A key aspect of ‘the network effect’, which I have blogged about quite a lot in recent times, is the need to simplify route structures. Auckland’s current bus system looks very much like someone threw spaghetti at a map, a million different routes going all over the place – with the key thing being that because there are so many bizarre routes, they can only be run at fairly low frequencies. But there are other disadvantages of having a complicated route structure:

  • It’s really confusing. One wonders how many people are put off using public transport because it’s simply “too hard”. For my local route it was damn difficult even finding where its CBD bus stops are, it bizarrely uses Hobson Street, it changes its name and route in the evenings and at weekends, some buses terminate at Herne Bay, others at Westmere, and so on.
  • Having many different route numbers probably leads to unequal loadings. People flock to a particular service because perhaps it’s a bit quicker as it skips a detour, or perhaps it goes via a minor detour that gets people 50m closer to their house, but at the cost of 5 minutes of travel for everyone else.. Unequal loadings is not an efficient use of the resource.
  • The legibility of the route becomes unclear. There are distinct advantages in intuitively being able to understand the public transport network without needing to research it. Why does the 007 use the Mt Albert crosstown arterial for some of its length, but the Balmoral crosstown arterial for other parts? Why does the 008 wind its way through back-streets rather than sticking to an obvious and legible arterial?

So there are two main elements to what I call “route simplification”, one is to get rid of the vast array of route numbers that are only ever-so-slightly different to each other: and the other is to make the route more easily understood by its users. If we look at the Sandringham Road bus routes, I think that both methods need to be used to simplify this route. Let’s start by looking at its map:

So we have 10 different routes that constitute the “Sandringham Road buses”. The 202, 240, 241, 243, 246, 247, 248 and 24F. Even trying to make some sense out of the difference between these routes is a challenge: at the city end the 202 goes via Bond Street and Great North Road into the CBD, the 240 and the 24F go via Ian McKinnon Drive (which means that the routing on the main map is actually wrong) and the rest go via Symonds Street.

The 24F and the 240 differ only in their stopping patterns: in that the 240 starts its run at Mt Albert Road whereas the 24F is non-stop between Mt Albert Road and K Road (I think). The 241, 246 and 247 seem to be the most obviously stupid routes of the lot, in that they’re exactly the same except for slight differences in where they end (and slight differences in the route that the 241 takes going via Owairaka Ave). The 233 goes via St Lukes, but is otherwise identical to the 243. The 248 takes the most higgledy-piggledy route imaginable through the back-streets of Blockhouse Bay before terminating there, while the 249 for some bizarre reason continues on from Blockhouse Bay to New Lynn – although you would have to be absolutely retarded to ever even consider taking the 249 between New Lynn and Midtown, as it spends half its time heading in completely the wrong direction.

If this was an isolated example of the completely unnecessary complexity of Auckland’s bus network, then perhaps I would think that it was just funny. Unfortunately, almost every single route in the city is stupidly and unnecessarily complex like this Sandringham Road example.

So what can we do about this? I think if we start at the city end and work our way outwards it should become obvious that these ten routes can be simplified down significantly. To start with, having both the 202 and the 240 as ‘alternatives to Symonds Street’ seems pretty unnecessary. Having caught the 202 a few times, it seems that it basically serves as a “bus to Auckland Girls Grammar School”. Perhaps that means it’s necessary to retain, but certainly I don’t think we need both the 202 and the 240.

Working our way further out, the next snag is the “what the heck to do about St Lukes?” problem. While the 233’s detour to St Lukes is useful if you’re visiting the mall it’s a damn pain in the neck if you’re not. Hopefully in the longer term we will be running very frequent cross-town buses along Balmoral Road so this detour won’t be necessary, but in the meanwhile probably our fundamental off-peak distinction along Sandringham Road will be between St Lukes buses and non St Lukes buses. It’s annoying, but because the mall does generate significant patronage there’s not much that can be done.

Further out, we see that the routes start to diverge like mad after the Sandringham Road/Mt Albert Road intersection. As I mentioned above, I think the 241, 246 and 247 could be completely eliminated and folded into other routes as they add nothing but complexity to the system. The 249 is similarly pointless, as the Portage Road, Bolton Street areas will (hopefully) be served by local routes feeding into the New Lynn train station from its surrounding areas.

Which leaves us with two main routes: the Tiverton/Wolverton route of the 233/243 and the higgledy-piggledy route of the 248. Now I always thought the 248 was completely necessary as it was “the Blockhouse Bay bus”, but actually it turns out there other buses go to Blockhouse Bay: most notably the 258 which uses Dominion Road. Comparing the times between Blockhouse Bay and the city along both the 248 and the 258, and it does seem the 258 is a bit quicker (and hopefully if the Dominion Road bus lanes are improved and not destroyed, that speed advantage would become even greater). So do we even need a Sandringham Road bus route that goes to Blockhouse Bay? Perhaps not, especially when the Green Bay buses provide an “along Blockhouse Bay Road” bus route (although they’re another example of a series of routes that desperately need simplification and better frequencies:

What all this effectively means is that we can have one core Sandringham Road route: from New Lynn to Britomart (I think it should go to Britomart rather than Midtown so enable connections with the rest of the network). This is shown in the map below:

During the opening hours of St Lukes Mall probably every second off-peak bus would travel via the blue detour, at least until we figure out a better way of dealing with this issue. And during peak times probably a few buses could travel via the green route rather than the red one. But otherwise pretty much every single bus would be travelling this core route, which is “anchored” at each end strongly by the CBD and New Lynn (the reason I have it terminating at New Lynn is so it’s useful for passengers wanting to travel there from along the route).

Apart from a rather annoying detour to ensure that it passes through the Owairaka shops the route is also pretty damn straight, which means relatively quick travel times. At peak hour I think you could run some of your services as “express”, meaning that they wouldn’t stop on the city side of Mt Albert Road, and complement them by running services that start at Mt Albert Road, but ultimately that is just messing with stopping patterns and the route simplicity itself is maintained.

There does remain one rather under-served “hole” around Boundary Road, but otherwise I think between this route, the existing 258, the Green Bay routes and the 009 cross-town route the area is fairly well covered, and in a far more simple way than we currently see. I think we end up with four route numbers (core route, route via St Lukes, shortened route between Mt Albert and city and route via Bond Street and Auckland Girls’ Grammar) plus an express version of the core route. That’s a lot better than what we have now I think.

Now we just need to get bus lanes along the whole thing!

Share this


  1. Maybe they could institute a rule that said every seperate route number must be drawn with a seperate line on the timetable maps, and if it gets to the stage where you can’t tell what road the bus is actually running on then the routes need rationalization 😀

  2. Makes sense axio! Also makes me wish that Auckland had a simple grid road structure.

    I’m not quite sure if I’ve got this 100% right. A few questions:

    1) Do we really need to provide a “202 route”? Could the AGGS students just walk?
    2) Do we really need to link to St Lukes? Could people just walk from Sandringham Road?
    3) Do we need to find a better way of serving the Boundary Road area? It’s about 1km between this bus route to the north and where the 258 goes to the south.

  3. One thing to make your route better is the Sandringham Rd Extension that was built as part of the SH20 section through Mt Roskill. For some reason the motorway isn’t showing on google maps however if you look at the satellite images you will see it there. Basically Sandringham Rd carries straight on after Stoddard Rd for a few hundred meters where it hits a roundabout and then travels northwest slightly to link in with Maioro St. This link cuts the about journey around Stoddard Rd in half as it is only about 750m compared to 1.5km for the route you show. Whats more because it of the way it is built it has less lights to contend with so it is probably even faster still.

    I don’t know what will happen to this once they finish Waterview but I hope they keep it in some form.

  4. Yeah I did think of that possibility Matt, but I do think it’s worthwhile taking the route through Owairaka shops at that is meant to be a future development node, and it could be a fairly good passenger generator. There’s always the balance between serving as many areas as you can, but keeping the bus route as quick as you can.

  5. Mapping Auckland routes is made more complex than it needs to be by using that ugly Australian practice of giving separate route numbers to every single route variation and short turn.

    For example other networks hide some of this complexity by using the same route number for all trips on a line regardless of whether some stop or start partway along the route. The destination name on the bus tells the route end point and usually if someone does catch the short running version the wait for the next bus that continues on will be short.

    Even special workings of a route that divert off the main route can be given the same route number with a comment code or letter if the route variation serves the same stops as the base route and no one would be stranded by accident if they took that bus.

    Hell if most rail network maps gave a diferent route designation to every special working, express version and short run they would be hugely complex.

    I do agree that Auckland could do with route simplification. However some gains could be made just by packaging it better. There is some good guidance on this in the HiTrans 2 manual.

  6. LX, do you have a copy of those manuals? I’ve been trying to track them down for ages (without having to spend near 100 euros one them).

    Miggle, good point. Just turn it into a school bus I agree.

  7. HiTrans 2 is fantastic. I have access to a copy at work. Great graphics to explain a range of network design principals that make them clear to anyone. I think you would find it a money well spent if you can get a hold of it. It covers and supports most of the issues you cover. I can’t comment on 1, 3, 4 and 5 as I haven’t had my hands on them.

  8. As someone who used to live in Kingsland I would alway use the 240, 202 or 205 i.e. Bond Street/Great North Road services. They are WAY faster than routes which go up Symonds Street and past the University. Real bottlenecks on that route are on the approach to Mt Eden Road and Waterloo Quadrant in town.

    I whole heartedly agree with you Jarbury on reducing route complexity. However what is the larger picture here? Why would we want a service between CBD and New Lynn, the network effect is based on a grid, not the outdated radial model our current bus routes follow. Surely the grid should be north/south and east/west, not some strangely skewed grid across the isthmus?

  9. Scott, I think the solution to the bottleneck issues with the Symonds Street route could be solved by having bus lanes along that top of Symonds Street stretch, and also along Waterloo Quadrant (or simply making the buses terminate at Britomart). I used to catch the 240 quite a lot when I lived in Sandringham and worked in town, but that was because at that time the Central Connector was under construction and Symonds Street was a mess. Now that it’s up and running it makes sense to utilise it, plus the university is a huge drawcard for patronage.

    In terms of what kind of grid to create, as Auckland doesn’t have a perfect grid street system I think we need to be a bit more flexible, particularly in the outer-isthmus areas. I have chosen the route to terminate at New Lynn because I think it’s a strong anchor and can attract patronage from other areas along the line – ideally you want good counter-peak patronage as you need to shift the bus and driver along that corridor anyway. New Lynn is meant to grow a lot in the future, both for residents and as an employment hub. While the service does run from New Lynn to the CBD, in my mind it doesn’t really duplicate the western railway line that much, because it serves a quite separate corridor. I would be very surprised if people caught a bus along the whole route.

  10. I think a grid made of radial and circumferential lines is perfectly appropriate for Auckland, operationally this would achieve the same thing at a north-south square grid. Trying to put such a square grid over Auckland would be kind foolish, as it would bear no relevance to the existing routes on the ground nor the land use patterns. Fact is that most travel of any kind in Auckland follows a radial corridor (not necessarily to the CBD of course, but along some radial corridor), because the city has grown along radial corridors.

  11. So any ideas when AT will undertake such a review? I’ve always hated Auckland buses, and generally always end up asking the driver to make sure he/she is actually going where I want to go, mainly because of their inherent complexity and the endless list of bus numbers.

  12. I’ve talked with people at ARTA about the bus routes and there appears to be a “rule” that nobody should live more than X metres from a bus stop … well, these people have never been too happy about that but didn’t have the mechanisms to change the routes (or bus stops). At leas the will is there!

    I personally think that the issue of the bus routes is very related to the integrated ticketing thing. For a “network” of buses to be effective, the cost of switching buses must be ZERO. The only thing that “may” control the fare is where I’m going to and from … the route I take must be irrelevant… example:
    I want to go from somewhere around Parrs Park in Glen Eden to Britomart.
    The options I (fit an healthy) have are:
    1.- Walk to the train station (Sunnyvale or Glen Eden) and take the train in ($5) (train trip ~45min)
    2.- Take a bus to Henderson (+5min) or Glen Eden (+25 min) and take the train there ($1.7 + $5 = $6.7) (train trip ~45min)
    3.- Take a bus to New Lynn (+30min) and take the train there ($1.7 + $4 = $5.7) (train trip ~35min)
    4.- Take a bus straight to the city ($5.5) (bus trip ~60 min)

    And this is for a “commuter” kind of trip that the current system is set out to satisfy!

  13. The Central Connector has sped up 240s no end…

    I haven’t posted for a while Josh and I apologise for that but I’ve been thinking about this problem a lot and what I’ve come up with is that the big problem of moving to a network effect is the usual inertia but mainly also contracts and commercial runs, NZ Bus et al don’t want to kill the goose that lays the golden eggs, so how to rearrange things so all profit..?

    Here’s what I think:

    – Research the network effect to show that we could easily double or triple bus patronage using existing resources
    – Design and plan about 100 routes Auckland wide with timetables integrated with each other and the future 10 minute rail frequencies
    – Introduce zonal fares and a paper integrated ticket (if electronic not completed)
    – Ask the Bus Companies to voluntarily cancel their existing contracts and claims to commercial runs
    – Replace these contracts with tendered routes that cover all the Bus Company’s expenses plus a 3% profit
    – Add a clause to the contract that if one of these 100 hundred routes makes a profit on top of the tendered contract expense (and they will) the Bus Company operating that route receives 25% of the profit, if it makes a loss then they pay nothing

    This provides an incentive for the Bus Companies to change, to lobby for improvements (such as more bus lanes, faster times all day equal more profits) and to work together with the council and we as the consumer will get a much better deal…

  14. Interesting thoughts Jeremy, and I do think that we need to get the bus companies on-side somehow. Hopefully some people at NZ Bus are reading this blog!

    I plan to do similar posts to this one for all the “core routes” that I think would make up a network: probably just sticking to the isthmus for now as that’s what I know best. The idea then is to stitch it all together into a rather more refined version of the network map I did a few weeks back.

  15. Maybe we could establish a set of operating criteria and target one region of the city each and work them together for a comprehensive network plan, I could revise my plan for the routes on the Shore for example.
    We would have to get some good patronage and travel demand data, but building our own model of a transfer based network might not be an impossibility.

  16. If I remember correctly NZBus tells you how many buses they have at each depot on their websites. Assuming the don’t own much more than they need you could probably estimate it from there.

  17. Hi,

    I have found these posts about simplifying bus routes very interesting, and agree that the network is a crazy bunch of spaghetti.

    There are just under 400 routes in the whole network, this includes bus, train and ferry. For buses, the total is around 380 I think.

    In reply to Scott, fleet lists for Auckland buses are available on this site,

    The major operators are NZ Bus(Go West, North Star, Metrolink and Waka Pacific), Howick and Eastern, Birkenhead Transport, Ritchies and Urban Express. Typically an operator should have enough buses to run all their services, school services plus 10% extra vehicles in case of maintenance or a breakdown.

    Looking at the fleet lists, the major operators have about 1033 buses between them. So take away 10% and you have around up to 900 buses running. However, this does include a number of buses which would be allocated to school services.

  18. Looking at the fleet lists at http://www.tranzinfo.net.nz, the major companies NZ Bus, Ritchies, Howick and Eastern, Birkenhead and Urban Express have a total of about 1033 buses. Operators should have a 10% surplus of buses to cover breakdowns and maintenance so there should be about 900 buses operating each day. This includes school services too and does not include operators who only run school services such as Bayes.

  19. Whangaparaoa has a total of 8 separate bus route numbers on esentially two bus routes which mimic each other. Route 896/897/996/997 terminate at Manly and 898/899/998/999 at Army Bay. Terminate them all at Army Bay or Gulf Harbour ferry terminal and have one run to Orewa/Waiwera and the other to the city via the Northern Busway and you have a workable and simple system! Chance of that ever happening? Well, I aint holding my breath!

Leave a Reply