Monday, September 02, 2024

Book review: Without Delay by Simon Lane (2nd edition)

 


It's rare that transport industry practitioners write full-length books on their trade, especially those with something to say about the Melbourne scene. Books are more likely to come from academics instead. James Murphy's The Making and Unmaking of East-West Link and A Very Public Solution and Transport for Suburbia from the late Dr Paul Mees are examples. 

Academics and their students may also present research via papers or at conferences such as ATRF. Other parties that have recently published research or advocacy material on transport include Committee for Melbourne, Climate Council, Infrastructure Victoria and the former Transport and Cities Unit of the Grattan Institute. And historical material may appear from enthusiasts-practitioners via organisations such as state bus and coach societies or the Australian Timetable Association.  

DTP bureaucrats, transport operators and to some extent consultants are more constrained in what they can say. So a book appearing from one is a special occasion, often delivering a perspective you won't find elsewhere. 

Intended audience

Without Delay is by a rail manager for rail managers. But safety regulators and franchise managers should read it too. Especially those in this state given its pointed comments about them near the end. 

Advocates, enthusiasts and those interested in efficient public service delivery more broadly may also benefit. Though beware; you are not its primary audience. Hence Without Delay has jargon, abbreviations but no glossary. Neither is there an index. So it's best read with a highlighter in hand for easier later reference. 

The author

Who is Simon Lane? You could call him an international rail trouble-shooter. Brought in when service delivery is poor and there is a desire to do better. Or there's a major event coming when trains absolutely have to run properly. Notable Australian experience includes Melbourne's Met Trains under Minister Alan Brown (1994-1997) and the period leading up to and including the Sydney 2000 Olympics,

Later he was asked by the Victorian Minister for Public Transport to review Melbourne's poor train performance during the hot 2009 summer. After Connex lost the franchise to MTM he became interim Chief Operating Officer in 2010-2011 before Singapore beckoned.  

Like many rail managers in Australia, Lane hails from Britain. If you've wondered about that, it's not just our better weather and wages; he explains that Australia and (perhaps surprisingly) Singapore have been suckers when it comes to hiring senior managers who don't meet that country's qualification requirements (p60). That we expect train drivers, signallers and technicians to be qualified, but not senior managers, is one theme of the book. 

Importance of statistical analysis

What are some other Without Delay tips? 'Stretch targets', even those thought impossible, are important. Rail managers may know but not understand the network, especially its vulnerabilities and where it can be improved easiest. Timetable people may blame rolling stock people for faults and vice versa.

To get the truth on what to fix first it's essential to use data to ascertain where delays happen and their real causes. You can play along at home with (some) of the former through Track Record's interactive dashboard. Or you can go back through old annual reports (like I did in 2006). Such a statistical approach is essential to identify the lines with the worst performance, with some groups performing better than others. Variability is also important. When improvements from one initiative (eg fixing train doors) become discernible there is less 'noise' in the numbers, making the next problems to fix stand out. That gives useful guidance for the following year's work program.   

Attend to common faults

While major network meltdowns (like Oaks Day 2008) get the most publicity, it's the common everyday faults, such as might arise from fragile timetabling, operational sloppiness or unreliable train doors that contribute most to mediocre daily performance. Lane says that too often this is accepted in a culture of 'learned helplessness' that accepts achieving a franchise target as 'good enough'.

This can be exacerbated by a desire (arguably an obligation) for a franchisee to maximise shareholder profits rather than provide the best service it can be - ie risks and delays as low as reasonably practical). It takes two to tango, so this mediocrity can only happen if the network owner/franchise manager (in our case the state government through DTP) is too accepting. 

Accountability

Past Melbourne on Transit articles (eg thisthisthis and this) have noted the 2003 - 2010 fall, the 2011-2014 rebound and the slow 2015 - 2019 decline in metropolitan rail's operational performance. 

Who is at fault? Lane said that all but one of the eighteen ministers he has encountered have been open to his advice (p339). Franchise operators are legally bound (by Australian corporations law) to put their shareholders first (p309). This puts the onus on the contract owner (ie DTP) to uphold asset stewardship and operational performance. Any slackness is quickly noticed and exploited by the operator, who become masters at knowing what they can get away with ('the normalisation of deviance') and in pleading mitigating circumstances.

Noting the post 2015 deterioration, he said that today's contract managers were too tolerant, and those of the late 1990s would have driven better performance such as was actually achieved as per graph below (p311).  


Would-be rail managers (whose role should be to build systems that endure after they leave - described as the 'flywheel effect') are encouraged to seek qualifications and avoid what Lane calls the 'amateurish good bloke' theory of management. Reading inquiry investigation reports for air, sea and rail disasters is also essential - these have lessons that apply for reliability as well as safety.

The three cultures

There is a focus is on avoiding 'loss of control' events through reference to a 'three culture model'. The three cultures roughly correspond to 'bad', 'mediocre' and 'good', with the bad one being described as 'denial or professional recklessness'.

Falling in to this category, according to Lane (p218), was the deterioration in Melbourne train performance from 2004 as the then government blamed the private operator before eventually acting (but too late to electorally save itself in 2010). Lane was particularly critical of 'Meeting our Transport Challenges' (wrongly cited as 2007 - it was actually 2006) as its rail projects, including Dandenong's third track, did not boost available train paths. Based on Lane's classification, I would count fare evasion on Melbourne buses as a more recent sustained 'loss of control' event, including some denial as described here.  

'Administrative or professional negligence' was the second culture. This is what Lane encountered in 1994 (p229) when starting in Melbourne. The difference between this and the first culture was that shortcomings here were less blatant and needed a skilled person to see. There was also an appetite, though not necessarily the understanding required, to improve. MTM's cessation in performance improvement after the 2012 rebound (and the slow decline after) is also cited here (p230).   

Third and best is the 'high performance culture'. Lane says that Met Trains was up to this level by December 1997 (p237) with peak train punctuality close to 95%. Unfortunately there's only a paragraph on this, though other references to this era appear in the book. As the book traces the author's experiences there's significant gaps. For example I'd have liked to have read much more on the 2011-2012 performance revival (which contemporary Melbourne readers and managers will remember better). 

Government versus franchising

Given his background, my expectation was that Lane would prefer a privatised franchise model. However his conclusions are based on evidence that shows the state-run rail networks of Perth, Sydney and Brisbane are performing better than Melbourne's franchised model on reliability and customer satisfaction. Although he didn't mention Brisbane's poor patronage performance, poor frequency and (arguably) less efficient staffing practices compared to say Perth.

In relation to franchising, if you are going to do it he suggested the Network Owner (ie DTP) give franchisees more autonomy in how they run a good service, but being firmer on them in performance demands. 'Partnerships', such has become the fashion, have lead to blurring responsibilities, poor accountability and overstaffing. Page 310 says that the pre-franchising Department of Transport (ie circa 1999) had 25 staff compared to 1200 now. 

I would like more evidence of like-for-like comparisons before accepting this statement on departmental staffing. This claim is not unlike that from the late Paul Mees who said that the department was grossly overstaffed and said that it could manage with about 30, as he said was the case in Zurich. Plus there were some important transport functions in 1999 that were just not done, eg maintaining bus stop signage and timetables (which had to await Metlink's arrival in 2004 to happen), not to mention timetable data, websites and journey planners.  

Infrastructure priorities

While most of the book is about 'O and M' (jargon for operations and maintenance), there is a section on planning for new projects. Unlike the current 'infrastructure-first' state government, Lane is a 'sweat the asset / service first' man. Big builds are the last resort, done only after possibilities such as boosting frequencies, adding more carriages or modernising signals have been been exhausted. 

He notes that some major projects (including metros in Sydney and Melbourne) are marginal or negative BCR. Rigour would be improved if proposed projects were subject to detailed pre-feasibility studies and were compared against good quality options, rather than against what he said were rail's equivalents of nails, broken glass or manure, as in this Goodies clip (which the book frequently quotes). 


Something the book doesn't capture, but which could be a problem is that administrators tend to be able to only juggle a few balls at once. In our case it is possible that, without special attention, our large building program risks distracting attention from operating the service well, especially on lines where replacement buses are the rule rather than the exception. That has impacts on patronage and fuels a 'tyranny of low expectations'.   
   
What I would have liked

Without Delay gives rare insights from an industry practitioner. It is worth purchasing for this reason alone. Especially if you're a problem solver who prefers raw to varnished. 

The second edition had improvements over the first. But if a third edition was to come out then better editing is needed. Might I be so bold to say, by a professional, not an amateur. This is to avoid the current edition's (a) heavy repetition, (b) use of jargon and abbreviations (with no glossary), (c) appearance of some factual errors (including the Goodies skit dog food options), (d) spelling mistakes (including both American and English on the one page), (e) better quality/more readable diagrams and (f) need for more detail in some case studies presented.

Better editing would also return its cost by making Without Delay accessible to a wider audience including  advocates, academics, policy analysts, administrators and media. Many such people will not have  direct industry experience but may still be influential in advisory or franchise management roles.

Notwithstanding its limitations, which are shared by many (almost) self-published limited-market books, I still recommend Without Delay as an essential read for anyone interested in safe and reliable rail operations. And, if it's any comfort to those struggling, I found the last half an easier read than the first half.  

Buy Without Delay via Amazon

Note: Affiliate Link. I receive a commission from sales. No extra cost to you. 

Thursday, August 29, 2024

UN 183: Melbourne trams versus the world


Some interesting commentary on how our trams compare with other countries' here (from the same people who brought you those SNAMUTS city maps showing public transport's haves and have nots). 

It starts by asking whether Melbourne trams are the slowest in the world (spoiler: they're not - that honour goes to Toronto, which is so good with buses but so bad with trams) and finishes with many other graphs and numbers. 

Speed

As can be seen below, Melbourne's weighted tram speed, in both kilometres per hour and relative to other cities terms, fell between 2006 and 2016. There was however a rebound in 2021, possibly due to pandemic-related factors. 


Our most recent average of 15.5 km/h is far better than Toronto's at 10.7km/h. Zurich, often praised by transport academics and commentators, has performed poorly on tram speeds, dropping from better than Melbourne to worse than Melbourne. Sydney, and some northern European systems, operate faster than ours. 


Catchment density

Another graph compares job and resident density around tram routes. In 2006 Melbourne was, with Adelaide, at the bottom of the sample but had increased significantly by 2016. Sydney fell then rebounded, with the large changes likely due to its growing tram network extending to less dense and then much denser areas. Increased density is good for potential patronage but may slow trams if there are also more cars, bikes and pedestrians. 

Frequency

This is yet another addition to the mountain of evidence that Melbourne is deteriorating on service frequency, especially when measured in per capita terms or compared against Sydney.  

Melbourne averages 8.8 trams per hour. This is higher than the typical route's frequency (5 or 6 trams per hour) because the calculation factors in sections with multiple routes (mostly in or surrounding the CBD). 

Despite its decline, Melbourne still occupies a respectable middling position on frequency. However note that they are based on weekday interpeaks, where individual routes are at or close to turn-up-and-go frequency. However if the graph was based on maximum waits across the week, Melbourne would be a poor performer with 20 and even 30 minute intervals between services the rule at nights and certain times on Sundays. 

Conclusion

There's two lessons out of this data if we want a world class tram network. That is: 

1. Run trams faster through priority measures that give spatial or temporal separation from car traffic

2. Run trams more frequently, especially during popular night and weekend times

We also need to exploit cascading opportunities, such as presented by the Metro Tunnel's relief of the Swanston St corridor, to reform the tram network so that routes more evenly reflect CBD activity and there is better connectivity in CBD fringe areas, including to new stations such as Arden, Anzac and new routing via Victoria St and Park St


See other Useful Network items here

Thursday, August 22, 2024

UN 182: An always there network - the two principles public transport needs to work


Always there

If you want people to use public transport in big numbers it needs to be 'always there', 365 days of the year. 

The concept of mass transport operating on fixed routes to fixed timetables was devised by mathematician Blaise Pascal in the mid-1600s.

This was revolutionary; the concept of running a service even if there were no passengers wanting to ride seems counter-intuitive and wasteful. 

However it started the idea of an always there service that people could rely on and even base life decisions, like where to live and work, on. You could call it the 'always there' effect, with no need to book.

An extension of this idea is that service exists over a wide span of hours from early to late, ie not just in and between the peaks. Hence when Melbourne boosted many of its local buses to run 7 days until 9pm in the 2006-2010 period, patronage on these routes increased, including at times when there was already service.

Extend the idea more and you get a frequent service. Which in the best cases mean turn-up-and-go. Public transport then becomes something like water flowing out of a tap, there whenever you want it. Frequent service is true 'on demand', ie on your terms not theirs, provided you're willing to walk a little distance to it. But the trade-off is that by being on fixed corridors you can give it priority over other traffic and build shelters with seats for the short time that you are waiting.  

Over 400 years on and there's people who think they're smarter than Pascal. They back 'demand responsive' service like FlexiRide, which has neither fixed routes nor fixed timetables. Our state government embraced this two or three years back but has since cooled on it as limitations became apparent.

For example it is using GAIC developer funding to replace Tarneit's struggling FlexiRide with more suitable fixed routes. And it hasn't progressed moves to introduce FlexiRide in Greensborough/St Helena, initiated when FlexiRide hype was at its peak. Melton people, who have suffered from their FlexiRide not reliably meeting hourly weekend trains, will be hoping they follow Tarneit soon with more fixed routes. 

So Pascal got the last laugh; compared to fixed route/fixed timetable service, flexible routes are either poorly used or max out to become unreliable if more than a few people want to use it. And the idea of  an 'always there' service remains powerful today. 
 
Network effect

As cities grew it was found that not every destination could be accommodated on a single route. This necessitated transferring for some trips. It was found that further ridership gains were possible by making this easy.

This requires several ingredients: (i) good route geometry to provide adequate speed and coverage and minimise backtracking, (ii) the shortest possible distances between stops of intersecting routes to aid physical transfer, (iii) integrated fares to remove the financial cost of connecting and (iv) either very frequent service or timed connections to minimise waiting times. 


The above rules are not always obeyed by those responsible for building the network. For example  some rebuilt stations (eg Mentone and Edithvale) and so-called accessible tram stops are further from road intersections and buses than those they replaced. And Infrastructure Victoria keeps peddling its modal fare moonshine. A good thing the state government is ignoring them on this.  

Another example of the network effect can be seen in the success (in terms of inducing more driving and dispersed land use that induced it even more) of the US interstate highway system, as explained here. This showed that providing a ubiquitous connected network was more important than trying to assess cost-effectiveness on individual sections or reactively relying on existing demand for future planning. Whether it's public transport or roads, if you make something better more people will use it, with these gains being consolidated as the improvements shape habits and longer-term decisions. 

The benefits of major infrastructure projects are enlarged if other modes are considered. Sydney's new Metro got a lot of limelight this week. However it was accompanied with significant bus network and timetable reform. The aim of this is to reduce network duplication and provide some connectivity improvements to the new Metro stations. Hopefully we get bus reform of a similar scale for Melbourne's Metro Tunnel, with some concepts for Watergardens to Dandenong discussed here

An always there network

Marry these two ideas to get an always there network

Go to Doonside in Sydney's outer west and check its train timetables. You will see a 7 day service about every 15 minutes even at 5am Sunday or midnight any night. Ditto for most other stations in Sydney. That's what an 'always there' network looks like. It's legible and easy to communicate. And it helps explain why Sydney's train patronage is much higher than Melbourne's despite having fewer stations. Sydney's Metro has lifted the bar even further with 5 to 10 minute frequencies common. 



You can't use Sydney's density or traffic to explain all this away either. Other city comparisons demonstrate the centrality of the 'always there' network, even if other factors are unfavourable. Brisbane versus Perth is the most notable Australian example. Despite having half the lines, half the stations and a lower population, Perth's rail system outperforms Brisbane's on patronage due to operating its trains every 15 minutes  all week versus 30 minutes for Brisbane. Unless it's after mid-evening there'll always be a train soon in Perth. And the planning of its bus routes, with them feeding rather than paralleling trains gives Perth a network effect that Brisbane lacks. 

Even when it decides to substantially increase metro rail services (rare events insufficient to arrest declining per capita service), Melbourne has taken a different tack to Sydney. Instead of going for the 'always there' approach it only boosted services during times that it thought would be busiest, typically midday and not early morning or evenings. 

This habit goes back years. An example was the 11am - 7pm increase in Sunday train and tram frequencies while Jeff Kennett was premier in 1999. 2012 saw a further big increase to every 10 minutes on weekends for Ringwood, Dandenong and Frankston (under another Coalition government). However the busier lines of Ringwood remained with 30 minute gaps weekend evenings and Sunday mornings, with the quieter but marginal seated Frankston line getting a modest upgrade to every 20 minutes at all times but early Sunday morning in 2021 (under Labor).

What this means that even on our busiest line (Dandenong) you get frequent service between about 11am and 7pm. But outside this service falls sharply to every 30 minutes on weekends, especially after 7pm and Sunday mornings. 


The sharp fall-off means that if you are making a return trip of more than a few hours there is a high chance that you'll get a frequent 10 minute service in one direction but face 30 minute gaps in the other.  This is very different to a flat Sydney-style timetable with 10-15 minute intervals early to late. Hopefully the Metro Tunnel fixes this for Dandenong - Watergardens but there is not yet word on whether other lines will get relief from their 20, 30 and 40 minute gaps at key times people travel.  

Service designs with inconsistent frequencies or sharp fall-offs also hurt messaging. Overall a consistent 7 day 15 minute frequency, like at most Sydney stations, is more sellable than what's found at Dandenong, which could be every 10, 20 or 30 minutes. It also makes early start/late finish trips much more attractive, with weekend V/Line tourism to the regions and airport trips the key beneficiaries. 

Sydney's approach of a frequent service until after midnight is obviously the gold standard. But smaller cities, such as Perth and Auckland, have a pattern of 15 minute daytime Monday to Sunday frequency on their key routes, aiding legibility. Melbourne doesn't have this consistency, with early weekend evenings and Sunday mornings being a particular unresolved issue over many decades. 

Conclusion

Public transport is not just about building infrastructure, as prominent as this is at the moment.  

Design and operation of the network is equally vital. 

This should be informed by the concept of an 'always there network' as outlined above. 

This would make public transport services useful in many more peoples' lives, and maximise the benefits of the infrastructure we build. 

See other Useful Network items here