A month and a half ago, when I wrote about a possible restructure of trolleybus service on the Queen Anne-Belltown-Downtown-First Hill-Madrona corridor, I promised “within a week” to explain why the restructure could deliver so much more service with roughly the same amount of money. Obviously, I’m several weeks late in doing this, but I hope you’ll forgive me.

To tackle this subject, I have to introduce a some planner jargon:

• Clock-face schedule. A schedule that attempts to place a bus at a stop at consistent times past the hour (e.g. :07, :27, :47 for a 20-minute headway service). For routes operating at headways longer than 10 minutes, these schedules are the most comprehensible to riders; at shorter headways, riders tend not to worry about schedules.
• Cycle time. This is the amount of time a bus takes to run a complete round trip and be ready to start out on the next trip. This includes driving time, required layover break time, and schedule padding, but does not include the deadhead time from the base to the starting point.

Let’s take a simple example. Suppose we have a route that takes 27 minutes to drive each way. Metro’s union rules require a five-minute break at the end of each cycle, and let’s suppose five minutes of padding per cycle are required to make the bus keep time reliably. The cycle time of this route is 27 + 27 + 5 + 5 = 64 minutes. Knowing this, we can work out that a pattern that maintains a clock-face schedule with 30-minute frequency would require three buses in service at once; 15-minute frequency would require five buses; and 10-minute frequency would require seven buses.

This simple example leads us to an important insight: routes that operate more frequently can almost always be scheduled more efficiently than less-frequent routes. In our example, going from 30-minute to 10-minute headways is a 300% increase in service that requires only a 233% increase in service hours; the difference is made up by shorter layover times. Of course, in some cases, route timing will line up just right so as to make the efficiency savings of increased frequency quite small; I chose the number 27 as a somewhat extreme example to make my point. Nonetheless, awkward route lengths occur quite frequently in the real world.

Knowing this, and looking at a map of current Queen Anne-Madrona service, we can see the problem:

Metro’s daytime service on this part of the network consists of one 15-minute route, plus four 30-minute headway routes that are interlined together to provide 15-minute headways in their most popular sections before wandering off into the neighborhoods. During the week, additional trips run from Virginia to 21st Ave, effectively creating a sixth route. Having all these different termini wastes a tremendous amount of driver and bus time; and this is where the restructured service is brilliant:

By splitting the 2S into a separate 15-minute headway route, and bringing all the other routes to a common terminus at the north, those routes can be scheduled as if they were a single route with sub-10-minute headways rather than four or five different 30-minute routes, leading to dramatic cost savings analogous to our simple example above.

These cost savings are enough to allow for the extension of frequent service out to Madrona* and into Sunday and evenings in East Queen Anne. This analysis also addresses the objection that this potential restructure over-serves the north side of Queen Anne. Certainly, nothing there justifies sub-10-minute service, but it is just a pleasant side-effect of the cost-saving one-terminus routing. Extending frequent service to Seattle Pacific, along with extending the 2X on Nickerson also allows some possibilities for improving Route 17; I will discuss this in a future post.

One objection this analysis does not address is that it might be better to maintain all-day service on Route 2N, extending the trolley route to the terminus of Route 13, splitting headways between 3rd 6th Ave W and Queen Anne Ave. To justify that change, we’ll have to examine the question of whether riders prefer higher frequencies versus wider geographical coverage, and I’ll examine that question in depth in future posts.

* There’s actually even more to it than that: additional service out to Madrona is almost free (versus an alternating turnback at 21st Ave providing 30-minute headways) because of how the cycle-time pencils out in either case. Reducing the south of the network to two terminals  also saves money, just as reducing the north from four to one does, although the savings are not as dramatic.

## 50 Replies to “Why Current Queen Anne-Madrona Service is Inefficient”

1. Chris Stefan says:

One thing I’ve seen in other proposals is to loop the current route of the 1 and the 2 together on Queen Anne. That would solve the problem of providing local service on 6th W and W Galer.

1. Bruce Nourish says:

For the 1 I would:

* Break the through-route with the 36, through-route it with the 14S.
* Reduce midday frequency to 30 minutes (from 20).
* Increase peak frequency to 15 minutes (from 20).
* Change back to the late-night shuttle routing that existed prior to Bridging the Gap.

Extending the 1 somewhere more useful would make it a slightly better route, but I don’t think the differential ridership would be enough to justify it.

1. Aleks says:

I’m veering off topic quickly, but are you still in favor of a 36/70 through-route? (I ask because the TMP seems to be taking a contradictory position with Corridor 3.)

2. Also, does this mean you agree with axing the 14N?

2. Gordon Werner says:

what about the Rt. 1 and the Rt. 36 which share these buses?

3. Gordon Werner says:

[ot]

1. Gordon Werner says:

[ot]

2. Mike Orr says:

[ot]

1. Aleks says:

[ot]

1. Aleks says:

[ot]

4. Mary says:

I think to make this work for customers in west Queen Anne during the day, there would need to be some sort of van service making a loop over to the main bus routes. Many of the passengers outside of rush hour commute times are elderly or have small children. The walks from 7th or 8th down to 10th W for the 1 can be steep, and south of McGraw the walks between Queen Anne Ave & 4th – 8th W are up & down. Perhaps easy enough for a fit adult, but not so simple for others. It doesn’t matter how great the service is on QA Ave (or 15th W, for that matter) if it’s a major hassle to get there :)

5. I’m still not sure there’s all that much on the 3/4N corridor. I’d buy the argument if Taylor were as dense and busy as Queen Anne Ave, but the main benefit is probably serving both sides of Seattle Center.

On the other hand, this is a powerful argument against turnback routing, and maybe it can convince some people in Madison Park to stop opposing putting trolley wire back in… OK, I’ll get serious now.

1. While there is little or no retail along Taylor/5th Ave N, there is plenty of multi family housing.

3/4 runs on 5th Ave N below Valley, where there’s a QFC and various shops. Actually there are some restaurants and coffee shops on Taylor below Ward St too. But as Oran said it’s really all the apartments, and if you happen to live near Galer St in SLU you can cross Aurora on a footbridge there.

3. Still not convinced the effect in shared termini is worth it. I don’t see any way Metro gets any efficiency gains out of it unless some buses that arrive as 3/4s leave as 13s and vice versa, and I don’t know if they do that anywhere in the system.

Part of what throws me off about Taylor isn’t so much the lack of retail as the lack of intersections. Also, Galer’s a bit north for me to consider it “SLU”.

1. Bruce Nourish says:

“I don’t see any way Metro gets any efficiency gains out of it unless some buses that arrive as 3/4s leave as 13s and vice versa”

That is precisely what would happen under this restructure. Equipment for these routes is completely interchangeable.

6. Aleks says:

Thanks for the post. This is a really good explanation.

One further question I have is, what effect do you get from having divergent route lengths? That is, suppose that the 3’s takes 7 minutes longer than the 13 to get from 3rd and Cedar to QA Ave and McGraw. It’s clear that you’ll still get most of the benefit you describe here, since buses can arrive at SPU as 3s but depart as 13s (or vice versa). But conversely, frequencies from McGraw to SPU will have to be erratic, since it’s not possible to synchronize them for both of their shared segments.

Are there any other downsides of having diverging routes like these, from a scheduling standpoint?

1. Bruce Nourish says:

I don’t know exactly how it would play out if the route run-times were divergent; I suspect we’d still save some money. Between 3rd & Pike and Queen Anne & Boston, where the 3 and 13 would be partly separate, the difference in run time is only a minute or two. So it isn’t a big deal in this case.

1. d.p. says:

The 3 is quite a bit faster than the 13. Spends less time in Belltown, avoids the Broad Street jag, zips up the medium grade while the 13 crawls up the cliff, and doesn’t have nearly as much traffic to deal with at the all-way stops as the 13 does.

Under current conditions, if your destination is the north slope of the hill and you miss a 13 just before a 3, there’s no question that you can take the 3, beat the 13 to Boston, and switch.

Still an excellently-reasoned idea, Bruce, but this minor kink would have to be factored in (likely just leading to a staggered service pattern on the SPU segment).

7. Stephen says:

I think it would be even better to eliminate the 3N/4N entirely, and improve the pedestrian connections to Aurora (build a couple more pedestrian bridges), and to add a few stops for the 5/ RapidRide E. It’s only a 1/10th of a mile from Taylor to Aurora, and then all of people who currently take the 3N and 4N would have easy access to extremely frequent services (I believe that both the 5 and RRE will have 15 minute headways, which means you have 7.5 minute headways all day, even more frequent at peak). The 7.5 minutes you save more than outweighs the 3 minutes it would take to walk the 2 blocks, plus that then saves you from having too frequent service to SPU.

1. Matt L (aka Angry Transit Nerd) says:

Eliminating cost-efficient service in favor of bus stops as inviting as this one?

Good luck with that.

1. Stephen says:

Yes, that’s why I think we need to spend some capital money to improve the pedestrian connections, and there is already RapidRide money allocated to make the stop into a more usable “station” along that corridor. The point is that some capital money spent now can allow metro to save service hours in the long run.

Also, just because service is currently cost-effective doesn’t mean that it’s the optimal configuration for the network as a whole. Just because you can have 2 frequent parallel routes 2 blocks from each other, and still both have good ridership, doesn’t mean that you can’t be even MORE cost-effective by combining those two routes.

2. Bruce Nourish says:

Yeah, I don’t think anyone familiar with the topography and built environment of East Queen Anne/Aurora would suggest this.

It also misses the point that the purpose of urban bus routes is to connect neighborhoods to their local retail districts and community center, not just get people downtown.

3. Mike Orr says:

The top of Queen Anne is flat from 7th W to 3rd N, and from Galer to McGraw, except the very southeast corner. 7th W to 4th W is also flat to Raye and the cemetary. Beyond that it gets steep quickly. That’s why Taylor, Aurora, Dexter, and Westlake have parallel bus routes, because even one block is steep. Each street is effectively a separate neighborhood, as in West Seattle. On the other hand, there may be people walking from 8th W to the 2 to avoid walking to the 1, but I don’t think that’s enough justification for keeping the route.

4. And then there’s the proposal to move the 5 to Dexter to replace the 26/28.

2. Aleks says:

What Matt said.

Honestly, I would sooner scrap all bus service on Aurora south of the ship canal, and reroute the 358/RRE along Taylor/5th (plus a new road/bridge linking 5th with Aurora).

South of the ship canal, Aurora just isn’t a good street for buses in any way. The only reason we run buses there is that it’s the best way to get to the northern parts of Aurora. It’s definitely not substitute service for anything to its east or west.

1. Aleks says:

Note that I’m not saying we *should* make that change to the 3/4/358, just that it would make more sense than the change you’re proposing. In practice, I don’t know that those changes are worth it, at least not right now.

2. Alex says:

Um, you can’t put an Aurora route onto Taylor/5th. There are no bus-usable connections between Queen Anne Drive (south end of the Aurora Bridge) and Valley (where the 16 exits Aurora currently). I don’t know if a 60′ coach could exit at Queen Anne Drive because the exit is basically a 90 degree turn. There are a couple more tight corners before you get to Queen Anne Avenue. It would be terribly time consuming to riders coming from farther north on Aurora.

3. d.p. says:

The only reason we run buses there is that it’s the best way to get to the northern parts of Aurora.

The 358/future RapidRide E serves an important purpose in that it connects a busy-yet-distant corridor to downtown. Once it breaks free to its limited-action section, it does so quickly but with exceptionally poor reliability.

It’s poor reliability and quirks of geography combine to make it pretty useless to anyone below about 73rd Street.

Sure, every once in a while I’ve gotten lucky, and OneBusAway has shown a 358 bus 11 minutes away when I was about an 8 minute walk from the stop on the 46th St overpass, but the areas it passes through — 45th/46th corridor included — simply aren’t developed in a way that makes it a logical departure or transfer point. Add in the unpleasantness of waiting by the highway, and the route’s 0% timeliness once it reaches the lower double digits (something the kneecapped RapidRide can only be expected to marginally improve — and the appeal of its “express-ness” is negated.

(Even in the northbound direction, most of my attempts to break free of downtown’s orbit quickly via a 358, followed by a walk or a connecting bus, have backfired spectacularly.)

When I submitted for consideration that east-west subway spur map, I decided it was more important to create an excellent connection to the 5 corridor than to provide a transfer to the Aurora RapidRide. I provided the line a much greater reach, giving it easy access to the entire populated walksheds of Greenwood, Phinney, and Fremont.* The northern reaches of Aurora, meanwhile, would do better to be given better access to the Link system at Northgate. There is simply too little you can do to make the Aurora corridor more than incidental to peoples transit movements in the middle third of this city.

*(While I agree with Bruce’s reservations about routing the 5 over the Fremont Bridge under present conditions, I think the subway would effectively transform the 5 from a long-distance workhorse into a feeder/local, making a route that gives lower Fremont and Dexter immediate access to the subway worth the trade-off of dealing with the drawbridge.)

4. d.p. says:

Wow, “limited-action?” I almost don’t want to know what other typos I made if I started out with that doozy. That’s “limited-access,” obviously…

3. Matt the Engineer says:

Others have said this, but I wanted it to be clear. Queen Anne is one of Seattle’s tallest hills, and the route from the top down to Aurora involves many, many stair steps. I’m in reasonable shape, and can build up a good sweat on this route on a cold day.

Unless “pedestrian improvements” includes a funicular, this will never be a popular pedestrian route.

1. Mike Orr says:

And at night it’s pitch black at some of the stairways, and a pile of wet leaves can make it even more dangerous.

2. Eric says:

Darkness is a solvable problem if we’re willing to invest money on more streetlights.

3. Bernie says:

“Clearly” appropriate lighting is important. But “darkness” is not a problem. Light pollution is!

A number of cities in the U.S. have developed standards for outdoor lighting to protect the rights of their citizens against light trespass.

Can’t find the link just now but I think it was New Zealand that had a town leading the fight. Not only is it obnoxious to have a street light illuminating our living room at 3am it’s also a huge waste of energy. Street lights are programmed to turn on when it’s still completely light out… gee, I wonder who’s benefiting from that bill???

4. Bruce Nourish says:

I assure you, darkness is indeed a problem on the steep, narrow, and often tree-overgrown stairways of Queen Anne.

That said, light pollution does indeed suck (I paid to get out of an apartment lease once, for that reason) and it should be added carefully so as not to be a nuisance to stairway neighbors.

8. GuyOnBeaconHill says:

Bruce, I’m thinking of ordering some “SAVE ROUTE 2–CALL 555-5555”, when do I need to have them ready?

I don’t see any compelling reason to delete the West Queen Anne leg of the 2N. The WQA leg is about 1 mile long and takes about 4-5 minutes to run. Overall the route has outstanding ridership and, according to the stop data documents you presented before, each bus on that leg serves about 9 passengers per trip during the middays. That means 36 current passengers per hour will have to hoof over to the 1 or 13. Not likely to happen.

Sending 8 to 12 buses per hour to SPU is ridiculous. Ridership from SPU is barely higher than what Metro gets on the WQA leg, so based on that fact alone, why does Metro want to quadruple service to SPU? To get more efficient scheduling? If the 3N runs to SPU 4x per hour and the 13 runs 2x/hr, Metro will have plenty of chances to wring out the excess layover time.

Let’s look at reality (and hope my coding worked). The commercial district on Queen Anne Avenue betweend Boston and Galer is an auto-centric hub that needs to undergo a drastic transformation before it can be considered a transit corridor. If the top of Queen Anne were to embrace Broadway style redevelopment, there would be a need for more buses. But in the current neighborhood on top of QA we find 2 large supermarkets with suburban style parking lots and a number of small single story commercial buildings fighting for parking. The commercial district on top of QA is basically a strip mall with very little adjacent housing.

Overall, I don’t see the greater benefit of deleting the 2 WQA to route more buses through the bottleneck of Queen Anne Avenue and on to SPU.

1. Bruce Nourish says:

“Overall the route has outstanding ridership”

That has nothing to do with the ridership on that leg.

“[E]ach bus on that leg serves about 9 passengers per trip during the middays. That means 36 current passengers per hour will have to hoof over to the 1 or 13. Not likely to happen.”

Huh? There are two buses per hour during the midday on the 2N. Moreover, many of those passengers are deboarding at stops within a quarter-mile of the 13 alignment, a distance that we can certainly expect passengers to walk, a fact that I discussed in that post.

“Sending 8 to 12 buses per hour to SPU is ridiculous. Ridership from SPU is barely higher than what Metro gets on the WQA leg, so based on that fact alone, why does Metro want to quadruple service to SPU? To get more efficient scheduling? If the 3N runs to SPU 4x per hour and the 13 runs 2x/hr, Metro will have plenty of chances to wring out the excess layover time.”

Did you actually read the post?

“The commercial district on Queen Anne Avenue betweend Boston and Galer is an auto-centric hub that needs to undergo a drastic transformation before it can be considered a transit corridor.”

Yes, the supermarkets are car-centric, but if you’ve been on Queen Anne Ave lately (have you?), you’ll see that the west side of the street is in the slow process of redeveloping as a mixed-use residential corridor (which is unsurprising as it’s now zoned as an urban village). I’m also willing to bet that many more people would take the bus to it if the bus doubled in frequency, as riders care more about frequency when making spontaneous short trips.

1. Matt the Engineer says:

The Met Market is about to build up as well. That pretty much only leaves Safeway and 7-11 without a street wall.

2. Bruce Nourish says:

That’s great to hear. It’s also funny to read the “I look forward to leaving this neighborhood” comments. Don’t let the door hit you on the way out…

3. GuyOnBeaconHill says:

Bruce, I read the post, studied it carefully, thought about it overnight and made some comments this morning. And yes, I’ve been on QA regularly.

If the 2 runs at 30 minute headways, then it makes 2 trips to WQA and 2 trips back from WQA. That’s four trips/hour @ 9 passengers = 36 riders/hour losing their service. Whether or not they will walk to another route is debatable, but the rest is arithmetic. Also, your decision to delete ridership at the first and last stops on Galer is somewhat questionable, given that those stops aren’t exactly right at QA/Galer.

4. Bruce Nourish says:

The stops on Galer nearest Queen Anne Ave are about 1/4 mile away from the nearest 13 stops. That is absolutely, positively within walking distance for any rider capable of walking to the existing 2 stop in the first place. In fact, my numbers overstate the strength of the 2N, because almost all the stops north of McGraw on the tail are within a similar distance of the 13 stops on 3rd Ave W.

It’s true that deleting the 2N would leave a small gap in all-day service around 6th & Galer. That’s not ideal, but it’s no worse than many other comparable parts of the city, and the loss to those people will be greatly outweighed by the improvements in service to the much denser 5th/Taylor and Queen Anne Ave corridors.

Moreover, I suspect many of the able bodied riders beyond the standard 1/4 mile radius of remaining stops will, in fact, continue to use the 13 if the 2N goes away. The idea of consolidating routes from neighborhood milk runs to frequent-service trunks is not a new one: it has been done in a number of places around the city, and I am going to write about the results of those changes in future posts.

2. Mike Orr says:

People want to walk to a 15-minute or 10-minute bus. They don’t want to wait for a 20- or 30-minute bus, or walk from one stop to the other to see which one comes next, or sit impatiently wondering if they should walk to the other stop, or worrying that their bus will be late and the other bus won’t be. That kind of aggrevation suppresses ridership. Ballard has the same problem, which is why the 15 and 18 are being consolidated into one route.

1. Eric says:

Extremely well said!

3. Mike Orr says:

“If the top of Queen Anne were to embrace Broadway style redevelopment, there would be a need for more buses.”

If there were more bus service, there would be more impetus to make it a Broadway style redevelopment. Shops depend on foot traffic, and foot traffic depends on frequent buses. If there’s no frequent buses, people are more inclined to drive, and to insist on free parking at the shop.

1. GuyOnBeaconHill says:

Bus ridership is driven by people going from their residence to their place of work or school. Shifting route 3 to SPU is an excellent idea, it connects the school with the hilltop and lots of apartment buildings on the east side of QA, then the Seattle Center and downtown. Current ridership on the 3N, 4N and 13 isn’t heavy, why add more buses?

BTW, Broadway has actually lost transit service since its transformation. Before the change, route 7 ran 6x/hour and route 60 ran 2x/hour. Currently the 49 runs 4x/hr and the 60 runs 3x/hr. Net loss of 1 bus/hour.

2. Bruce Nourish says:

“Bus ridership is driven by people going from their residence to their place of work or school.”

It’s driven by a lot of other things, too. The mindset that buses exist solely to get people from their houses to their work is the source of much of the crap design in Metro’s network.

3. Mike Orr says:

There’s also riders from outside the neighborhood visiting a particular shop. It may be the only shop of its kind, or it may have better quality than similar shops. (A “shop” can be any kind of business or club, including a Zen meditation center or boxing gym.) This is what increases commerce in the neighborhoods and raises the city’s overall employment. Transit provides one part of this; the shops themselves provide the other part. Without good shops, people have no reason to go there. Without frequent transit, it’s too inconvenient to get there, and good shops have little incentive to locate there. This makes the difference between a successful “streetcar suburb” and a decaying has-been where visitors don’t come and residents increasingly turn to cars. Queen Anne Avenue is densifying in spite of the mediocre transit, but it could be even more successful and reach its potential if it had frequent transit as Bruce is proposing.

4. d.p. says:

Completely on point, Mike. Every single word of that.

So why are we so insistent on turning our real Rapid Transit system into the Great Northwestern Mall Shuttle system (with people still bitching about having missed one.

Just how fast do I need to get to how many Abercrombie & Fitches to make my life complete?