is neither here nor there. It doesn’t follow existing bus routes, nor does it follow a coverage pattern that is easy to remember. For example, without looking, how far up 15th NW does the bus go (and which direction is it going)? How about 15th NE? To be fair, switching over to the D-Line helps answer the first question (Holman road) but the second question is still as confusing as ever.

        So, if you are going to have night owl only routes, they should be simple. Ideally, they should reuse existing routes. So, for example, let me offer up the 40, the 5, the 44 and the 73 for north of the ship canal. Don’t like it? Then how about the D-Line, plus night only buses that follow Phinney/Greenwood and 15th NE. That would be really obvious, and would provide just about as much coverage. In some cases, like when you are on the east side of Greenlake, you would have to walk a long ways. But my guess is there aren’t that many people in that area.

Comments are closed.