UN 238: Another look at our most productive bus routes

There is significant unmet demand for bus services.

Boost service on a existing popular (but underserviced) bus route and usage will rise in proportion. Higher frequency and longer hours induce new demand, just like the 200% rise in Saturday use of Dandenong's Route 800 bus when weekend trips were added. 

Such underservicing is common even on our most important bus routes, with headways of 30, 40 and even 60 minutes widespread, especially on weekends. I looked at this back in 2023, when 2022 data was available. 

Our most productive bus routes

Last week Mohan Wadia released similar analysis, based on data from 2023. 

This compared patronage with service hours per route to arrive at a productivity number. This averages 18.3 passengers per live bus hour across the network with wide variations. 

Though we need to be careful with patronage numbers. The Auditor-General found that, even after accounting for non touch ons, DTP bus usage data was inaccurate and actual bus ridership could be substantially higher than stated. 

As for service hours, there are two things that can make that number high. That is the time a route takes from start to finish and the number of trips per week. Both are high for the long SmartBus orbitals. 

The analysis starts by listing the 20 highest productivity/highest overall hours bus routes. Starting with the most productive first they are: 

902, 182, 703, 900, 907, 733, 170, 901, 495, 623, 506, 485, 160, 150, 497, 903, 630, 481, 406, 246. 

These can be separated into the following: 

* Orbital SmartBuses: 901, 902, 903
* Other SmartBuses: 703, 900, 907
* High productivity regular routes: 246, 406, 506, 623, 630, 733
* Wyndham routes: 150, 160, 170, 182, 495, 497
* Sunbury routes: 481, 485

I pointed out back in 2019 that high productivity, such as all these routes exhibit, is a sign of underservicing. All justify higher service but, for the longer routes, not necessarily along all of their length as segments can have widely varying productivity and/or inefficiently overlap other routes. 

I will now discuss each grouping of routes and recommend actions arising to further improve productivity and service levels.  

Orbital SmartBuses

While 903 has the highest patronage (and some sections with more intensive service), it is beaten by both 901 and 902 on productivity. This is likely due to 903 west of Northland being inefficiently overlapped by multiple other routes (eg 527, 465, 411, 232 and more) as well as having some low patronage industrial catchment in the west. These factors make it ripe for breaking up as an orbital with western and northern parts merged with other routes to provide a consistent seven day 10-15 minute frequency on most sections. 

The 901 also suffers from some overlap with 902 in the Templestowe and Greensborough areas, serves an industrial catchment in Dandenong South and has territory between South Morang and Greensborough that likely do not justify a frequent service. 

902 likely tops the productivity list due to its extremely high usage around Glen Waverley and Springvale as well as duplicating fewer routes on its way across to the north and north-west. It may however benefit from some improved directness near Greensborough and operation to Melbourne Airport instead of Airport West Shopping Centre (whose catchment is more localised). 

Usage on the orbitals has been the fastest to rebound after the pandemic with usage in 2023 within 5% of 2019's number. It may well be higher now due to the reduced attractiveness of driving due to higher fuel prices. 

Shorter term recommendation: Operate short workings to boost the busier parts of the SmartBus orbitals to every 15 min during the day on weekends. Add after 9pm Sunday service on these portions. Existing service levels on Route 907 could be adopted as a template on both these counts. 

Longer term recommendation: Split SmartBus routes into more manageable shorter segments typically operating every 10 minutes or better on their busiest portions. Consolidate with overlapping routes for best legibility, economy, productivity and evenness of service. 

Other SmartBuses

All three (703, 900 and 907) are in the top five performers for productivity. Route 907 has steadily got service increases over the years. It has long operating hours including (rare for Melbourne) 15 minute weekend frequency and after 9pm Sunday service. Its next step is possibly an improvement to daytime frequency to 10 minutes and late evenings to every 15 or 20 minutes. 

The other two far lag the 907 in service. 900 was meant to be the substitute for a promised railway to Rowville. However it does not have train-standard frequencies on weekends and evenings with 30 minute gaps typical. It also stops around 9pm Sundays. The 900 gets most of its productivity from the busy Caulfield - Chadstone - Monash area. Productivity at the Rowville end is hampered by poor quality feeder services east of Stud Road and bad walking connectivity between the shopping centre's bus interchange and the Route 900 stop. 

703 is even further behind in service with this the only pilot SmartBus that did not get upgraded to the full SmartBus service levels on weekdays during the day or in evening finishes. While it does not serve any big shopping centres it does serve Monash University which contributes greatly to its high productivity. 

Recommendation: Upgrade operating hours and weekend frequency on Routes 703 and 900 to match 907's current service level. Upgrade 907's service level to every 10 min or better during the day all week and every 15 - 20 min at all other times. 

High productivity regular routes 

These routes are the unsung heroes of the bus network. They didn't get the special branding of SmartBus but have well above average productivity. Most are on straight alignments or serve major destinations such as universities and shopping centres. With the exception of the 733 (part of which got a boost to every 15 min on weekdays) none of these routes has had significant timetable improvements in the last few years. 

406, 623 and 733 are most notable for the large shopping centres served. Their weekend loadings are particularly high indicating that their typical 30 to 60 minute service is insufficient. Despite inconvenient access to trains at Richmond (because the nearest station entrance is kept closed at non-event times) the 246 is a strong patronage performer due to its very direct main road alignment along Punt Rd/Hoddle St. It is also one of a handful of routes that runs a 10 minute frequency on weekdays with 'rapid running' or headway-based timetables. 

630 has a weak western terminus but still performs well due to Monash University at its other end. 506 is an east-west route centred on Brunswick. It is the only route out of the top 20 not to run on Sundays.

The above confirms, as did the 2021 Bus Plan, that in Melbourne there is often a mismatch between a bus route's patronage (or patronage potential) and the service levels provided. Cities such as Perth, where timetables are continuously reviewed and there is a conscious effort to form a seven day high frequency network every 15 minutes or better from high performing routes, suffer from this less. 

Shorter term recommendation: Upgrade weekend services levels on all routes to match weekday frequencies. Extend operating hours to midnight.  

Longer term recommendation: Network reform and straightening on all routes (406 more direct between Footscray and Highpoint, 506 removing Brunswick West kink, 623 potential amalgamation with part of 624 and routing via Caulfield Station, 630 a stronger western terminus, 733 southern terminus extended through Clarinda to Cheltenham instead of Oakleigh) with further service upgrades. 

Wyndham routes

These are also high productivity routes but are uniquely concentrated in a local government area that got almost its entire bus network redrawn in 2013 (Point Cook) and 2015 (Hoppers Crossing/Werribee/Wyndham Vale/Tarneit). 

Five of the six routes (150, 160, 182, 495 and 497) drop to a 40 minute service outside peak times. While there have been Wyndham area bus upgrades in the last few years they have not benefited these routes despite scope to deliver 20 minute headways (such as their routes run in peaks) with the existing bus fleet. The continued existence of productive main road routes running only every 40 minutes lends weight to those who assert that Melbourne's west has been short-changed by Daniel Andrews political strategies that prioritised swing seats in the east. 

The other Wyndham route in this list is the 170. Unlike the others this has a 7 day 20 minute service. It's a main road route between Werribee Station, Werribee Plaza and Tarneit Station. It possibly justifies a tram-like 10 minute frequency given its usage. 

Recommendation: Upgrade routes 150, 160, 182, 495 and 497 to every 15 min peak (to match proposed late 2026 Werribee line train timetable with its 7.5 min service) and 20 min off-peak seven days. Extend operating hours to midnight to match service levels recently implemented on routes such as 153, 154 and 192.  Upgrade Route 170 to run every 10 min day, 20 min night to form a tram-like corridor. 

Sunbury routes

481 and 485 are short local feeder routes in Sunbury, typically operating every 40 minutes. I'm surprised they made the top 20 most productive as Sunbury demograpics are slightly different from the high bus using mix seen in areas like Wyndham, Brimbank, Craigieburn and Greater Dandenong. 

Better version of above chart is here.

Other comments

The analysis also plots routes on a scatter graph. 

601, the Monash University - Huntingdale express shuttle, stands out as the most productive route. This is achieved by short run times, high frequency and a high volume corridor. It didn't make the list of 20 charted above as its shortness means it doesn't require many bus hours, despite its high frequency.

Other strong performers noted include the 402, 737 and 767 with the 737 (which serves major destinations in the east such as Monash University, Glen Waverley, Knox City, Boronia and Croydon) singled out as being worthy of improvement. 

Route 788 (Frankston - Portsea) is identified as an example of a low productivity route out of the top 50 (as measured by service hours). This can be attributed to several factors. First of all it is a long route - so its service hours will inevitably be long even if its frequency is not high. Parts are slow, with speed humps and significant local traffic in parts. That's important as slowness hurts productivity. 

The 788 is also overlapped by significant service on routes 781, 784 and 785 between Frankston and at least Mornington. Parts of its catchment is low density, whether inland, the part hugging the bay (fish don't tend to ride buses) or parts west of Sorrento. The route is also partly paralleled by an express bus (887). 

However the 788, as the Mornington Peninsula's main public transport spine, goes to many useful places and has a steady hop on and hop off usage. It is possible that the reformed Mornington network that started on 5 July will nudge more people to get the 788 as the Barkly St bus interchange now has fewer buses. That, and the potential to make a new connection from Hastings due to the new 886, should somewhat improve the route's productivity. 

I mentioned the 18.3 passengers/hour average productivity. It should be noted that there is what statisticians call a 'long tail'. Many routes are significantly below average. Sometimes this is unavoidable if it is agreed that coverage in a hard to reach area should be retained for equity reasons. But on the other hand there are parts of the bus network where inefficiencies such as overlapping routes or excessive service levels (in a few cases) could be tackled to improve productivity. Some of the latter is the sort of more complex bus network reform that successive governments have often found difficult. 

Summary

The pattern of high productivity bus routes remains similar to last reviewed. That is routes that serve major shopping centres, large universities and run directly between train stations on at least two lines are highly productive. 

But just because a route is already productive does not mean it can't help even more people get around. This is possible by cutting maximum waits, increasing operating hours, and, particularly on orbital SmartBuses that serve weak catchments or overlap other routes, simplifying the bus network so that corridors have a smaller number of routes running more frequently. 

There has been welcome acknowledgement by government that there is a place for routes that are consistently more frequent 7 days with longer operating hours as shown by funding in recent state budgets for service upgrades to some Wyndham buses and other key routes like 508 and 561 to run every 20 minutes all week. 

However an opportunity remains for Melbourne to (i) both expand this frequent network to more routes,  (ii) present a more consistent service offer (our main routes still vary hugely on service levels, reducing legibility), (iii) adopting a more frequent turn-up-and-go standard for our top tier routes, and (iv) publicly defining and selling the frequent network more like we used to do with SmartBus.

Sydney's All Day Frequent Network (every 10 min or better), Perth's High Frequency Bus Services (every 15 min or better) and Auckland's Frequent Transit Network (every 15 min or better) all represent best practice Monday to Sunday examples from this part of the world.

Perth's model is most notable for how it forms its frequent buses - it continuously reviews the bus network and goes out of its way to optimise resources to upgrade or consolidate regular routes until they are frequent enough all week to qualify for High Frequency status (after which they get a 900 route number). This means that even in times when there is not a lot of new funding for bus services new frequent routes can be added, as happened earlier this week when existing Route 100 was upgraded to the more consistently frequent 920.

With Melbourne's historically much slower rate of bus network reform, the opportunities to do this here are far higher here than in Perth, with political will seemingly the only barrier.  

See other Useful Network items

UN 237: A new Melbourne PT frequency map

Melbourne's public transport has its share of disruptions, wildly varying service levels and (for buses) limited operating hours and indirect routes. 

But parts of the network are pretty good, with tourists awed by the size of our train and tram systems. 

Every route is equal - how Transport Victoria maps

However, although charged with network marketing, DTP through its Transport Victoria brand has a habit of allowing the good parts of the network be lumped with the bad parts. 'Use the journey planner' is their normal refrain, with no thought given to expanding would-be users' ideas of what is possible first. 

Especially for buses, which have the highest patronage growth potential of the three main modes yet low social licence (as revealed in internal Bus Plan documents tabled in parliament), fixing these perceptions requires (i) changing reality by chopping out the bad bits through bus network reform and (ii) accentuating the positive through good information and marketing. 

Auckland has done both. Its bus network is probably now the best of any similar city in Australia/NZ and, with 15 minute 7 day frequencies on its main routes, certainly beats Melbourne's. AT also understands messaging with it defining a 'frequent promise' with main routes highlighted on maps.  

DTP/Transport Victoria presents buses more or less the same, regardless of how useful or useless they are in reality. This is exacerbated by our wide variation in bus service levels. The Knox example below displays the main road 901 (every 15 min weekdays, 30 min weekends over long hours) at equal prominence to the 757 (8 trips on weekdays, none on weekends). The former is useful for many trips, the latter for only a few.  

If you're house hunting and want to know whether a bus route is useful without having to pore through hundreds of timetables, you won't get much help from DTP/Transport Victoria. They provide basic information (equal for all bus routes) on a take it or leave it basis. Unlike a more dynamic private sector company or even a government business enterprise they appear not to have much of a stake in whether people use buses or not.  

Frequent routes are more useful - some independent maps 

Fortunately there are independently-produced frequency maps for Melbourne's public transport that start where TV's local maps finish. 

There's my interactive frequent network maps here (that has just had an update with Bus 140A added). 

But today's topic is to introduce the amazing new Melbourne frequency map developed by Adam Bain. Available at ptmapmelb.com it is based on midday weekday frequencies. Switchable layers exist for 10, 15, 20 and 40-60 minute frequencies. It is inspired by similar maps for Seattle and Miami.

Just like with Melbourne, Seattle had a disconnect between what was important for passengers and information that was published by the siloed transit bureaucracy. Creating a gap filled by independent map makers and activists.  

You really need to spend some time with ptmapmelb.com. Achieving something that a hundred timetable lookups or journey planner searches will never do, it will change how you view Melbourne's PT network. You can zoom in for more detail but at no point does it become overwhelming. 

To get an idea of high and low service areas I suggest first ticking the 10 minute box and looking at the network. You'll see many (but not all) tram corridors and a few train lines. It's overwhelmingly a radial  inner suburban network with only a handful of cross-radial routes. You'll need to look carefully to see any buses. 

Ticking 15 minutes adds the rest of the tram network, some extra Metro lines in the east and quite a few buses, with the SmartBus orbitals being most prominent. But remember this is a weekday map with services on the latter (especially) collapsing on weekends. 

A substantial growth, particularly in the west and north comes when 20 minutes is selected. Apart from the outer-east this substantially completes the Metro train network, Melton and Geelong V/Line and adds many bus routes. But you will still see large populated areas without coverage. Most prominent gaps include Point Cook, Melton, large parts of Brimbank, Wollert, Epping, Thomastown, Mernda, the entire outer east, Greater Dandenong, Pakenham and Frankston. 

Adding 30 minutes brings up the Belgrave and Lilydale Metro lines as well as large parts of the bus network in the north and east. This include a lot of odd frequencies like 22 and 25 minutes around Epping and Reservoir where connections with trains are particularly poor. 

The much finer drawn 40-60 minute range includes most of the rest of the bus network including the areas listed as missing out on 20 minute service. In the west, north and outer south-east this may be due to an unwillingness to schedule buses every 30 min as they do not meet trains typically running every 20 minutes. If there is not the interest to run buses every 20 minutes then this confines large areas with (at best) a 40 minute service. As well as being good for its own sake, moving to a Metro rail network based on 10 minute core frequencies (as per the NDP) allows a wider range of connecting bus frequencies. 

The final two boxes are for limited service or FlexiRide routes. These routes are common in parts of the outer east (eg around Knox) eg 681, 682, 757, 758 and more. Although developed around 30 or more years ago these areas never got a full bus network then and still don't have one today. If an area has a lot of complex peak, limited service or FlexiRide routes it is likely to be crying out for bus network reform, as in the Reservoir and Knox examples below.  

The frequency finder

I said before that Adam's map was based on midday weekday frequencies. That's a problem because weekend frequencies in Melbourne can vary between 0.25 and 1.5 times weekday interpeak frequencies with Sunday service sometimes a fraction of Saturdays. 

Frequencies may also vary across the day between peak, interpeak and night time bands. Buses (especially) lack consistent operating hours so services may not run when you need to travel. 

However help is at hand through the companion Frequency Finder for trams and buses. 

Frequency Finder lets you find frequency by route by day of week and time of day. There's also a handy span guide and a map that changes colour (to indicate frequency) by line selected.

I deliberately selected a complex route (824) to test whether the graphic would show the more frequent Moorabbin part of the route (every 20 min) differently to the Keysborough end (every 40 min). It did.

However the frequency finder is limited to a single route and not a corridor. Thus overlaps like 250/251, 411/412 or 811/812 will not be shown as higher frequency here unlike they are on the map. 

Conclusion

This map and the accompanying frequency finder will encourage people to see the Melbourne public transport network differently with service have and have not areas really made apparent. This will make it a useful planning and advocacy tool. 

See other Useful Network items

TT 233: Route 140A joins the weekday 'Useful Network'

 

A quick welcome to new bus route 140A (Rockbank to Mt Atkinson) which joined the weekday Useful Network frequency map yesterday. Described by Transport Victoria here, hopefully they will add it to their local area map soon.

The 140A brings Mt Atkinson from having zero public transport to having better than average weekday frequency and operating hours for a Melbourne outer suburb with buses every 20 minutes across most of the day.

On weekends the 140A enjoys wide operating hours that also exceeds what most other new estates get. But at every 40 minutes (which matches the Melton line's limited weekend train service) it doesn't quality for inclusion on my Useful Network frequency maps for Saturday and Sunday.  

The commencement of Route 140A follows a long community campaign to bring public transport to an area that had much promised but nothing delivered in the way of shops or services (even a Westfield shopping centre was promised but they pulled out).

Providing buses was made harder by the area having an incomplete road network. Also some prospective termini (like Rockbank and Caroline Springs stations) are little more than train stations in the middle of nowhere. Until local facilities develop even simple errands that are elsewhere possible on a single bus trip need (as a minimum) at least one and possibly two changes between services, none of which are consistently frequent. I speculated on several bus route options for Mt Atkinson here. 

The 140 as designed gets Mt Atkinson people to Rockbank Station and, when extended, Tarneit Station. This will make it the first bus route that connects the Melton and Geelong lines. 

The long-awaited bus will make a big difference to Mt Atkinson's liveability. Though adding a new $2-5m bus route should be simpler and quicker than a major $200m capital works projects like a level crossing removals, institutional arrangements, funding means and plain political will often makes achieving the latter faster. 

But in Mt Atkinson's case community advocacy led to political urgency and the makeshift interim Route 140A being established. The 'A' standing for alternative route until the full 140 Rockbank - Mt Atkinson - Tarneit route can start later in the year. 

Mt Atkinson residents often refer to their suburb as an 'island' due to the lack of local retail, services and transport. They now have the 140A bus so they can leave their 'island'. But only to a point. The 140A bus itself still appears as an 'island' on the weekday Useful Network map, with no qualifying lines or routes intersecting. This is because while the Melton line is mostly every 20 minutes on weekdays it still has some much longer gaps that stop it qualifying. Hopefully this, along with a 20 min weekend frequency, can be sorted well before the long-promised Melton line electrification happens.  

As for the every 20 minutes Useful Network, what's next? At one time buses every 20 min in growth areas were almost unheard of. Now it seems to be quite normal planning practice, at least on weekdays. Funded 2025 and 2026 budgeted upgrades for trains and buses should give cause to be adding more lines to the Useful Network maps in due course. 

See other Timetable Tuesday items here