UN 240: The Suburban Rail Loop's Bus Problem

Executive summary: The state government risks setting the Suburban Rail Loop up to fail if it takes advice from a planning committee that has a poor understanding of the bus services needed to effectively get people to its stations. It should instead heed those with an understanding of multimodal public transport planning to be assured of maximising the patronage and mode shift benefits of the SRL.     

An expert committee set up to advise the Planning Minister on the Suburban Rail Loop and its precincts has published a series of reports on certain topics of importance to the role and success of the project.

SRL East Precincts Standing Advisory Committee 

Planning Minister Sonya Kilkenny referred several matters to the Suburban Rail Loop East Precincts Standing Advisory Committee (SRLEPSAC) in July 2025. 

One of those matters is mode shift. Which is effectively SRL's ability or otherwise to transform transport from being driving-dominated to a pattern that is more multi-modal with higher shares for public and active transport. 

The Committee includes one Lead Chair, nine (!) Deputy Chairs and fourteen other members. They are heavily drawn from urban planning and legal backgrounds with a few engineers thrown in. Many have worked for local government or as private consultants. However public transport operations or planning backgrounds appear rarer, something that may be influential as you'll see later.  

It has published nine reports so far. This includes overarching reports for (a) Summary and approach, (b) Common issues and (c) General matters plus specific reports for each of the six proposed SRL stations. 

These reports give an idea on official thinking and advice to the Planning Minister. So they need to be scrutinised for rigour and appropriateness, especially as their subject matter includes transport planning.

I'll review the General Matters report later after setting the scene. 

Transport and planning policy development

While by name a transport project the state government has stressed that the Suburban Rail Loop is as much if not more about planning and housing. 

Possibly because bringing in the planning dimension allows potential land use changes and value uplifts that (some hope) could part fund SRL through value capture. The housing element taps into current concerns about a tight and expensive rental market. And reframing SRL as a broader project subverts critics who regards it as poor value from a purely transport perspective. 

The recent megaproject era also coincides with some change in policy thinking. This is exemplified by the merging of the Planning and Transport departments (as DTP) and the preference for transport to be planned in a wider context. 

Hence specific rail, tram and bus plans have withered. They have struggled to attract substantial funding despite tens of billions going to stand-alone 'Big Build' infrastructure projects. There is also no overarching transport plan, something the Auditor-General noted as a Transport Integration Act requirement in 2021. Instead some transport themes (such as the publication of a map of strategic bus corridors) will be picked up in the broader Plan for Victoria.

If this approach drives DTP organisational structures, there is a chance that the staff resourcing for individual modal planning relative to broader precincts planning might change as the department faces pressure to deliver Silver Review cost savings.

If poorly done this poses risks including a loss of internal transport planning expertise and capacity to implement the greenfield rail timetables and bus network reform still needed to realise Big Build benefits. A DTP that is big but weak also risks it having insights that do not sufficiently flow to major projects such as the SRL.  

The Transport Integration Act 2010 is relevant here. Points that stand out include efficiency, seamless travel within and between modes, integrated decision making and the user experience. 

Inside the General Issues Report

Of most interest is SRLEPSAC's General Issues Report. Mode shift gets about 20 pages. 

I was particularly interested in what it said about access to stations (important if they are to get much use since there are only six on a 26km long line) and the role of various transport modes to enable this.

Bus services already good - no improvements needed?

The first mention of buses comes on page 15 of the General Issues report.   

In principle they like the early delivery of bus transport and infrastructure improvements. 

BUT

Without evidence the Committee asserts that "many of the precincts are already particularly well served by well-established bus networks".

Essentially they are saying is that buses around the SRL stations already pretty good. And that it's not worth investing more in them to deliver mode shift. So it was preferable to direct resources to active transport instead.  

There is no doubt active transport is important, should be improved and that its returns per dollar spent are high (likely more than the SRL itself). The same goes for bus reform. You'll get some benefits from an SRL without bus upgrades but they are limited. Add bus reform and the benefits multiply.

If you were given a choice of three projects (a) SRL only, (b) SRL plus bus reform or (c) bus reform only, no SRL then the case for the last two could be argued but there are no circumstances that an SRL only would represent good value. Essentially upgraded buses is a necessary condition for SRL to succeed with analysis by Mohan Wadia available here.

SRLEPSAC's view is basically the opposite. They don't see buses as an important part of the SRL project. Their comments about the areas around the SRL stations being "particularly well served" by buses are only tenable if they looked at a bus network map, counted the routes and were content at that. As opposed to examining matters such as route directness, operating hours, frequency and legibility to get a fairer assessment. 

If the learned planners who comprise SRLEPSAC did not deign to examine timetables for their widespread hour gaps or (horrors!) try riding some buses, they could have checked recent analysis from reputable sources for an informed view on the bus network.  

Indeed Victorian government agencies outside the SRL silo were continually churning out material indicating the need for improved bus services from 2021. Examples include:  

* 2021: Victoria's Bus Plan
* 2022 Announcement of major bus network reform (which in this form later got shelved)
* 2023 Infrastructure Victoria report on buses 
* 2024 Detailed bus network analysis by DTP and consultants with this tabled in Parliament 
* 2026 VAGO performance audit on bus planning

All revealed serious problems but also great opportunities for a revived bus network, many at low cost.  This is directly relevant to the SRL East because, with only 6 stations along its 26km length, many will need to rely on feeder transport such as buses to reach stations (remembering there will be no extra car parking and many live beyond both rail and convenient active transport reach of the SRL stations).

Potential for bus network patronage growth (which addresses both SRL station access and the mode shift question) was further expressed in a speech by then department Secretary Paul Younis who suggested that reformed buses across Melbourne had potential to carry as many people as trams. 

Real world evidence of such growth exists too. Route 800 Saturday usage rose 200% when its previously limited timetable was upgraded. And routes such as 733 and 767 (both of which serve SRL stations) had strong ridership increases when their services were modestly increased to 15-20 min on weekdays and 30-40 minutes on weekends in 2023. 

Additional rail (such as SRL) can increase rather than reduce bus trip making due to increased connection opportunities. This is demonstrated in cities such as London where, despite large Tube and Overground systems, bus ridership exceeds either individually. Versus the Melbourne case where, despite their much larger catchment area, bus ridership is lower than either trams or Metro trains. 

That the SRLEPSAC appeared so uninformed about the bus network and its potential for mode shift either indicates a lack of rigour leading to potentially poor advice to the minister. The result could be decisions (eg underprovisions for fast direct bus routes and poor interchanges) that result in the SRL failing to deliver its potential network benefits as part of the connected transport system envisaged in the Transport Integration Act. 

Bus infrastructure

The committee report considers bus infrastructure.

This could include matters such as bus stops being located as near as possible to SRL (and other) stations, sufficient capacity for expected service levels (even if not all provided immediately), convenient stop locations in the SRL area (noting that catchments can be 1.6km from stations) and, critically, provision for efficient through routes in multiple directions with strong bus priority.

The latter is extremely important and often overlooked; interchanges that require a bus to backtrack (eg Glen Waverley currently for key routes) waste thousands of hours that is better spent speeding travel and increasing service frequency. 

There is only one good answer to the last point about early delivery of bus infrastructure - yes! Because if it is not delivered first-up it never gets done. Which puts a permanent choke on the capacity and catchment of SRL stations and thus the SRL's usage and mode shift potential. Those wishing to take advantage of the SRL's activity centres to live a convenient low-car life are also affected if buses to surrounding local destinations are substandard.  

Bus infrastructure requires some idea of a future bus network to define some high capacity corridors. This part under 8.6.3 identifies work done. 

So far so good. Until we get to this, which immediately follows:

Mr Gnanakone is a professional traffic engineer and transport planner. Earlier he said that it was critical that active transport improvements be delivered early to achieve mode share targets. He was also supportive of early bus network improvements. However he warned that the timing for most bus improvement works was medium to long term, and even after SRL East was operational.

This should not be regarded as an inevitability; a benefit of buses is their ability to respond quickly and scale up to changing travel needs so it is purely a political choice that bus service upgrades and network reform have historically been so slow with multi-decade lags in some cases.  

Unfortunately the Committee does not take Mr Gnanakone's advice regarding the need for early delivery of bus upgrades on board. They are pretty soft, saying it is merely desirable rather than an essential of equal if not greater importance than the SRL itself. 

As noted before SRLEPSAC (wrongly in my view) thinks the existing bus network is fine and overlooks the need to make provision for the sort of fast and frequent bus networks SRL stations will need. 

I agree with Mr Gnanakone about timing of bus improvements. Buses should be upgraded as precursors to the SRL, not as something that *might* be done afterwards. Only this sequencing can bring forward SRL benefits early and maximise mode shift. 

Conclusion

It is concerning that the Suburban Rail Loop East has a committee of planning experts, intended to advise the minister on matters such as transport mode shift, that apparently: 

* Thinks that existing bus services are good enough ('particularly well-served' in their words)
* Has not considered others' analysis and policy eg DTP's Bus Plan, IV or VAGO reports
* Does not see buses as an essential means to get people to SRL stations or a contributor to mode shift
* Has rejected advice from transport professionals on the importance of bus access 

If this committee's advice on buses is followed then the SRL risks being a stand-alone project poorly connected to feeder transport that does not deliver anywhere near the patronage or mode shift possible if a more connected multimodal approach was taken. 

The Committee may well be good at town planning but for buses (and other SRL feeder modes) both the Planning and Transport ministers might do well to seek advice from elsewhere. 

As a reminder their General (and other) papers can be read here. 

See other Useful Network items

TT 235: Bus frequency of frequencies and the weekend cliff

Last week I graphed the frequency of Melbourne bus routes at midday weekdays, aided by data from Adam Bain's Frequency Finder. 

Matching the common reputation of Melbourne buses, this graph showed that just 7 out of over 350 routes ran every 10 minutes or better while a further 18 ran every 15 minutes. More ran every 20 minutes but the vast majority were still every 30 or 40 minutes. 

However that data didn't tell us whether weekday peak service is much better than off-peak. Or how weekends fare. 

Fortunately Adam has since added peak and weekend charts to his item on bus route numbers. So I'm back to look at what these tell us about the frequency of bus frequencies at these times. 

Weekends (every 40 minutes is king)

Below shows number of bus routes in each frequency band at midday. It's like the graph last week but I've added Saturdays and Sundays plus a 'No service' grouping. 

You will notice that while the 31 - 60 minute grouping was tallest on weekdays there was still a substantial number of routes in the 21 - 30 and 16 - 20 minute frequency bands. 

That is not so on weekends. On Saturday and (especially) Sunday most frequent and semi-frequent weekday bus routes become infrequent. That introduces an inconsistency of service that is not present with trams and trains.  

For example on Saturdays the 31 - 60 minute column is five times taller than the next. This overwhelmingly comprises routes either every 40 or 60 minutes. Even semi-frequent service is rare on Sundays, with barely 10% of the total (or 38 routes) operating every 30 minutes or better versus nearly double that (78) not operating at all. For some reason the 901 and 902 orbitals (nominally every 30 min) made the 31 - 60 minute grouping on Sunday, possibly due to a slightly uneven timetable. 

Personally I'd have split the 31 to 60 minute groupings so that 40 and 60 minute frequencies were in different ranges. If this was done you would find that 40 minute column is tallest, ie it becomes the mode (ie most common) off-peak frequency for Melbourne's buses. This would likely apply on both weekdays and weekends though there are routes that are every 40 min on weekdays but 60 min on weekends, especially Sundays. 

Be aware that this generalised data also doesn't tell us the very large differences between routes. On some service levels collapse between weekdays and weekends, others are frequent Monday to Saturday, collapsing on Sundays while a third group run the same frequency all week.

Examples of routes (that mostly but not always have legacy timetables) with sharp weekend frequency falls include 410, 472, 828 and even the new 241. All these drop to one-third or one-quarter weekday frequencies on weekends, particularly on Sundays. 

Drops to half frequency are common on most SmartBuses that drop from every 15 to every 30 minutes on weekends. Similar applies to new or upgraded growth area routes in Craigieburn, Diggers Rest, Melton, Mt Atkinson, Cairnlea and Clyde where every 20 minutes weekday, 40 minutes weekend is the standard pattern.  

Conversely some of the newer or recently upgraded Wyndham area routes (eg 153, 154 and 182) have their flat 20 minute frequency operating 7 days to late with no weekend service 'cliff'. Similar (but at lower service level) applies in areas such as Pakenham and Mornington where a flat hourly service all week is the norm as can be seen from Adam's new interactive interchange and service map (with context here). 

Peaks (a little better than interpeak)

If midday weekends get far fewer buses than midday weekdays, what happens during the weekday peaks? This is the most challenging time to run buses, especially in the morning where school and commuter peaks coincide. Traffic volumes also make for longer or more variable run times, making it harder to schedule efficient yet reliable timetables. The graph below is the previous one with peak columns added for each frequency range. 

The range of frequencies operated in peak is flatter than midday weekdays and (especially) weekends. Most notably there is a substantial number of routes operating in the three 16 - 60 minute frequency ranges with 30 minutes being a fairly average frequency. However this is not quite 'turn up and go' with  only a lucky few enjoying 10 or 15 minute peak frequencies.

Many bus routes in established areas had higher peak frequencies prior to large service cuts in the early 1990s. Many have gained (or had restored) evening and weekend service but not all peak service has yet been restored. This goes some way to explaining why about 100 bus routes are still infrequent in the peaks.

Like with weekends there are variations. A welcome new trend on some new growth area routes (eg 154, 405, 451 and 475) is to run a 20 minute frequency peak and off-peak. In contrast some earlier routes (eg 150, 160, 182, 494, 495, 497 etc) have a ~20 min peak service but only a 40 minute off-peak service. The fact that these rate amongst the most productive bus routes in Melbourne make a 20 minute off-peak frequency a better fit. 

A few routes are listed as having no service in both weekday peaks and midday periods. The former are shopper type routes that don't run peaks while the latter are commuter peak only routes. 

A table showing the numbers in the above graph is below. 

Trams (the mode is the service)

Trams have a distinctive operating culture including operating over long hours and at generally high frequencies all week in a manner that buses do not. This is made graphic below (Sunday shown as a 'worst case' - better views of this and other days here). 

The diagram shows that all trams run every 10-15 min (or better) during the day all week (with lower Sunday morning frequencies). No buses have 10 minute maximum waits (white bar). And just 4 (out of over 350) have 15 minute maximum midday waits across the week. These are buses 234, 246, 905 and 907. 

The graphic doesn't analyse evening frequencies but if it did trams would look even better. Whereas trams have a maximum evening wait of 20 minute Monday - Saturdays, the figure for the above four routes is 30 minutes with 234 and 246 finishing much earlier on Sunday than trams. 

Buses 234, 246, 905 and 907 are the only ones that have somewhat tram-like frequencies on all days of the week. All are east or south of Melbourne CBD, none north or west (apart from part of 246). The one bus that offers a tram equal service is the small segment of the 732 between Vermont South and Knox City that was instituted to partly meet a 1999 political promise to extend the 75 tram to Knox City.  

This consistently high service is a major contributing factor to why the tram network, despite it serving only a small part of Melbourne, carries tens of millions more people per year than buses, which serve  millions more homes. With wide hours, decent daytime frequencies and a rarely broken 20 minute maximum wait*, trams have a sort of dependability built into the timetable that Melbourne buses lack. This regular service induces more routine usage for a wide variety of trips that buses could have but don't due to their often more limited hours and frequency. 

The tram/bus inequality demonstrates the significant gaps between inner/tram and middle-outer/bus suburbs in their public transport provision. This includes busy middle-suburban destinations (eg Box Hill/Chadstone/Monash) and denser outer areas (eg Tarneit and Craigieburn) that strongly justify frequent tram-like bus service all week compared to their existing back-street type service levels, particularly on weekends. 

 (*) Sunday mornings and evenings excepted. 

Trains (mostly every 20 minutes BUT)

Adam's analysis did not assess trains but there's not many lines so I made a frequency of frequency table for completeness (excluding peaks due to complex stopping patterns on some lines). 

There is overwhelmingly a concentration of service around every 20 minutes midday. Unlike buses, there is no pronounced weekend service fall-off on most lines. Indeed Belgrave and Lilydale enjoy better midday weekend service (every 20 min) than on weekdays (every 30 min). 

Why are 21 lines shown when our metropolitan rail network has only 15? Firstly I've included the two heavily outer suburban Geelong and Melton lines. Secondly I've doubled up on lines like Sunbury, Hurstbridge, Belgrave/Lilydale and Pakenham/Cranbourne that have outer suburban branches or frequency fall offs to convey a fairer impression of frequency much of their catchments have.

The data behind the graph is below. 

Metropolitan train stations mostly (but not always) have a 'safety net' 20 maximum wait, at least during the day. The major exceptions are Belgrave and Lilydale on weekdays and Melton on weekends with 30 and 40 minute gaps respectively. The quieter outer part of the Hurstbridge line also has a lower service.  

The biggest gap between trams and trains are the latter's 30 (and sometimes 40) minute intervals at nights and on Sunday mornings. 2025 and 2026 state budget measures will go part way to closing these gaps to 20 minutes on the Craigieburn and Upfield lines (all times), Mernda and inner Hurstbridge (weekends) and Belgrave, Lilydale, Alamein and Glen Waverley (weeknights). There has so far been no commitment to closing the 30 minute weekday gaps for Belgrave and Lilydale and the 40 minute weekend gaps for Melton. 

Summary

This data says several things about buses in Melbourne. The typical route runs every 30 minutes in peak and about every 40 minutes off-peak. Many Melbourne bus timetables have a weekend cliff where services are much less on weekends than weekdays with a significant minority still not running seven days.

In addition there is not a large defined group of frequent all week routes with similar service levels to trams. Instead only a handful run frequently seven days with the most common pattern being frequent weekdays and infrequent weekends, particularly Sundays.

Melbourne's lack of a premium bus network of maybe 50 main road bus routes operating at tram-like frequencies seven days (as proposed in the Bus Reform implementation approach that the Auditor-General said was unachievable at the current rate of progress) remains (along with low all day frequencies on certain train lines) the biggest block on patronage growth. 

The steep service drop-offs show that lack of bus vehicles is not the problem; rather it is the lack of a will to work them harder for more of the week due to recruiting too few drivers. And in some areas there is also significant scope to make better use of existing bus and driver hours by simplifying routes and removing inefficient overlaps.

Attention to these service matters would transform public transport's usefulness in ways that the 'Big Build' so far has not in most suburbs.  

See other Timetable Tuesday items here

SPECIAL: When is your bus going contactless?

On Thursday the state government announced the date of contactless payment 'tap and go' roll-out on Melbourne buses. That date is Sunday 26 July 2026 for everything with some starting tomorrow on the 19th. 

This is a big deal because tap and go makes payment for bus travel easy for any full fare payer with a credit or debit card. Trains and trams have already been contactless for some weeks prior. 

Contactless will remove a significant barrier for spontaneous trips. Right now there is a need to find a Myki outlet, buy a card ($6 before any travel), add some value and then keep it topped up. 

You forfeit the card cost if not making any more trips. And claiming refunds on unused balances is a hassle. Even worse was that buses stopped doing cash top-ups during the pandemic and never resumed them after. Myki's clunkiness (along with relatively high fares for some short trips and insanely cheap V/Line fares) even undermined the social licence of charging and enforcing fares in some peoples' minds. 

Admittedly Android users could have 'Mobile Myki'. But some people prefer to keep their finances and phones separate for security and budgeting reasons so that was never a complete solution. Not to mention it being unavailable for iPhone.  

Summing up, Myki worked well for regular rail commuters but was inconvenient for spontaneous bus and tram travel. Myki arguably depressed both the number and character of bus users. Honest people who were willing to pay but found this inconvenient avoided buses while its inconvenience hardened some into habitual fare evaders who faced little resistance with lax enforcement.

The seeds of this were sown by an early decision in the Myki specification to have expensive to produce cardboard smartcards rather than cheap paper for disposable tickets. After their trial in Geelong (and elsewhere) the former was not proceeded with when the previous government descoped Myki. 

Contactless Tap and Go promises to fix that by restoring an easy payment option across all modes that we haven't really had since cash fares. Trains were first, then trams, and now it's the turn for buses with a phased roll-out. When it's done you'll be able to travel without buying and keeping topped up a Myki card.

An example benefit is being able to travel from Melbourne Airport to the city for under $3, as I discussed with ABC's Charlie Pickering on Thursday 16 July (approx 6 min in). 

Concession passengers should continue to use Myki to get the cheapest fares. Contactless roll-out for concession passengers is complicated as the Auditor-General reported. Though the temporary halving of fares means that general concession payers using contactless to pay full fare are paying the same they did earlier in the year.     

A staged roll-out

The media release places greatest prominence in the 26 July date as this is the one when the roll-out will be complete across all buses. However it lists certain areas that will get Tap and Go a week earlier, ie 19 July.

This advice is safe and simple for the average person. But transport geeks and early adopters will want to know route number level of detail. Especially as it's a bit more nuanced with some services going live sooner. 

You will find this information in a special edition of the Victoria Government Gazette. More precisely No. S 409 Thursday 16 July 2026. From this we learn that roll-out will be by four dates: 19, 21, 24 and 26 July, with this largely ordered by operator.

I've listed the roll-out below with route numbers given for services in the metropolitan series. 

19 July 2026 implementation

Bus routes in Bacchus Marsh, Ballarat, Bendigo, Geelong, Kilmore, Latrobe Valley, Seymour, Warragul. 

Metropolitan numbered routes: 433, 434, 452. 475, 481, 483, 485, 486, 487, 488, 489, 683, 684, 685, 789, 790, 791, 792, 795, 796, 798, 799, 881, 890, 891, 892, 893, 894, 895, 897, 898 and 965. 

The above are run by operators including Christians Bus, CDC, Cranbourne Transit, McHarrys, McKenzies Tourist Services, Sunbury Coaches and more. 

21 July 2026 implementation

150, 151, 152, 153, 154, 160, 161, 166, 167, 170, 180, 181, 182, 190, 191, 192, 194, 241, 301, 343, 356, 357, 358, 381, 382, 383, 384, 385, 386, 387, 388, 389, 390, 400, 402, 404, 406, 407, 408, 409, 410, 411, 412, 414, 415, 417, 418, 419, 420, 421, 422, 423, 424, 425, 427, 428, 431, 432, 439, 441, 443, 444, 453, 454, 455, 456, 457, 458, 459, 460, 461, 462, 463, 465, 467, 468, 469, 471, 472, 476, 477, 478, 479, 482, 484, 490, 494, 495, 496, 497, 498, 501, 511, 513, 514, 517, 518, 524, 525, 528, 529, 530, 531, 532, 533, 534, 536, 537, 538, 540, 541, 542, 543, 544, 554, 555, 556, 557, 559, 561, 564, 566, 569, 570, 577, 578, 579, 580, 582, 601, 605, 606, 612, 623, 624, 625, 626, 630, 941, 943, 947, 949 and Flexiride Rowville, Lilydale, Croydon, Mooroolbark and Melton South.

The above are run by operators including CDC, Dysons, Transit Systems and Ventura. 

24 July 2026 implementation

200, 207, 215, 216, 220, 223, 232, 234, 235, 236, 237, 246, 250, 251, 270, 271, 273, 279, 280, 281, 282, 284, 285, 293, 295, 302, 303, 304, 305, 309, 318, 350, 364, 370, 426, 429, 503, 504, 505, 506, 508, 509, 510, 512, 546, 552, 553, 558, 567, 600, 603, 604, 609, 668, 669, 901, 902, 903, 905, 906, 907, 908, 922 and 923.

The above are run by Kinetic. 

26 July 2026 implementation

201, 526, 527, 548, 549, 550, 551, 627, 631, 663, 664, 670, 671, 672, 675, 677, 679, 680, 681, 682, 688, 689, 690, 691, 693, 694, 695, 695F, 696, 697, 699, 701, 703, 704, 705, 706, 708, 709, 732, 733, 734, 735, 736, 737, 738, 740, 742, 745, 753, 754, 755, 757, 758, 760, 765, 766, 767, 770, 771, 772, 773, 774, 775, 776, 777, 778, 779, 780, 781, 782, 783, 784, 785, 786, 787, 788, 800, 802, 804, 811, 812, 813, 814, 816, 822, 823, 824, 825, 828, 831, 832, 833, 834, 835, 836, 837, 838, 839, 840, 841, 842, 843, 844, 845, 846, 847, 848, 850, 857, 858, 861, 862, 863, 885, 886, 887, 888, 889, 899, 900, 925, 926, 927, 928, 929, 951, 953, 959, 967, 978, 979, 981 and 982.

The above are run by Ventura (partly in 900's case). 

That's the sequence of contactless tap and go roll-out according to Victoria Government Gazette. The big date, as announced, is Sunday July 26th but roll-out on many routes has been gazetted for the days prior. Happy tapping! 

UN 239: Assessing transport projects by interchange point: SRL vs Airport Rail vs Frequency uplifts

Airport rail vs Melton electrification

I've never warmed to the version of airport rail as proposed in Melbourne. 

Sure it connects a major destination (that most people might go to a few times of the year) but it  transforms neither the network nor the city. 

With the likelihood of a premium fare nor it would not have much helped airport workers (most of whom live in surrounding suburbs anyway). Their needs are simpler - decent buses to surrounding areas that run at least semi-frequently over a wide span of hours.  

Before Ben Carroll MP got a Keilor East station added to the plans, Airport Rail, like what got built in Brisbane but unlike Sydney or Perth versions, threatened to be a stub with no wider network benefits. 

Keilor East station would rescue Airport Rail from complete uselessness by providing Metro rail connectivity in a large area without it. There is an additional cost but Airport Rail's value is magnified. But only by a little bit; $12 billion for one (1) extra suburban access point for Metro rail is in no one's world good value.

To bolster its sorry case, the Airport Rail project must put on a marauder's hat and poach benefits that other projects (ill-advisedly for them) have not exploited. 

Such benefits remain on the table due to us not electrifying Melton as early as we could have (although who knows, maybe it will still beat airport rail), not increasing all day suburban frequencies on lines such as Craigieburn to aid connectivity and not exercising the option to run Metro Tunnel every 5 minutes off-peak and thus enable CBD tram reform. 

If we were to do the last with those extra paths going to Airport Rail then most of the benefits stop right there. Whereas if they go to something like an electrified Melton line then the wins ripple much further from their origin. For example the prospect of new suburban stations, relief for the Ballarat line and more balanced loads for trains entering Sunshine.  

Because almost everyone sees themselves as using Airport Rail occasionally (while having some vague faith that it brings business to the state) this, like level crossing removals, has topped public polls on the popularity of transport projects (though to be fair major bus network reform has also rated well).

Melton electrification has historically suffered by its benefits being geographically concentrated. Few from Mornington or Montrose are likely to use Melton rail much. Whereas they might occasionally use Airport Rail. So despite Melton rail upgrades helping thousands of people every day it has made front pages far less than the showy but ultimately less beneficial airport rail. 

However the story is not over. It's a thrilling race. Who gets the spoils of all that unlocked capacity in the Metro Tunnel first? Most would have predicted Airport Rail. But interest has ebbed and flowed. 

Also the political environment has changed. There are no safe seats anymore. The Werribee by-election shocked Labor. And more recently the rise of One Nation in the polls threatens to split the non-Labor (and some of the Labor) vote in regional areas and some outer suburbs. Areas like Melton are in the box seat for some of these forces as a long-term government faces an election in barely four months.

I am often asked my opinion on various transport topics. The two hottest in the last six months have been 'free' public transport and airport rail. My answer to both are the same - both carry massive opportunity costs and there are far better uses for the money (which could fund an amazing network all across Melbourne). 

I can't help thinking that some of the gloss has gone off airport rail. People realise it won't actually save very many much time relative to its cost and greater needs elsewhere. Add population growth and changing politics, I think Melton should get first dibs on funding and train paths.  

Suburban Rail Loop - cheaper per interchange point than Airport Rail?

If $12 billion for what is effectively one new suburban interchange point is poor value where does that leave the Suburban Rail Loop and other initiatives such as rail frequency upgrades and bus reform?

Don't think of the Suburban Rail Loop East as an underground orbital railway for a moment. Instead think of it as six new interchange points. As far as interchange points go they are pretty good; all six are key destinations in their own right, four intersect with Metro lines, two with trams and one with V/Line. 

A crude way of comparing SRL with airport rail is the cost per interchange point created. And by interchange I'm assuming one where there's access to 7 day frequent service for a standard fare. 

Melbourne Airport Rail is like $12b per interchange added whereas SRL is $34.5/6 ie $6b per interchange added. SRL ends up being half the cost per interchange created.

$6b per interchange is still a huge amount of money and it is reasonable to debate the SRL's value, timing and opportunity costs. This is especially when Melbourne must create dozens if not hundreds of interchange points in the suburbs to make PT useful, a feat that is unaffordable at $6b a pop. 

Scalability matters. If we can do more for less then we can have more, with bigger benefits. That's a major need in Melbourne since our network has so few interchange points. And remember adding route coverage linearly increases catchment whereas creating interchange points multiplies travel opportunities so the benefits of the latter are disproportionate. 

Especially given that Melbourne's rail network is so steadfastly radial, an SRL at $6b per interchange point that fundamentally transforms the network is better value than Airport Rail (at $12b per interchange point) which (at best) does not change the network structure and (at worst) takes core area train paths from potentially busier lines such as an electrified Melton.

One can fairly be concerned about SRL in its current form but the also expensive but even lower benefit airport rail (not to mention various large road projects) should also not escape scrutiny.

In my view there is more substance in the view that both Airport Rail and SRL are low priority than the stance, expressed by some SRL opponents, that Airport Rail should be our biggest public transport priority, when in reality it contributes very little benefit to peoples' everyday travel relative to its cost.

Whereas potential exists for a well-designed SRL to have substantial benefits but these would be even greater if a substantial package of other measures (involving a tiny percentage of SRL's cost) were done in the decade leading up to its opening.  

A 5am formula - can it work for SRL?

It's hard to compare project A versus B with a single formula but it might be easier to compare variants of the same project. 

As I mentioned before SRL could be conceived of as six linked interchange points costing $6b each to build. The connections between them will be faster than existing driving or public transport options. This speed has value in extending the range of destinations available within a given time. 

Also of value is the number of interchange points. More interchange points means more origin and destinations possible. And, like how the number of handshakes in a room increases disproportionately with the number of people, the number of potential connections rises in a triangular fashion. But only to a point; have too many interchange points and they eat into each other's catchment and slow travel speed. 

Below was my first attempt to make a formula to reflect this. 

Merit is just a number, not a unit. The higher the better. 

The formula effectively multiples the overall  improvement in speed (which would be end-to-end for a project like the SRL) by the number of interchanges possible for a line with a given number of interchange points (p).  

As an example a new line with three stations (p = 3) has an option of travelling: a. first to middle station, b. first to last station, c. middle to last station. Thus, like handshakes, there are three combinations. But unlike handshakes you can do reverse trips, which I've counted, giving a total of 6. That makes the formula simpler as there is no longer any dividing by 2 as with handshakes. 

Below is a simpler version of the same formula with the speed ratio changed to a time ratio to put it in the denominator. 

The Suburban Rail Loop has 6 stations so that's our value of p. 

The travel time is 22 minutes end to end versus 55 minutes currently, giving a ratio of 0.4. 
The merit number is 30/0.4 or 75. 

What if one were to consider an SRL East variation? It has some long gaps between stations so let's add a few more - say at Warrigal Rd, Clarinda and a second at Monash.

While we're at it let's also replace Glen Waverley with Mt Waverley to improve directness but keep the length the same by terminating it at Doncaster rather than Box Hill. 

This changes the SRL to a 10 station project, undoubtedly adding costs (unless offsetting savings can be found elsewhere). Travel time is slower due to the extra intermediate stations - let's call it 0.5. But it's still very fast. 

That gives a merit number of 90/0.5 or 180, or more than double what we had before. SRL costs would undoubtedly rise with this approach but not double, meaning that an SRL option with more stations could have higher benefits relative to cost. 

Anyone can dream up random formulas at 5am so I'm not going to claim this is particularly rigorous. But I think something like this vaguely makes sense when initially comparing options. 

It should be acknowledged that some of the extra stations (eg Warrigal Rd) are never going to be the busy hubs in terms of land use that say Box Hill or Clayton will be. However they do provide connections with more routes, thus multiplying (not adding) travel opportunities. 

If one was to do a journey plan a huge number of trips would include the SRL, especially one with well planned multimodal interchanges, as one leg of it, even in cases that might not be immediately obvious. This is because its large time savings over existing buses (and for that matter driving) draw a lot of trips (including some induced demand) into it, even if an extra interchange is required. 

How adding frequency can multiply interchange points

The formula isn't just useful for infrastructure-oriented megaprojects.

If you take the view that a location only qualifies as an interchange point if routes intersect at high frequency, have easy physical interchange and the geometry is sound, you will find that the vast majority of potential interchange points are hampered by low frequency and/or poor physical access. 

As an example just four out of nearly 400 bus routes in Melbourne run every 15 minutes or better seven days per week during the day. That figure at night? Zero. 

Boost frequencies and you create reliable interchange points where mere crossings of routes used to be. 

Compared to the SRL the number of interchange points created when you upgrade a bus (especially one that intersects with a large number of already frequent routes such as routes like 241 or 508 in Melbourne's north or 246 in the inner-east) is very high. A snowballing effect starts if you upgrade existing trains and trams as well as parallel lines become a frequent grid. 

It does not require upgrades to that many trains and bus frequencies to create 60 or even 100 new interchange points, ie ten times more than the Suburban Rail Loop. And instead of the cost per each interchange point being $6 billion it is a tiny fraction of that (sometimes even zero), albeit mostly operational expenditure from running that more frequent service. 

Travel speeds uplifts aren't as high for each interchange point created but because you are upgrading so many interchange points (with so many new connection possibilities) for such little cost the scale of benefits if done at scale could easily equal that of a megaproject like SRL and far outstrip any gains from Airport Rail. So much so that 10 min frequencies on multiple lines and routes should be done as soon as bus and train drivers can be recruited. 

Buses are not necessarily directly substitutable for the SRL but their great scalability and very low cost per improvement makes them an important part of any network development strategy. That is because we don't have the capacity to add 100 connection points through SRL like means in the next decade but we certainly can (and much more) with bus and train service upgrades. Also such network reform can complement the SRL and is arguably even necessary for the SRL to work with reasonable usage as it makes getting to its stations easier from a large catchment. 

Conclusion

To summarise, something like the SRL East might be triple the capital cost of Airport Rail but it might benefit many more that in terms of the number of trips. Along with land use influences, which Airport Rail on a premium fare will not at all.

Although this is contingent on really good interchange design across all modes (people have expressed concerns about 'Cheltenham'/Southland in particular) along with bus reform and consistently frequent service on intersecting train and tram lines. 

While they're very different projects to SRL East and Airport Rail, rail and bus frequency increases, and for bus network reform, a method based on interchange points and travel speed savings may still be useful to compare the merit of various network proposals. 

In summary the results from my crude 'analysis' are: 

* Train frequency uplifts, bus frequency boosts, bus network reform: Low cost per interchange point created, making it highly scalable with dispersed benefits. Should be done yesterday, next best time is now. 

* Suburban Rail Loop: You can make good cases either way as it radically changes the network's topology from radial to a web but at substantial cost per new interchange point created, tending to concentrate benefits. Good interchange design is critical. Variants may increase the number of interchange points with higher benefits. 

* Airport rail as currently proposed: Benefits hard to quantify the benefits as it changes neither the network nor the city very much. Individual benefits don't stack up either; saving 2 million people 2 hours each per year rounds to nothing whereas saving 20 000 people 200 hours per year is transformative (even if total hours saved in each case is identical). 

What do you think? Comments are open and I'd love to read them. 

See other Useful Network items

TT 234: How often? A look at the midday frequencies of Melbourne's buses

Most cities have bus routes that vary from infrequent to frequent. While the majority of routes may be infrequent and a minority frequent, you will typically find it is the frequent network that caters for a disproportionate number of trips. Either due to them serving a popular or densely populated corridor or because their better service induces higher usage. 

The quieter less frequent routes tend to have a 'safety net' coverage function while the busier frequent routes present a stronger case for further service and infrastructure upgrades that further boost usage. Examples include higher frequencies, wider operating hours along with reinforcing land use policy such as location of major destinations, denser housing, road space allocation and parking requirements.  

Frequent routes are are also more marketable - something important for cities that wish to grow bus usage. Marketability is one reason transit agencies may differentiate frequent routes from the others. Examples of differentiation may include special bus livery, better bus stops, greater prominence on network maps, and/or distinctive route numbering.  

Route number memorability and frequency

Last week Adam Bain discussed whether some route numbers are more memorable than others and whether the more memorable numbers also happened to be the more frequent routes. This analysis included a chart colour coding each route, listed in order, by its midday weekday frequency. 

High frequency routes included the high frequency university shuttles (mostly x01 numbers), Smartbuses (900-908), inner urban routes (mostly 200s). Distinctive patterns were noted for 20 minute headways (some middle northern, middle south-eastern and outer suburbs), 30 minute intervals (north, east and south-east) and 40 minute gaps (outer suburbs). 

My video here talks you through this in more detail. 

The frequency of frequencies

While people sometimes describe public transport networks in terms of number of routes, that tells us nothing about how long and how frequent each route is (and thus its population catchment and usefulness).

As I mention here fairer metrics to give a high level idea of the network's usefulness include annual service kilometres and the proportion of people near frequent service. Dividing annual service kilometres by metropolitan population tells you about service intensity provided while patronage (and especially patronage productivity) data reveals something about whether the service provided is being used. 

Last week I looked at the patronage productivity of various bus routes in Melbourne. Amongst the most productive bus routes were often those that were more frequent. From this you could deduce a non-linear relationship between service levels and patronage. Not only do frequent routes carry more passengers as they are frequent but because they are also more productive the number of passengers is disproportionately higher. 

That causes you to get a pattern where a small proportion of bus routes (say 20% - I haven't crunched the numbers) account for say half or more of bus passengers. And in Melbourne's case it's helped but not entirely due to our three very long orbitals. (Those who study tax systems would likely find a similar pattern with income taxes - in a progressive system most taxes come from the top 10 or 20% of regular wage earners, not only due to their higher incomes but also because they pay a higher proportion).  

Jarrett Walker often mentions that under conditions of constrained resources a bus network optimised for coverage has a high number of infrequent routes while one optimised for frequency might sacrifice coverage for frequency. Most cities would have bus networks that sit somewhere in between these extremes with a mix of frequent and infrequent routes.

One (admittedly imperfect) way of analysing a bus network is to examine the number of routes in each frequency band. That helps you understand the character of a network, ie one that's simple, direct, frequent (and possibly more widely spaced) versus one that is complex, indirect, infrequent but with short walking distances to the nearest (sometimes poorly used) route. 

Through manual counting I converted Adam's coloured chart to a frequency distribution for midday weekday bus routes in Melbourne (below). 

The big picture is that Melbourne has very few all day frequent bus routes. The network is characterised by a large number of infrequent routes, a feature which marginalises buses compared to trains and trams despite the bus network's extensive coverage.  

Most notable is the dominance of routes operating in the 31 to 60 minute frequency range. The vast majority of these are either 40 or 60 minutes with 40 minutes far more common than 60 minutes. Melbourne (and Victoria's) love of 40 minute bus frequencies is notable compared to Brisbane and Perth which have a lot of weekday 60 minute frequencies. 

Frequent routes running every 15 minutes or better are uncommon in Melbourne, accounting for just 7% of the total on weekdays. On weekends this slumps to under 2% of routes with only the 234, 246, 732 (part), 905 and 907 complying. 

While it is common for cities to have more infrequent routes than they do frequent routes, Melbourne's fall-off is particularly extreme. This can be put down to:

(a) Some skewing of data due to the use of trams on corridors that in other cities would be frequent buses and our long orbital bus routes.

(b) Our weaker political and institutional support for bus reform than Auckland and Perth who prioritised the creation of frequent routes from multiple less frequent routes. This reluctance has caused us to have fewer 10 minute frequency routes than is otherwise possible within existing budgets. 

(c) A tendency to favour 20 over 15 minute frequencies, largely due to our base 10/20/40 minute headway hierarchy on most railway lines. 20 minute service is a big improvement over 30 or 40 minute service but does not generally qualify as turn-up-and-go frequent service like 15 min might and 10 min probably will.  

On weekends this is the most influential of the lot: 

(d) legacy of a 'small city' bus funding culture that like Adelaide (but not Sydney, Brisbane or Perth) schedules even main routes at half frequency on Saturdays and (even moreso) Sundays. 

The graph's numbers are tabulated below (with routes every 20 min or better indentified): 

Every 61 min or worse:  34

Every 31 - 60 min: 167

Every 21 - 30 min: 72

Every 16 - 20 min: 57
153 154 170 180 190 192 200 207 232 250 251 270 390 406 408 420 460 465 467 471 475 477 501 503 505 506 508 510 512 517 527 528 529 533 537 540 541 546 561 570 582 603 604 605 630 732 733 767 791 798 800 825 828 831 893 897 898

Every 11 - 15 min: 18
201 216 220 223 234 305 472 552 670 703 900 901 902 903 905 906 907 908

Every 1 - 10 min: 7
235 237 241 246 301 402 601

Note: 695F excluded, 745 counted as 1, 900 series night network excluded

A reminder that the above relates to midday weekday frequencies. Over half of those listed fall to 30 or 40 minute headways on weekends.

The number of routes that are better than every 15 minutes or better seven days is so low as to be counted on one hand (234, 246, 905, 907 and part of 732). 

Frequency of frequencies with reform 

Bus network reform such as required to massively increase bus usage would almost certainly change the mix of service frequencies. There would be fewer limited service routes with long gaps while more routes would run more frequently. 

The chart below indicates potential directions of change. 

Early priorities would likely concentrate on closing the longest waits, addressing operating hours issues such as early finishes or lack of weekend service and boosting weekend service on otherwise frequent routes.

Weekend services key

While that might not cause the above weekday-oriented graph to change much I would rate the creation of more 7 day frequent routes as being key to to boosting bus usage. Indeed going from 4 to say 40 all week frequent routes would transform the bus network with these routes driving patronage growth. The graph would likely still contain high numbers of less frequent routes but with the majority of bus trips being made on a more defined frequent network with an Auckland-style maximum wait service guarantee the passenger experience would be transformed. 

Network reform and route consolidation

Further gains are possible (with even lower costs per route) with network reform. This would seek opportunities to consolidate routes with overlapping catchment into more direct and frequent routes. The biggest single shift here would likely be the creation of many more 20 minute frequency routes from existing 21 - 30 and 31 - 60 minute routes. Some of this would be straight service upgrades while in other cases network reform would reduce costs.

Going from every 30 to 20 min needs 1 extra bus per hour added. As does moving from 20 to 15 min. Whereas moving from 20 to 10 min needs 3 extra buses per hour. Expressed in another way, assuming you're starting at a 30 minute frequency you can have three times as much route upgraded to run every 20 minutes as every 10 minutes. 

Still the travel experience of 10 min frequencies are so good (and rare for buses in Melbourne) that all opportunities to do it cost-effectively with network reform such as may be possible along corridors like Murray Rd and/or Bell St should be explored, while also reviving probably the single highest patronage potential element of the now sidelined northern and north-east bus reform. 

The above would also be consistent with Victoria's Bus Plan whose priorities were local reform followed by buses being a mass transit option, ie more routes running every 10 minutes seven days. This would be a radical step compared to now where the current network has no seven day routes every 10 minutes or better and just four seven day routes every 15 minutes or better. 

Interaction with train frequencies

Auckland opted to reform their buses before it reformed train frequencies. While some connections were messy (20 min trains to buses every 15 min) the bus reform was an overall greater good with patronage rises. Perth already had 15 minute train frequencies compatible with its 15/30/60 min bus hierarchy (though it now also has some 10 minute bus frequencies). 

In Melbourne's eastern suburbs the potential pace of bus reform is constrained by weekday train timetables which remain on a 15/30 min off-peak pattern, contrary to the 10/20 minute pattern applicable elsewhere (including the Sandringham line from later this year). This is because if trains remain at current frequencies the case for improving their busier bus routes from 30 to 20 min is weaker than with harmonised 10/20 minute frequencies.

There is also a general fall-off in population density when you go more than 1km east of Stud Rd such that it is hard to argue for 10 or 15 minute frequencies but 30 minutes is inadequate. In this context 20 minute bus frequencies on main roads are justifiable especially if supported by similar train frequencies. Which in turn meshes well with the very strong case for stations like Box Hill and Ringwood to have 10 (rather than 15 minute) midday rail frequencies.

The older parts of Melton is a contrasting example. In their case trains got boosted to every 20 min on weekdays. Later its trains got boosted from every 60 to every 40 minutes. No bus upgrades accompanied its weekday train boosts. Thus some buses that had 2 trips per hour were a poor fit for the upgraded 3 trains per hour. The opportunity to upgrade these to 3 buses per hour (as well as simplify routes to run along main roads) was lost, unlike the case of Wyndham in 2015 that got a new bus network when Regional Rail Link commenced.

However much more recently when weekend Melton trains improved there was a genuine attempt at multimodal planning with many bus routes improving from every 60 to every 40 min to harmonise. Creating an interesting case where some bus routes are more frequent on weekends than they are on weekdays (where an hourly service still remains).  

Overall Melbourne has been weak at this sort of multimodal thinking. There have been many cases of infrastructure upgrades (such as level crossing removals) that have not been followed by commensurate rail service boosts or a reappraisal of bus networks to take advantage of road improvements such as grade separations. 

Summary

Melbourne has very few frequent bus routes. Operating more of them, whether through straight service boosts (particularly weekend) or bus network reform is key to unlocking the potential of buses to carry millions more passenger trips per month. 

While it has limitations, analysing frequency data per route can help us quantify the network's extreme bias towards low frequency and less productive services and drive interest in getting more from our investment in bus services.  

See other Timetable Tuesday items here

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