A few weeks ago I took a look at public transport ridership in 2024. In today’s post I’m going to be looking a bit deeper at bus ridership.
Buses make up the vast majority of ridership in Auckland with 70 million boardings last year out of a total of 89.4 million. That’s about 76 percent of all trips, quite a bit higher than the 65 percent of trips it was prior to COVID thanks to bus usage recovering faster than other modes. Boardings are currently at 93 percent of what they were in February 2020. By comparison, trains are at 65 percent and ferries at 79 percent.
So let’s break these numbers down a bit, which we can helpfully do thanks to Auckland Transport regularly publishing a bus performance report which contains data on all services provided by AT but does not contain data on school services, buses-replace-trains services, special event services, or exempt services. I used this data back in 2020 to look at our busiest bus routes prior to COVID and some of the analysis I did back then has been updated with the latest numbers.
Bus Ridership by the type of service.
Auckland Transport have long broken their services down by layers based on the minimum level of all day/all week frequency a service has. By and large these classifications are a useful high-level way to understand the quality of a service but there are always some exceptions. One good example is the 866 runs from Albany to Newmarket along the Busway and then via Ponsonby and the hospital. In the morning peak, in the peak direction, it runs as frequently as every 7.5 minutes, better than many all-day frequent routes, but it doesn’t run that frequently counter-peak or off-peak and doesn’t run at all on a weekend meaning it’s only classified as a local service.
AT present much of the data using these classifications already but I’ve made a few tweaks to split the NX1, NX2 and WX1 out to the Rapid category. A few of the minor ones, like On-Demand, as well as the figures for services not included in the report, such as school services, have been grouped into the ‘Other’ category.
In total over 60 percent of all trips take place on Rapid or Frequent services and notably on the whole these services are back above their pre-COVID levels. In many ways this is a good advertisement for the power of the frequent network, that people will continue to use public transport where it’s frequent and convenient to do so.
The biggest drop in usage comes from Peak Only services as changes in usage since COVID mean they’re often just not needed as much anymore. In some ways slightly less peak usage can be a good thing as those services were often some of the most expensive to provide due to needing additional buses and drivers just for a few hours a day. The resource freed up by a slight reduction in peak demand likely supports a lot more service elsewhere.
Bus Ridership by Corridor
One of the issues with the data is that it is for each individual route. As an example, instead of just being 25 that runs on Dominion Rd, it has both the 25B and 25L separately, as well as the peak only 252 and 253.
To get a better idea as to our busiest corridors, I’ve grouped the routes together where possible. As a general rule, I’ve grouped any services with the same base number or that share the same corridor for more than 50% of their length. There are a few exceptions to this to this though.
In total there were 188 different routes that ran at some point in 2024, though that includes a few where a change has occurred, for example the 650 was upgraded to the 65 but both get included in the data. I combine those routes down to around 146 different corridors and the top 60 of those are shown below. These top-60 routes represent over 80% of all bus trips.
Here are a few things that stand out to me:
- As with last time, the impact of the busway is huge and accounts for over 10% of all bus ridership. Even if they were separated out, the NX1 is clearly the busiest single route and the NX2 would be the third busiest after the 70.
- The 70 has long been the second busiest corridor and that’s in part because it’s serving a few different trip types, such as a feeder to trains at Panmure and Ellerslie Stations for a faster trip to parts of the city, is used for trips to the University, it serves trips along the Gt South Rd corridor and also shopping trips to Pakuranga and Botany. The route will also be the prime service using the Eastern Busway once it is completed.
- Arguably some parts of the Gt North Rd corridor are even busier than suggested here as some of the key Northwest routes, like the 11T/W also use the section from Waterview to the city.
- The Northwest corridor has been the biggest mover in the order following the changes in November 2023 which saw the WX1 introduced. In this chart it has moved from being the 13th busiest corridor to the 9th – a little bit more on the Northwest later in the post.
- The Pakuranga Rd corridor (the 72 group of routes between Howick to Panmure) is the highest used corridor that doesn’t go to the city centre (assuming most 72X users are also going to Panmure). This is a great example that people are prepared to transfer.
- Likewise, the 83 (Takapuna to Albany via Browns Bay), does well and it too doesn’t touch the city. I suspect most of the usage is flowing to/from the busway stations.
- The 120 (Henderson to Constellation Station via Westgate) is the busiest non-frequent route and outperforms a number of existing frequent routes. It has been on the plans to upgrade for a few years now but looks like it will finally happen in April.
- The 38 is the least used frequent route (that has existed for a full year). This was previously combined with the AirportLink as the 380.
I’ve also taken a look at how these routes have changed since COVID (not all existed prior to COVID). Here you can see that while the busway is still far and away the busiest service, it’s also seen the biggest fall in usage in total terms with 1.2 million fewer trips compared to before COVID. However, while that represents a drop of around 14 percent, the biggest fall percentage wise with the CityLink was down nearly 35 percent on pre-COVID.
The 74 shows the strongest growth. This route, which links Onehunga to Glen Innes via Sylvia Park and Panmure, was upgraded from the 743 to be frequent in March 2023.
Ridership by region and destination
Auckland Transport group routes into different regions and this shows how ridership by region has changed. The West is the standout here with a big change since the WX1 and other Northwest changes were introduced in November 2023.
Compared to pre-COVID, the routes in the North have had the weakest recovery.
I’ve also noticed in other data that ridership seems to be at or above pre-COVID levels on weekends with weekdays holding us back. I wanted to test the hypothesis that it is routes that start and end in the city centre that are not doing as well. So for each route I assigned whether the route travelled to the city centre or not.
It seems the hypothesis is correct. City centre focused routes still carry more people overall with 37 million boardings vs 29.3 million for buses that don’t go to/from the city. However, those buses to other parts of the region are above pre-COVID levels while the city centre ones are only at about 86% compared to Feb-2020.
Comparing the Rapids
Finally, the introduction of the WX1 has clearly had a big impact on the usage of services in the Northwest and I can’t help but see the similarities between it and the Northern Busway when it first opened.
As a quick reminder, the first Northern Express services started operating at the end of 2005 using just the Albany and Constellation stations and running on bus shoulders on the motorway south of Constellation while the proper busway was being built. The proper busway along with the three other stations were opened in February 2008.
The WX1 is somewhat similar, running along the motorway shoulders and only stopping at a few stations (Westgate, Lincoln Rd and Te Atatu) and the stations are of much lower quality than Albany and Constellation were.
As you can see below, the WX1 service is currently running well ahead of where the Northern Express was at the same time. If only the actual busway was already under construction.
Boardings per Service Hour
Looking at pure boardings is important but it’s also useful to compare that to how much resource is being thrown at the route. For example, the Outer Link carries more people than the Inner Link but it also has a much longer route. So AT have included this metric to give an indication of that. The service hours for each route is the scheduled ‘in service’ time of all services on weekdays i.e. if you have a bus that takes 30 minutes to run and it it is run four times a day, that’s two service hours.
The Busway is the best performer in this metric too, averaging over 60 boardings for every service hour it operates. Notably this is about the same as it was pre-COVID which shows that the level of service provided has been scaled down to meet current demand rather than just running all of the same service that used to be run. The same applies for most other services too and in most cases the boardings per service hour remain fairly consistent with what they were before.
There’s so much more that can be gleamed from this data so if there’s any specific analysis from this data that readers really want to see, let me know and I might add it depending on how much work is involved.
If you add Dominion, Sandringham, and Mt Eden roads together, most of those users would be within a reasonable walk of Dominion Road surface LR. Then add in all the transfers that would happen from the crosstown routes, and the increase in users due to much shorter travel time, and it could overtake the busway. And all they need to do is put back some tracks that were ripped up 70 years ago.
Yes all they need do is prevent any intensification in the entire corridor for another 10 years while they dick around with overly expensive options of heavy light rail, light heavy rail, light metro, or insert whatever BS takes your fancy, none of which will ever be funded. Hope you like what you have because that is what you are getting, the money all got spent digging some holes for CRL.
Or another option is to replicate the Dominion Road part of LR with buses. Almost all the gains of LR have nothing to do with rails and everything to do with less stops, direct route down Queen Street, 24×7 dedicated road space, traffic light priority, all door boarding, etc. Just need a bit of paint, some signs, better buses and a few changes to traffic lights.
But then the number of passengers require a large number of buses per hour and the length of platform needed in the city centre becomes significant. You might need to tie several buses together to carry and load that many people. I wonder what that might look like?
What happened to Simeon Brown’s trackless trams?
Symonds Street is already handling Sandringham Road, Dominion Road, Mt Eden Road, and many others. Using Ian McKinnan and Queen Street for Dominion Road buses would probably be less busy than Symonds Street currently is.
Until the busses reach their destination and cause a log-jam of busses in the CBD (one of the major problems surface level light rail was designed to overcome).
Their destination could be Albany…
Agree, to an extent. BRT would be a good start if not a long term solution.
The signal priority thing doesn’t really work when you have more than about 15 vehicles per hour, nor do in-line stops work, which is one reason to go with LRT because you can move a lot of people in 15 vehicles per hour if they’re 66m long each.
Agree traffic light priority doesn’t work at peak, maybe off peak though. There aren’t that many lights between Mt Roskill and K Road so I doubt it would have much saving.
In-line stops could work with all door boarding to speed up the process and less stops.
Jimbo exactly. And shows how entirely moronic it was to ignore the unchanging geographic fact of Dom Rd’s centrality through the isthmus and propose to send rail down Sandringham instead.
Dom/Queen St is one of world’s the ideal surface LR routes in the world. It will happen.
Thanks for the data. Can we also look at the money matter, ie, how much money buses services earned versus how much money spent. How much money subsidized by Auckland Council (our money). By looking into the money side then we can has a better understanding of Auckland Bus Services. By looking into the matter, can we also look at the load factor of bus services, ie, how is the usage factor vs how much seats available. By looking into this, we can also know better which and where should be improve.
Sorry by just asking, but we should also look at the financial side as well for NZ cannot just do business by wishes at the cost of the residents.
Thanks you!
>Can we also look at the money matter, ie, how much money buses services earned versus how much money spent. How much money subsidized by Auckland Council (our money).
And then we can compare the subsidies on the Busway routes with how much AWHC and all the road widening that project requires would cost, and realise that providing good public transport saves us money.
From memory the Busway makes an operating profit.
Can the figures more clear? We just can’t make an assumption. Bus services is a public service with money support from Auckland Council, the financial matters must make it clear and regularly disclose to public.
Bus service is a public good. Do you expect parks to make money or want numbers for maintenance released continuously?
Hum, that’s basically boardings (ticket revenue) by service-hour (operating cost).
Anything over about 20 boardings per service hour is making money, more or less.
My local bus doesn’t seem to hit any of those minimum frequencies lol.
Would be interesting to see the opposite of this. Those routes that aren’t used
Service hours is an expression of cost. While there will be some differences from contract to contract, by in large the cost of a service hour will be similar across the region. Therefore the boardings per service hour chart is useful for understanding the efficiency of a service.
But note, AT also do provide some services they know will be lowly used but are provided for equity purposes e.g. serving small rural communities. This is part of the reason they have the different service classifications.
Occupancy levels of transit services is important information for understanding resource efficiency, though is more meaningful to measure whole service occupancy average, rather than obsess about that one empty bus someone saw.
Of course we do already know that average car occupancy struggles to break much beyond 20%, which is extremely inefficient. And driving is heavily subsidised service, through rates, taxes, and negative externalities.
Contrast with bikes and scooters which are always at 100% occupancy and sometimes beyond it! While also being much more spatially efficient, causing little wear and tear, and being minimally funded.
+1000
As no doubt is the intention, these statistics prove that Light Rail is justified. The NX are popular because there is no other option for the North Shore. Dominion Road is popular because it should have Light Rail on it’s surface. The 30 and 309 are popular because Mangere has no good connections, and the Onehunga train is infrequent.
Everything still screams Light Rail, but our last government was too distracted to make it real, and now we have the car people in charge again.
Sad for our kids, good if you own a bus company, but irksome if you believe in efficient, rapid, mass transit.
bah humbug
Your comments about the Manukau Rd routes aren’t even close to being true. There are few if any people on the 30 by the time it gets to Onehunga and it’s similar numbers on the 309. Nearly everyone has gotten off either on Manukau or Pah Roads. This would justify a light rail route at some point.
Most passengers in Mangere already catch a more frequent bus service to Otahuhu or Papatoetoe train stations (or get dropped off there) and train into town (especially outside of peak times), rather than endure the 30 minutes minimum journey to Onehunga.
Yes 309X does enjoy good patronage but it’s a limited service which is able to make up time through use of the bus lanes.
The plan is for the 120 to become the 12 and for the 902 to become the local service starting at Kyle Road through upper Glenfield towards Constellation and the 12 to use the motorway from Greenhithe to Constellation. More info here: https://haveyoursay.at.govt.nz/greenhithe-bus
I think that the 12 will double the current numbers, if the service becomes a lot quicker between Westgate and Constellation. It should be feasible to get to the city from Greenhithe in 35 minutes, which is a massive improvement.
Interesting analysis.
Since you ask, does the data allow analysis by passenger kilometre? Your story about the success of the 70 related people making short journeys along its route. How would your comparisons stack up if you used PKT per route rather than boardings per route?
Going further, is there any data that helps us understand the effectiveness of the overall transport system? It would be helpful, one would think, if analysis showed if the expected inverse correlation between ‘boardings’ and ‘delay’, or perhaps ‘PKT’ and ‘delay’ exists. In this case I would suggest including both bus and car delay to calculate total person delay, in the analysis of congestion.
There is no passenger km data that has been released at this level, only as a mode total.
There is however reliability and punctuality data at a route level
Yes, I use the 18, 195, 11T or 11W, whichever comes first. All take me home (near MOTAT). I only need to wait a few minutes most of the time. So these all should be included in the Gt North Rd corridor.
Makes sense to separate them as they cover two quite different areas once the separate. The 30 and 70 share the same corridor for the inner part of their route too.
I’ve also been slicing and dicing the patronage stats, dividing them generally along the same lines as Matt. In particular, I’ve been looking at which routes are “growth” routes and which are holding steady (or even declining). I’ve focused on comparing the 12-month rolling total passengers for each route in December 2024 with that in December 2023. Route changes etc make this an inexact comparison for some routes, but subject to those caveats . . .
Amongst the frequent services, the big mover is Route 11/110. It’s not an entirely fair comparison because of the route changes etc, and the fact that Route 18 duplicates part of it. But that aside, Routes 11/110 are up 127% year on year, and not at the expense of Route 18, which is up 16%.
A surprise for me was Route 20 – up 36% year-on-year. It’s still off a low base, but the growth has been consistently high over the last two years, suggesting there’s long-term growth potential here.
Even more spectacular was the growth on Route 64, up 54% year-on-year. The route has been growing steadily at more than 20% pa for the last couple of years, and the pace of growth appears to be increasing. I’m not sure if this route is still free – or even whether the majority of passengers know that it is?
Other big movers include Routes 670/67A/67B (up 30%, with minor impact as yet from frequency changes, so more growth likely in the future), Route 68 (up 20%) and Route 83 (up 22%), and the Airport Link (up 22% but still off a low base).
An interesting feature of these routes are that none of them (aside from 11/110 and the City Link) touch the city centre – most of them are cross-town routes which have been introduced since the New Network was implemented. Kudos to the team at AT for identifying these routes and making them work.
Most of the radial routes serving the city centre are also growing, but at a much slower pace, generally 10% pa or less (though Route 27 is a standout at 17% growth). On the less spectacular side, Route 32 patronage was flat, Route 70 recorded just 1% growth, and the Outer Link patronage dropped 1% (almost certainly due to the truncation of the route). It’s too early to be giving year-on-year comparisons of the WX1; but the NX1 grew at 13% and the NX2 at 10%.
Among the “connector” services, there are a number of North Shore routes that have had quite spectacular growth (845, 885, 901, 906, 907 and 926 in particular – all in the 50-70% pa growth range, though the growth is slowing from its peak around mid-2024).
Growth generally seems to be slowing from the post-covid surge, but it’s still very significant growth. Any other business that’s growing at 14% on average year-on-year would be considered worthy of serious investmnent – especially by a government that claims to be business-focused. How about it, then?
94 (942) and 941 Takapuna – AUT – Northcote – Highbury – Beach Haven – Glenfield – Wairau Park – Smales Farm – North Shore Hospital – Takapuna seems to be doing well, with their onward connections to City and elsewhere. Thanks to Cllr Chris Darby for backing that 2-part loop.
I’m not surprised that 32 is flat even though they is huge building growth in the area. It constantly gets stuck in traffic and not just at peak times. There are almost no bus lanes or bus priority on Massey Rd, even though they is room in some places. It’s a slow journey for not a huge distance.
Taking the 325 is probably just as fast.
The issue with just looking at growth is 2023 was impacted quite a bit by the bus driver shortages and they weren’t even across Auckland. Going forward that will likely be more appropriate.
I’m also skeptical about claiming growth when usage is still below what it was pre-COVID (except for the individual routes that are above it.
Yes and no. Around 7% of services were cancelled in Feb 2023, and the shortage of drivers was declared “over” by July that year. So while there was an impact, it was by no means significant when compared with some of the growth that has been observed since then.
And yes, we are still in the “recovery phase” from Covid – but growth in numbers is still growth in numbers. Given some of the structural changes in work patterns during and since Covid, I’d suggest that passenger numbers pre-Covid are of much less relevance than numbers post-Covid.
One reason that Route 18 is bucking the radial to CBD trend is that is probably carrying a good chunk of demoralised Western Line rail customers – I am one of them. In the afternoons, Western Line trains are timetabled to just miss my onwards bus connection. In the early morning, Route 18 is typically faster than the train from New Lynn to Downtown. In the PM peak, despite struggling to get out of the Albert St-Wyndham St-Hobson St-Pitt St mess, the headways on Route 18 are way better than the Western Line.
CRL will change this question for many west Aucklanders – can’t wait.
Yes will be super interesting to see CRL impacts on certain routes.
Being from the South I’m always surprised to see how the 33 stacks up. Considering it’s a key service local service, follows the train line for much of it’s length, and with basically zero bus priority provided.
Would be great to see more bus lanes / signal advances put on for the 33 and the cross-town 66.
33 bus Would be more popular if it continued to city . I missed that bus when it was numbered 471 472.
No way – that would be doubling up on the rail line – Central city trips can be accommodated easily by the train, and CRL will have multiple city centre stops, making it much more appealing to many people.
There’s so many easy transfer points from the 33 to rail, Manurewa, Manukau, Otahuhu.
Symonds Street is already handling Sandringham Road, Dominion Road, Mt Eden Road, and many others. Using Ian McKinnan and Queen Street for Dominion Road buses would probably be less busy than Symonds Street currently is.
Was meant to be a reply to an above thread.
I’d love for the 856 to become the 85. I recently spent $40 on an Uber to get to Albany because I couldn’t to wait 25 minutes for the next 856 or 861.
We caught the NX1 into the city on Sunday, was rather surprised the bus was 80% full leaving Albany – it isn’t usually that busy when I catch it on a weekday!
856 won’t become 85. It is too long and complex. At half-hour intervals, it often ends up 15 minutes ahead or behind schedule due to having the most stops of any regular Auckland bus route (I haven’t checked that statistic 100%, but I’m fairly sure). Stopping at every stop slows it to a ridiculous extent compared to stopping at none of them, varying between 45m and 1h20m. There are already short routes solving the infrequency of that route, meaning that it’s covered at a 15-minute frequency during peak hours. That’s probably the best it’s going to get.
Most “frequent” routes carry around the 500,000 pax/year mark or greater; the outliers are Routes 20, 38 and 64 which are in the 300,000 pax/year bracket (though 20 and 64 are growing rapidly). There are a number of “connector” routes which fall in the 400,000 pax/year bracket, and which may be candidates for conversion to “frequent”. These include:
105 – at 475,000
195 – at 576,000; but it does however duplicate the 18 and 11 along Great North Road
325 – at 481,000
814 – at 580,000
856 – at 511,000
861 – at 417,000
866 – at 540,000; this was under the previous RPTP scheduled to become the NX3, but that plan appears to have been dropped. Two other factors, though: (a) many Inner Link passengers also use this route which would artificially inflate the numbers; but (b) it operates only weekdays, and would probably gain additional weekend passengers if it became the NX3
917 – at 476,000
941 – at 559,000; interesting because the 942, which was converted to the 94, actually carried fewer passengers than the 941.
I’m hoping that some of these routes will be converted to “frequent” status in the next few years – including your 856 and 861. As always, budget will be critical.
I think the Boardings Per Service Hour is a much more important statistic than overall boardings, so I would have liked to see that graph ordered and coloured instead of the overall boardings graph. Otherwise a very informative and interesting article as always.
33 bus would of been more popular if it continued to city( i miss it when it was 471 472). 33 bus should merge with the 323 to panmure to connect with the 70 bus
Nah I think it’s good broken up like that. The 323 is more a local route that meanders more out to Panorama & Carbine Rd. 32 is more across and up from Mangere to Sylvia. For longer trips you really should be getting the train.
If the trains is actually running or delayed. And its a waste of taxpayers on a ghost bus
Really makes you wonder with how many trips are on north shore buses that we cannot even prioritise 1 lane northbound in the afternoon (close curran). Presumably more people travel by the buses than 1 lane takes in single cars?
There are about five thousand bus passengers an hour heading out of town over the bridge in the afternoon. That’s worth two and a half lanes of motorway traffic.
But you know, cars.
Crazy! And was on a decrepit ONewa rd bus yesterday – no aircon, rattled like it was about to fall apart. Not sure what is going on with Ritchies but it’s like being on a school bus in the 90s.. Increased fares for that crap.
They should close Curran and Wellington St. Could even be the emergency gates that they could open up one or other if they had maintenance ongoing. Would mean 0 merging from Spaghetti Junction until Onewa Rd. Sucks for locals, but is only a few more blocks for each. Meanwhile for the network it’d be vastly more efficient. Similar reason why they put ramp signals on onramps, about speeding up the main flow.
If Curran St was bus only the onramp, the NX3 would be vastly faster, and it’d allow a bus lane from the city across the bridge.
Not surprised to see dominion rd down since covid, as they have taken away most of the direct 252/253 buses. The 25B/L takes about 15minutes longer to get to the city (but handier for those going to symonds st)