Wellington Scoop

Hub and spoke, or point to point?

by Lim Leong
A good public transport network design will invariably involve many design tradeoffs. The key difference between a good, an average, and a bad designer lies in his/her ability to pick the right tradeoff under the shadow of political influence. Make no mistake, politics always plays a part on how a decision is made. Take the Wellington bus network as a real example.

Following are some of the possible design criteria/considerations:

1. To get people from A to B with no transfers (or minimal transfers)?
2. To maximise the number of reachable end points?
3. To reduce traffic congestion on key routes caused by buses and single lane infrastructure?
4. To service destinations where there were no service before?
5. To minimise costs?
6. To maximise efficiency (eg. Passenger loading)?
7. Avoid having to spend too much on the underlying infrastructure?
8. Will the design fit with the longer term public transport plan hence minimising regretful spend?

If you ask a commuter/customer what does she/he want, the answer is probably going to be (1), (2) and (4). If you ask the political stakeholders what really matters, the answer is probably going to be (5) and (6). If you ask a strategic city planner, the answers may be (3), (4) and (8)

There are many stakeholders in a transport network. Some of the design considerations above are mutually exclusive unless money is no object which is not likely. For example, to achieve (1) and (2) simultaneously is going to require a largely point to point design which will mean operational costs will escalate quickly as you add more end points into the network.

So what is the correct design? “Hub and Spoke” or “Point to Point”? IMHO, a “hub and spoke” design is more aligned to criteria (2), (3), (4), (5), (6). Whereas a “point to point” design is more aligned with (1) and (7). From a pure customer’s perspective, the ultimate is always going to be a point to point network from anywhere to anywhere but clearly the operational cost of such network will be prohibitive.

The solution is often a hybrid network combining the best elements of “hub and spoke” and “point to point”. The question of how much “hub and spoke” design to include versus “point to point” design should be a co-design/co-creation exercise between the network designers, customers and bus drivers. The customers and bus drivers can provide a real world perspective on usage and route timing. The real world perspective should be guiding what the design tradeoffs should be.

And finally, politics always play a part on the final decision despite the best intention of the designers.

In summary. There is actually quite a lot of science/engineering behind a network design. Often there is no “best” design because there are tradeoffs and compromises which have to be made. The ability to make the right call on the trade off is what distinguishes a good designer from a bad designer.

Lim Leong was a management consultant with KPMG and is now a senior Programme Manager responsible for complex delivery programmes at Spark NZ. His previous article for Wellington.Scoop was: “Why a Hub and Spoke network will not work in Wellington without the necessary underlying support infrastructure”.


  1. Corinna Connor, 18. November 2018, 9:08

    Day 1 of the new Kilbirnie hub. Only the 18e could pull up at stop C, so the 3 had the choice of blocking the petrol station driveway, or waiting further back. Doesn’t work Sunday 7am…how will it work at peak time? Ingenious planning from greaterwgtn and metlinkwgtn [via twitter]

  2. Morgan Clare, 18. November 2018, 9:11

    It is utterly embarrassing when a group of tourists get on the bus and half the seats have been removed. [via twitter]

  3. Sara Free, 18. November 2018, 9:12

    I was on a route 2 yesterday with seats removed and not enough handrails to hold on to. Quite frankly it’s dangerous. [via twitter]

  4. Peter Kerr, 18. November 2018, 11:10

    Notes from Wednesday 14 November.
    Bus Stop ID: 7014 (Adelaide Road at Basin Reserve)
    17:26 – arrive at bus stop
    17:37 – board bus
    Interchange at Courtenay Place for #2 bus.
    18:18 – Destination reached. Bus Stop ID:5315 (Glenmore Street)
    Time elapsed: 52 mins.
    Equivalent time pre-improvements of July 2018 as per Metlink timetable: 25 mins. Someone has blundered.

  5. KB, 19. November 2018, 10:59

    I like the numbered priorities and how they relate to users/stakeholders.

    Unfortunately for local government, the users will always prefer option 1, and with the looming wave of cheap to operate driverless ride-sharing fleets just around the corner, any “public transport” plans by local government are going to be made irrelevant as majority of Wellington users will choose the door-to-door on demand ride-share offering for only slightly higher cost than a bus/train ticket (without all the negatives like having to go to a stop, being ruled by a schedule, sharing with other humans, cramped seating, unreliable operators etc). I have yet to hear any sort of argument as to why the above will not be the end result. It’s only a question of when the tech is ready, not if.

    Given the above very likely reality – local government should focus their major public transport upgrades on the highest capacity routes where mass transit on dedicated route will still offer big time advantages over ride sharing point to point options. Eg: Airport to CBD rail, CBD to Hutt CBD rail spine.

    Major investments in the bus network to every suburb is short sighted in my opinion, and as soon as driverless ride-sharing arrives, these networks will see major cutbacks as demand rapidly decreases for buses. Electrify buses absolutely – but all these suburban hubs are going to be tumbleweed collectors in ten years time (maybe we will find an alternative use for them – maybe scooter/bike sharing hubs?)

  6. Traveller, 19. November 2018, 11:16

    I do not intend to ride in any vehicle without a driver. And also: there’s no problem about walking to a bus stop.

  7. Jonny Utzone, 19. November 2018, 13:24

    Traveller – have a go on the Docklands Light Rail or the Potsdam tram or the soon to be finished Sydney Metro. None have drivers… The Wellington Cable Car could easily become driverless too. No driver is the face of future of public transport.

  8. Traveller, 19. November 2018, 13:29

    You’ve persuaded me to reconsider. No problem with any driverless vehicle as long as it’s on rails…

  9. Gillybee, 19. November 2018, 15:01

    Historically reform works far better and causes way less upheaval than revolution.

    Lim Leong’s article presumes that the scorched earth policy employed by the NZTA/GWRC/Metlink needed to happen. It didn’t. Why try to reinvent the wheel when all it needed was adjustment?

    The management consultants who worked on implementing the new network lacked not only the specialised knowledge required to deliver effective transport policy but also the practical skills.

    Meanwhile “they” continue to throw hundreds of thousands of dollars of ratepayers’ money at the network in the form of a) international consultants’ fees to fix a problem so needlessly created and b) spin doctors to tell us the new network is “working” with each segment of the hub and spoke counted as a journey. No wonder patronage is “rising”, despite the evidence of our own eyes that traffic congestion in Wellington is at unprecedented levels as Wellingtonians opt out of using public transport.

    In the meantime electric buses get put on the back burner while pollution in the city climbs. The media silence is deafening.

  10. Frustrated, 19. November 2018, 18:29

    I don’t know how they are counting patronage but the traffic into the city is definitely and consistently worse at some times of the day. And as for complaints lessening. Of course they are. What’s the point? GWRC have been told everything they need to know again and again.
    As an aside, I’d like to know how much extra money all these extra ‘fixes’ and PR communications are costing us.

  11. Graham Atkinson, 19. November 2018, 19:08

    Gillybee I can assure you that electric buses are not on the back burner and you should be seeing the results of the testing and road trials very shortly with 18 hour per day all year round operation planned.

  12. Gillybee, 19. November 2018, 19:59

    How many electric buses are we talking about Graham? We currently have 400+ diesel buses running 18 hours a day of which many have clocked odometers. Their Euro rating isn’t worth the paper it’s written on.
    Do tell.

  13. DonP, 20. November 2018, 10:25

    The spin doctors will tell us that they are making excellent progress and now how many electric buses. BUT, are they being used and for how long? We should be asking “how many hours of on-road electric buses and how many hours of on-road diesel (which produce how much contaminants?).” Where are the Greens in all this – they are very quiet.

  14. John Rankin, 20. November 2018, 14:15

    Looking at @LimLeong’s criteria, I’d like to add the following:

    – Provide a reliably frequent service to all parts of the core network (ie on the busiest corridors)
    ‘Reliably frequent’ means at least every 15 minutes, at least 7am to 7pm, 7 days a week. Waiting 30 minutes for 2 buses to arrive at once (one of which is running late) doesn’t count.

    Improving frequency makes connections work better. “The bottom line: Ideally, connections are best addressed through operating trips more frequently rather than holding buses. Yes, this means there will be cases where you just miss a connection. And yes, there are cases where more time needs to be built into system schedules for crucial connections to be met. But in terms of overall systems, transit dollars are usually much better spent on increasing frequency than investing in buses sitting still in the hopes of a connection.”

  15. Pam, 20. November 2018, 23:02

    No worries. We can all bike everywhere. I really enjoyed my bike ride today in the pouring rain, I had a dedicated bike lane all to myself. Friday night I popped down to the supermarket on my bike and picked up the weeks groceries. Took me a while to get back up the hill to Wadestown, but still faster than a bus and I was seated, (well most of the way), when I wasn’t pushing the bike.

  16. Steve zDoole, 21. November 2018, 9:56

    Flat cities can have routes arranged as a Grid. London is a good example. Not really workable in hilly places though.

  17. Glen Smith, 22. November 2018, 1:51

    KB. The reason why ride-sharing (whether driven or driverless) won’t be the long term (or even medium term) predominant transport mode is congestion. Congestion in Wellington is already high and predicted to rise around 90% by 2041 due to a predicted 20% increase in total passenger trips. The consensus is that ride-sharing will exacerbate this congestion even further. This is partly due to mode change from PT to cars but also because ride-sharing requires significantly more trips to achieve the same transport outcome as individually owned cars. A simulation of a universal station based ride-sharing system concluded that 38% of trips would be empty due to return trips and rebalancing the network. An individual based ride share system (like Uber) would be even worse due to the added effect of drivers prowling the city for their next fare.
    As our city descends into the inevitable resulting gridlocked quagmire, people will come to recognise the obvious fact (long recognised overseas) that the only logical way to solve congestion and promote a liveable city is by moving transport trips from cars to more efficient modes (rail/bus/cycle). They will start to examine the enormous subsidies that car users currently receive (over 50% of the societal cost of each car trip) and will start to demand that car users actually pay their own way (likely by congestion charging) with funds being transferred to PT to produce a more level playing field.
    So as future ride-share customers travel at 5km/hr (or less) on choked roads while paying their expensive fare/ congestion charges, most people will cruise past on the cheap (or possibly free) high quality PT network that the central and local government are sensibly planning (fingers crossed they can do a better job than the bus debacle).