UN 119: What to do about poor performing bus routes

The overwhelming issue with buses in Melbourne is that service is too sparse rather than too generous. However there remain occasions where thought might be given to reducing or even removing service on a route. 

Three reasons for this include: 

1. The government wants to save money, and, with significant public subsidy, transport services can be a target. Depending on how this is done cuts can be indiscriminate with even popular routes facing reduced service. 

2. The route gets so little use that even the social rationale for its continuation is weak, especially if there are alternative services nearby. 

 

3. The route is quiet and there is a ‘greater good’ use for the money spent on it that would benefit more people (eg a frequency or coverage upgrade on another route)

The first reason can happen when state finances are strained. Major bus cuts occurred about 30 years ago in the Cain/Kirner era. Notwithstanding significant Bracks/Brumby upgrades about 15 years ago, the axe marks of these swingeing cuts remain in some bus timetables today whose services have yet to be restored. 

More recently we've seen per capita cuts with service kilometres lagging population growth. With about a million more people calling Melbourne home since it added the last SmartBus, the average resident is further and further away from frequent service.  

An example of (2) happened recently when Route 673 got deleted. This was an off-peak weekday route that duplicated all but one of the stops on two overlapping routes in the Lilydale area. It ran for years with very low patronage. 

Most interesting to students of bus reform is reason (3). Because we've had long stagnant periods in buses where little happened, there is a huge backlog in bus service reforms that could be desirable. In some cases there exist poorly used routes that are either overserviced or duplicative of other routes which got added later. Past risk-averse transport ministers have let them be but their retention indicates an opportunity cost that is stopping us getting the most from our bus network by putting buses on routes where they would be better used.  

Last year’s Victoria’s Bus Plan was largely a ‘plan for a plan’ in that it had neither committed funding nor concrete measures attached. However it correctly recognized the need for simpler, more frequent and more direct buses. 

A start was made in 2021 with a redistribution of service hours between quieter and busier Transdev routes. The Night Network reforms which saw special routes being consolidated into upgraded regular routes with longer hours marked a further improvement.

The more reform you want the tougher will need to be the appraisal of the existing network, especially if there isn't much additional funding. This makes identification and treatment of under-performing bus routes doubly important. If you don't do it right you won't be able to reform much and will get an Adelaide-style backlash if you try.  

Where are the quieter bus routes?

A long bus route may have many passenger boardings. A short one may have much fewer. But the shorter one isn't necessarily less productive. A good way to compare productivity is to use a figure like passenger boardings per bus service kilometre or hour. This compensates for route length and the number of buses used. 

Infrastructure Victoria's Draft Infrastructure Strategy has mapped our more and less productive bus routes (as they were in 2016) with the line drawn at 20 passenger boardings per hour. Because in-service buses operate (on average) at 22km/h this is about 1 passenger boarding per kilometre.  The orange routes are under 20 boardings per hour while the red ones exceed this in the map below (click to enlarge). 

It's hard to see individual routes on a Melbourne-wide map but the big picture is clear. Our productive bus routes are mostly in the outer west (notably Wyndham and Brimbank), parts of the north (notably Craigieburn) and the middle south-east around Box Hill and Oakleigh down to Dandenong. Quieter routes are overwhelmingly in the outer east and north-east with smaller clusters around Sunbury, Melton, Greensborough, Brighton, Frankston and Casey. One should however be cautious about using this map to draw conclusions about routes around Mernda as their rail extension wasn't open then and travel patterns may have changed.  

Reasons for differing bus route productivity vary

You need to know why routes are quiet or busy before deciding what you want to do about it (if anything). Reasons for below average performance might include: 

Some of the above is in the hands of the road planning and traffic management authorities. Other aspects can't be changed, such as catchment demographics, petrol prices, working patterns and other external factors. However most of the listed points are within the powers of those who fund, plan and contract public transport services to change. 

Not all points apply equally to each route. In some cases there's things about the route that could (and should) be changed for better performance while in others the route is about as good as it could be with external factors more decisive.  

I discussed how excess productivity can wreck a bus route and stifle patronage back in 2019. Also back then I covered the 10 quietest least productive routes. Today I'll stay on that theme but be more wide-ranging. 

Scope for remedial action

Werribee's Route 439 (visible in the south-west of the map) is one of the below average routes. However its catchment includes a lot of low density market gardening areas between Werribee Station and high-density Werribee South beach. It provides significant unique catchment with no other routes being nearby. It doesn't need many buses to run (maybe only one) and all but a couple of other City of Wyndham bus routes are above average patronage performers. You would likely leave this route as is, accepting its lower than average boardings per hour as the cost of maintaining coverage. 

Having said that scope may exist to boost 439's usage given the popular destinations it serves through better network information and promotion. This is a recurring theme across the network. Promoting service is more important than promoting infrastructure but most marketing budget goes to the latter. The Department of Transport also often gets information wrong and undersells benefits when service upgrades do occur (with frequency improvements on the 82 tram and 788 bus being recent examples).    

The abovementioned 673 (which no longer exists) is the opposite extreme. It had no significant unique catchment as its stops were (with one minor exception) were shared by other routes going to the same destination offering longer  hours and better service. Given the route's low existing and potential patronage the decision to remove it was fair. 

In between these extremes is the large 'grey' area occupied by most of the routes marked orange on IV's map. You wish to neither accept the underperforming routes and timetables as they are, nor simply delete them without offering a 'greater good' benefit in return. It is in this space that the real work of bus network analysis and reform needs to be done as solutions are less unambiguous. 

 

Ways to boost a route's patronage performance

There are several ways, listed roughly in order of complexity and benefit, to boost a route's performance. Earlier ways are narrower while later points take a broader view, considering the overall network rather than a route in isolation.    

1. Get the most from an existing route and timetable. 

Includes looking at issues with existing service delivery, consider whether passenger information or promotion could be improved, and checking that stops along the route are in the right places and are accessible.  Sometimes there need to be minor tweaks to the timetable to correct for run-time issues. Where buses are being delayed by traffic or lack walking access you need to call in the road authorities to upgrade priority and pedestrian access. While such improvements often have high benefits, our institutional structures and the political bias towards big infrastructure can make these slow and expensive to arrange at the hundreds of locations that need them.  

2. Examine both existing and potential patronage and the timetable's suitability for it. 

Then modify the timetable accordingly. Choices may include: 

a. Increase service on a route

b. Decrease service on a route

c. Redistribute service on a route from quieter to busier times and/or between different parts of the route

d. Redistribute service across a network, from quieter to busier routes and times

The first option (a) involves additional resources. A good example of when this was done on a large scale was when numerous bus routes gained evening and weekend service hours between 2006 and 2010 under the Meeting our Transport Challenges plan. Overall patronage increased by the same percentage that service kilometres did. One effect of adding evening trips was that usage increased earlier in the day as there was now a bus to take people home at night. Hence patronage can grow at times other than just when extra trips were added. 

MOTC upgrades have largely stalled in the last decade or more and there is a lot of unfinished business. There are still some quite popular routes that finish at (say) 5 or 6 pm rather than 9pm. Or they may not run Sundays despite dead quiet routes like 704 jumping the queue with 7 day service. Northern suburbs around Reservoir have many routes whose 22-24 minute frequencies do not evenly mesh with trains every 20 minutes. In this case a frequency upgrade could improve intermodal connectivity. 

Where a route finishes before the pm peak has finished there may be cases where adding a few trips (whether on a weekday or weekend) to increase the span by 1 to 3 hours on a route may disproportionately increase the route's patronage. Midday or early afternoon Saturday finishes haven't made sense since retail hours were deregulated in the '80s and '90s but in 2022 we still have bus routes with them. 

Many bus routes remain a gamble on public holidays for the lack of service on about ten days of the year. Reduced summer timetables can trip passenger up and make using buses needlessly complex. Particular opportunities for cost-effective upgrades exist in areas where routes are productive but underserviced such as in high needs/high patronage areas like Glenroy, Springvale or Dandenong. 

Point (b) is where routes are over-serviced relative to their patronage or patronage potential. The most glaring examples are relatively frequent routes through semi-rural or industrial areas. These are mainly found to the east and north-east of Melbourne and include portions of SmartBuses like 901 and regular routes like 695. 

In some cases trips can just be deleted though a wider network rethink is more desirable. Removing trips (especially if you have usage numbers to justify a decision) is always less controversial than deleting a route outright. This example from Perth shows how you can achieve the latter longer term by whittling away service on a route considered redundant over several timetable changes. 

While it is desirable that all routes run reasonably frequently seven days over long hours there may be locations with poor road layouts or particular demographics that could benefit from an interpeak shopper style of service to retain coverage. As this does not use peak buses such a route might have relatively low operating costs. This might not in itself carry many people but still performs a social and community connectivity role in an overall more productive network. 

Your analysis might find that a route performs well at certain times of the day but not at other times. The difference might be enough to consider redistributing service within a route (c). Some cities have done this, due to COVID-driven reduced commuting activity, by thinning out peak service to boost off-peak service for a better overall all-day service offering. I discussed this, mainly for trains, here. 

Some of this happened in last year's Transdev timetable revamp.  For example Highpoint's 223 got a 7 day service every 20 min or better in exchange for reduced late evening frequency. The outer part of Route 907 got a reduced peak frequency (enabled by starting some am peak services part way along it) but countervailing 'goodies' included improved weekend frequency including 24 hour service. Similarly the new Night Bus network reduced some overnight frequencies from 30 to 60 minutes but extended 24 hour service including on critical but too often ignored Sunday morning timeslots. 

This approach can sometimes be useful but is limited in application - most Melbourne bus routes do not have 'fat' in excess service that could reasonably be cut for redistribution. Where it exists (eg Reservoir's 552 that has a 15 minute Saturday morning service that could more usefully be reformed to provide a more uniform frequency that meshes with trains every 20 min) it has been on established area routes that have rarely been high priority for review.     

Point (d), still dealing only with timetables and not routes, is not quite a network revamp but is heading towards that thinking.  In its simplest form it keeps resources within the one bus operator, which is administratively simplest. A good example is again the 2021 Transdev timetable revamp, which saw service kilometres transferred from the poorly used but overserved 603 and 604 to busier routes like 279 and 907 that needed more service.  

A more sophisticated version involves transferring resources between contracted operators but this is rarer given DoT's preference to confining service changes to one operator where possible. Doing large-scale network planning on an operator rather than area basis may have contributed to the political failure of Transdev's 2015 Greenfields network as it failed to deliver compensatory upgrades to other operators' routes in the west and would have left passengers worse off. The most recent large area-based network reform involving multiple bus operators was Brimbank in 2014 (also in western Melbourne).   

3. Simplify or modify routes individually

Complexity can turn people off using buses. A complex route running to a reasonably frequent timetable can be a poor patronage performer. In this case it might be worth examining the route in more detail. Maybe the route runs a confusing loop or hairpin shape that leaves people wondering where it goes? Where not many people are travelling through an interchange scope could exist to split complex routes like the 280/282, 380, 513, 566, 736 and 834/835 into two easier to understand direct routes, even if timetables are unchanged. 

These types of splits really help network legibility and serve all existing stops, making them fairly uncontroversial. Despite this none have been reformed in recent years. Indeed we actually added the new indirect Route 469 in Transport Minister Ben Carroll's seat of Niddrie instead of having it as two simpler straighter routes split at Airport West.    

Other routes have kinks or occasional deviations that make travel slower than it needs to be. If that can be straightened without losing coverage then it might improve patronage productivity. In lucky circumstances the time saved may enable a frequency increase that assists connections with trains. This further boosts the attractiveness of the service. Relatively recent minor straightenings include routes 512 and 624, though there is scope to do much more with routes like the 506, 556 and many others.

Especially where routes are very long it is no good just looking at patronage productivity numbers for the whole route. This is because it is an average that conceals high and low results encountered on different sections of the route. Some routes pass through catchments varying from dense to sparse. The same bus can be anything from almost empty to overloaded depending on which segment of the route it is on. 

If this is a consistent pattern you may be able to lessen the number of trips that go end to end but add short trips so that a more frequent service is provided on the busier portion. Or if there are few through passengers, the route is long and there is a logical terminus scope may exist to split the route into two, with frequencies more aligned to existing and potential usage. 

Most passengers get an overall frequency gain so such a change would meet a 'greater good overall' test for service reform, while again keeping service at every stop currently served. The biggest scope for network reform of this type are undoubtedly some of the SmartBus orbitals that waste some of their high frequency and long operating hours in sparsely populated areas that don't justify the service while underserving dense or high social needs areas, especially on weekends. 

Splitting would permit optimisation of service levels and scheduled train connections (which is currently pretty much impossible due to unharmonised frequencies and multiple connection points). As well cases splitting creates ends which can be extended to bring major destinations such as La Trobe University Bundoora onto the SmartBus network for the first time.   

 

4. A real network review (involving multiple routes and potentially operators)

This approach also looks at routes but from a network point of view. It's potentially more controversial but the gains are higher if done well. Examples of wins possible include improved directness, higher frequency, longer operating hours and better coordination with trains. This is paid for by deleting or modifying duplicative routes. Timetable tweaks like discussed above are worthwhile but can only go so far. In many cases the best solution to underperforming routes is inclusion in a network review. 

Such a review can enable steps other than retaining, rescheduling or deleting the poorly used route. For example it could be extended to a more logical terminus (eg a railway station on another line) to make its usage more bidirectional throughout the day.  Or a route, instead of being simply deleted, could be replaced by an extension on another route or improved service on one nearby. 

Consolidating multiple routes, each operating every 30 to 80 minutes into a single simpler 7-day route every 15 to 20 minutes timed to connect with local trains may assist in areas with multiple overlapping poor performing routes such as Beaumaris, Brighton, Templestowe, Greensborough and Eltham. These types of changes make network reform more a 'swings and roundabouts' thing with an overall greater good rather than a straight cut and potentially more politically sellable. 

Areas like Brimbank and Wyndham, which had bus network reviews in 2014 and 2015, have very few underperforming routes, though this is partly due to local demographics being favourable for bus usage. Meanwhile areas that just got new routes layered over others (eg Greensborough/Eltham/Diamond Creek) and/or haven't had reviews for years (eg Croydon/Lilydale, Knox and Frankston South) have many poorly performing routes.  

In Rowville and Lilydale's case Telebus became FlexiRide with few or no changes to underlying regular routes. The result is that these areas have effectively two overlapping half networks with poor patronage productivity on each. Again consolidation opportunities exist, noting that the biggest single cost involved is driver pay.  

It is currently fashionable in parts of bodies like Infrastructure Victoria and the Department of Transport to suggest flexible route buses to replace poorly used fixed route buses or to avoid doing the work required in planning and reforming the latter. Some even think that flexible routes save money that could be invested into main routes. 

Unfortunately flexible routes often have fewer passenger pick-ups per hour than even quiet fixed routes, making such savings unlikely. You might accept flexible route services to provide for a small or specialised passenger base. However the high subsidies per passenger relative to fixed routes means they don't scale up to a cost-effective mass transit job. 

The world is littered with failed flexible bus route trials. Overall we are probably better off to stick with reforming local fixed bus routes, recognising that some will always have low productivity but accepting this as the price to pay for comprehensive coverage.   

Greater Dandenong, is another area with many unreviewed/limited service routes. Unlike in parts of the outer east (eg around Lilydale) they get good usage despite the limited service. This is an area where bus network reviews are desirable not so much to increase (already high) patronage productivity but to provide a much more useful network desirable on both social and patronage growth grounds.  

Similar comments apply on main roads where new routes have been layered over existing unreformed routes. A review may allow economical upgrades such as 10 minute frequency on simplified routes serving popular destinations across Melbourne's north or between Footscray and Highpoint. 

Conclusion

With so many unrealised gains it is not surprising that Infrastructure Victoria has recommended bus network reform, starting with areas served by the Metro Tunnel. Although timetable and route level reform should be pursued (and is easier to do) it is really only network reform along with adequate resourcing that can deliver the high bus usage growth that Melbourne needs. This needs to start by understanding why route under and over perform. That knowledge can then be synthesised with what is known about peoples' travel patterns and preferences to produce a reformed more useful and effective network.      

  

Other Building Melbourne's Useful Network items are here