Edinburgh Trams: Perception vs Performance

Edinburgh’s Trams may yet provide the ultimate test of perception over performance. I walked the entire route (by nearest footpath) on launch day, then rode the tram back. I preferred the walk.

The Experience

Inevitably there are teething problems on day one, however much things are tested. Many are trivial. For example, nobody seemed to have the correct key to unlock the gates at Murrayfield station, so passengers had to find the one gate that was unlocked – or just jump over them. The gates are presumably there to stop people in the stadium using the tram to get home from events. For as will become clear, the tram operations lack the redundancy to deal with unusually large volumes of passengers, and even a fraction of Murrayfield’s 67-thousand capacity will swamp the dinky little 250-capacity trams.

Fortunately Murrayfield is the only station with gates, or fences for the most part, else the trams might have spent their first big day carrying fresh air. Not that the air they are carrying is at all fresh.

Busy, but not overloaded, the environment on board reminded me of an old (London) Bakerloo line train in the summer: Largely devoid of air. Nauseating, not life-threatening, but the sort of experience where one is consciously glad to get off and breath in Essence of Car Exhaust. There was no effective air flow through the vehicle when filled with standing passengers, with insufficient open-able windows to offset the obvious deficiencies in the roof ventilation. The temperature outside was under 20oc. Did CAF‘s designers consider Edinburgh to be too far north to ever get warm? Perhaps whatever cooling system was supposed to be working, simply was not.

It goes without saying that none of the electronic information systems survived their first encounter with the enemy. The public transport industry has been failing to get such systems to work reliably for decades, so this is hardly a surprise:

By the middle of the day, the platform screens had ceased to predict the arrival times of the next trams, instead defaulting to a distinctly optimistic statement that trams were running every 10 minutes. The network-wide speaker announcements to the same affect became rather insulting, as any passenger with a watch could calculate that, having waited 15 minutes, and seen only one tram travelling in the other direction, trams were most definitely not, operating every 10 minutes. Problem, what problem? Most rail operators learned to keep passengers informed of delays by human voice within a minute or two. The voice is not just a surrogate for the visually impaired, it conveys something quite different. We accept that digital displays are machines and prone to actions beyond their control, but human voices are human, and should behave with respect and integrity. The Uncanny Valley demands it.

Then when the tram did finally arrive, it enthusiastically announced “Gogarburn” at every stop all the way to city centre. Damn stupid machines.

Other technology failings were more ambiguous. Did the “Not in Service” tram display that information via a large white wooden board in the driver’s cab because the swanky electronic destination display was unable to handle anything more challenging than “Airport” or “York Place”? And if so, why not just paint those destinations on the respective ends of each vehicle? Or because Lothian buses still maintains a tradition of using large white wooden boards to convey certain information? Perhaps as a tribute to old city trams, which I suspect is where these white boards originally came from? And when I say original, I wouldn’t be surprised to learn I mean just that.

Certainly the timing of the platform announcement warning of the approach of the aforementioned tram need fine-tuning: There was much amusement among the waiting passengers when the announcement triggered a few seconds after the empty tram had passed by.

But at least the wheels didn’t fall off. Nobody got run over and killed. The staff didn’t go on strike. And the sun shone brightly. The prime aims of the operation were successful. The carnival atmosphere of launch day shielded most passengers from the reality of what would, had it been any other day on any other public transport network, been regarded as a poor experience. I wonder how long the novelty factor will last, if the standard of the operation can be improved before the tram gains a negative reputation among users?

The Operation

While busy first days may not be the best judge of what is to come, from what I saw, operationally Edinburgh’s trams leave much to be desired. By the afternoon 20 minute gaps were appearing in a service advertised at 10, sometimes with another tram 5 minutes behind (the classic urban bus problem), but sometimes with another large time gap (suggesting a minor operational meltdown was in progress). Worse, these problems gradually snowballed through the day, as if there was no option to reset the service, for example by pausing delayed vehicles at the end of the line, while putting extra trams into service ahead of them.

I presume that the vehicle-to-service allocation mistakenly uses a railway model, not an urban bus model. In a railway model, if 6 vehicles per hour are run, a train more-or-less every 10 minutes is generally achieved, because operators have a high degree of consistency in, and control of, operations. In urban bus operation, one may need to run as many as 9 or 12 vehicles an hour to guarantee a vehicle actually arrives at each stop every 10 minutes, such is the lack of control on operating conditions that urban bus operators can experience.

While the city centre section is subject to excessive variability at traffic lights, most of the line should function like a clockwork railway: There are no complex junctions or service patterns, not contested (over-capacity) track or trains moving at different speeds to contend with. Boarding times were likely higher than anticipated on the first day, but those I saw remained well within the design of the vehicle, those massive doors and wide internal spaces are designed specifically to deal with mass boardings. None of the traditional bus boarding variabilities – like the stereotypical old granny who can’t find her bus pass, who remains blissfully unaware of the operational carnage she is inflicting – should apply to the trams.

Trams stabled at depot
Trams stabled at depot.

If there simply weren’t enough trams in operation to maintain the schedule, the travesty would be apparent to any journey passing the depot: The majority of the tram fleet were stabled at the depot throughout. I counted over a dozen vehicles stabled in the open, probably a few more in the sheds. This excessively large fleet is a relic of the line’s haphazard procurement process: A process of learning the hard way, what most of the British public transport operators already knew: Why specify such a unique design? Why not just lease what you are using? That much is “water under the bridge”. The concern is that much the same lack of redundancy seen in the line’s assets, also applies to its operations:

A rather finite number of tram operating staff appear to have been trained solely for that task, such that there are no spare staff available to cover excessively busy days. Days when one might want to use all those extra vehicles. Bus operators typically achieve far more redundancy in their staffing – able to move staff between routes and vehicles, tweak shift patterns to cover exceptional events, send managers into the heat of battle, do what has to be done to get buses into service. Knowing that small networks are generally very inflexible – meaning that when they have bad operational days, those days are very bad indeed – operational integration between tram and bus would seem logical. I can only presume the idea of a dual-skilled bus-and-tram driver/operators would have evoked a labour relations crisis, although I’m unsure which:

  • The spectre of RMT demanding twice as much pay than bus drivers, to maintain parity with train drivers – because trams are basically trains, right?
  • Or a return to the late 1980s minibus era, where new vehicle types primarily allowed new drivers to be employed on less generous terms than established drivers – because trams are far easier to drive than buses, right?

In the meantime trams will need to be over-staffed, or accept that on a proportion of days, the service quality will suck. As the most politically visible (and vulnerable) transport system in Edinburgh, under-staffing isn’t acceptable, implying more staff, meaning higher operating cost. That in turn makes the financial position of the trams even more precarious, increasing the chance of failure.

The city end of the line is not well designed. As far as I can see, there are no turn-backs or passing loops after the West End station. York Place uses a single platform terminus. The net result is roughly 20 minutes of running which is entirely sequential. Or bust. If anything blocks the line in the city centre, the effective service collapses. Effective service, since the majority of passenger journeys seem to go as far as Princes Street, so terminating trams short of Princes Street is not an attractive solution. Presumably in the rush to save costs by not building the planned route to Leith, nobody reassessed the operational requirements of the line, which almost certainly would have originally included a passing loop at Elm Row – the bus (and former tram) terminus, a short distance beyond the current tram terminus in York Place. Given the years of disruption and complex subterranean work required to build the current tram lines through the city centre, it is unlikely that anyone will be willing to make further engineering changes, such as adding a cross-over near Waverley.

The operators implicitly acknowledged their situation by keeping an engineering spare vehicle stabled outside Haymarket – engineering, not operational, spare, because it appeared to crewed by a couple of CAF staff, waiting in a nearby van. Clearly with half your fleet stabled at the depot, one can afford to have a multi-million pound asset shooting the breeze at Haymarket all day, right? I’m sure that impresses the bankers.

The only depot, home to both stabled trams and operating staff, is about 10 minutes from the Airport end of the line and 30 minutes from the city. Such positioning is notorious inconvenient, since trains can’t just be pulled in and out of service at the point that service starts – both vehicle and crew need to shuttle in and out of the depot first. The net result is that every crisis takes at least 10 minutes longer to react to than it would if the depot was somewhere more convenient.

Not that I’m impressed by the way routine staff changeovers are current handled. I was reminded of the sort of operating procedure that characterised Irish Railways in the early 1990s. A classic operator-first, passengers-second operational logic, which doesn’t belong in a modern public transport system:

In-service trams routinely stop at a private depot platform, holding there for around a minute while a crew change occurs. I presume this practice reflects the lack of crew facilities at either end of the line. I couldn’t find the most basic toilet or cabin-with-a-kettle staff facility. The facilities in the airport itself are too far away from the tram terminal. So it seems that every trip requires a crew change at the depot, which means every journey sacrifices about a minute of journey time for all the hundred-plus passengers aboard. Good job none of them are in a hurry (for as we shall comment on later, anyone in a hurry took the bus)!

Why not change crews at the nearby Gyle station, where the tram at least has to stop for passengers to board? Looks feasible, but as I discovered, there is no safe walking route across the Gyle roundabout. It’s a take your life in your hands and run across 3 lanes of fast-moving traffic affair. Indeed, I had just posed the question to myself, “who the hell designed this”, when I was rewarded with a detailed list:

Edinburgh City Bypass - Sighthill Section
Edinburgh City Bypass – Sighthill Section.

Whatever happened to that Alistair M. Darling chap? The bypass itself has (by regulation) no pedestrian access, so I doubt the engineers gave the junctions much thought. In fairness the area was probably considerably less built up in 1986, and the traffic flows likely much lower.

The Corners

When I first encountered the route of the Edinburgh Tram it was as a guided bus-way proposal. Indeed a part of the guided bus-way was built, and briefly operated, before being closed and paved over with tram tracks. Unfortunately the tram’s alignment is even more tortuous than the proposed guided bus-way, while trams, on account of them being on fixed tracks, have considerably less corning finesse than a Leyland National. Tight cornering is acceptable on city street sections of tram routes, where buildings constrain the route, and operating speeds are already low due to the presence of road traffic. We even implicitly accept new heavy rail alignments can not easily be forged through suburban environments. Edinburgh’s failing is that where a perfectly good – meaning straight and flat – heavy rail alignment is available, an excruciating number of tight corners or unnecessary gradients have been engineered in. Excruciating both on the ears and on the sense that the journey is slow.

Tram on the curve after Ingliston Park and Ride
Tram on the curve after Ingliston Park and Ride.

The section between the Gyle and the Airport is over beautifully flat, admittedly somewhat boggy, low value semi-agricultural land. It shouldn’t have been difficult to plan a line that was straight and fast, with gentle curves where required. Instead we have no less than four abrupt 90 degree turns in this section, each of which is traversed at walking pace to the audible pain of the wheel flanges. A walking pace that is twice as slow when one can simultaneously see the Airport bus flying past on the main road – a testament to the fact that highway engineers understand there is a correlation between route alignment and vehicle speed, something tramway engineers apparently do not.

The section between Haymarket and Edinburgh Park is reminiscent of a failed roller-coaster ride, in which the engineer has been challenged to make a straight, neatly graded railway alignment, as wobbly as possible. Why build an embankment at a fixed height when you can continually raise and lower the track every time you want to bridge a road? Why insist that the railway flyover has to be made at (roughly) 60 degrees to the railway below, requiring sharp turns on and off, when there seems plenty of room for a longer, smoother crossing? Both solutions are cheaper, but unlikely to have much impact on the £775 million ($1.3 billion) price tag of the entire line.

Tram approaching the Balgreen railway flyover
Tram approaching the Balgreen railway flyover.

The Edinburgh Tram system (as ultimately built) has minimal on-street running: Everything outside the city centre, from the Airport to Haymarket, is neatly segregated. There are remarkably few graded (level) crossings, and those few are generally over minor access roads. All this should give the system a heavy-rail quality, akin to tram systems in Manchester (at opening) and Birmingham, where trams move at railway-like speeds on former railway tracks in the suburbs, merely slowing for street access in city centres. Unfortunately the Edinburgh tram’s aversion to straight alignments reduces its function to that of a bus route. Albeit one using metal wheels and overhead power.

It comes as no surprise that the advertised journey time from the Airport to the city centre is 5 minutes faster by existing bus than by the new tram, with comparable frequencies (headways). Indeed, the tram has no discernible access advantage over the bus: One walks straight past the airport bus stand in order to reach the more remote tram terminus. Buses serving the Gyle Shopping Centre pull up right at the entrance, while the trams ostensibly serve the edge of Morrisons’ car park. Even the interchange with Waverley railway station is suboptimal, as the tram majestically swings right past the station entrance without actually stopping – while the airport buses line up right outside. In all honesty, the only place the tram serves better than the buses is Murrayfield Stadium. So long as you’re only there with a couple of hundred other people.

The Advantage

Lothian Buses have historically maintained a significant premium on fares to the airport, compared to all other journeys in the city. For the bus industry, this is a rare form of product differentiation, in which traditional politicised equality-based priced (broadly, same price for everyone making journeys of a similar distance) is eschewed to serve the needs of a wealthier market – one appreciative of a higher quality of service, and prepared to pay extra for it. The extent to which aviation still attracts wealthier people may be disputed in the age of budget airlines, but the differentiation remains unquestioned. Like buying food at the cinema, local residents instinctively know its a bit of a rip off, but they pay anyway, because like a trip to the cinema, a trip to the airport is invariably special.

The original assessment of the tram’s operating costs and revenue relied heavily on maintaining this airport premium. Otherwise the tram would be generating bus-like revenues, yet with train-like operating costs – surely a lose-making proposition: While trams carry more passengers than buses, they also require more staff to operate – crudely, every tram has an operator and conductor, while buses are normally operator-only. Bus operators share (in social spirit, if not necessarily in monetary transaction) their road building and maintenance burden with a wide range of road users, while trams are the sole users of their tracks, so tend to pay disproportionately more “way-cost”. And that’s ignoring initial investment costs, which tend to be significantly higher for trams than buses. In short, the passenger volume at which it becomes commercially sensible to switch from bus to tram provision is considerably higher than most transport planners like to admit. Since the airport tram route will struggle to conveying enough people to justify the tram commercially, even with some fairly optimistic growth assumptions, it is necessary to charge at least some of the passengers extra. Quite a lot extra.

Britain’s deregulated bus environment allows anyone (with the appropriate expertise, facilities and 28 days notice) to start operating buses between the airport and the city centre, charging whatever fare they wish. Access to the airport itself is more moot, since the airport owner could theoretically regulate access to their private roads, although pragmatism generally prevails. In practice, the only serious competition seen in Edinburgh has been strategic – which is a polite way of saying, dominated by irrational personal vendettas, which serve only to demonstrate the strengths (or otherwise) of the individuals involved. Still, reliance on a premium, which is likely well in excess of the cost of providing the service, is a vulnerability. Especially for a tram line, because if the competition becomes too intense, one can’t simply move one’s trams to Inverness and compete there instead, as certain bus operators have done in the past. Trams and tram lines are an extremely fixed asset by local transport standards, with few options for doing anything they weren’t originally designed to do.

In the best traditions of Stagecoach (notably whose Manchester Megabuses competed against both branded Stagecoach vehicles and other competitors, all on the same routes), Lothian Buses has decided to protect its position by competing with itself for different market segments. Currently it offers:

  • Airport tram – adult £5 single, £8 return, £9 day Edinburgh-wide bus/tram ticket
  • Airport bus – adult £4 single, £7 return
  • 35 bus, or walk to Ingliston Park and Ride then tram – adult £1.50 single, £3.50 day Edinburgh-wide bus/tram ticket

Both bus number 35 and Ingliston Park and Ride offer the (far cheaper) city fares. The 35 is direct, but takes almost a hour to reach the city centre from the airport. Ingliston Park and Ride has direct access to the tram, but adds a (rather exposed) 10 minute walk from the airport terminal. Both options have existed for several years (Ingliston Park and Ride was originally served by bus), but few people appear to use them to cut the cost by two thirds. In spite of two thirds being “quite a lot” in old money.

The cynic cannot help but notice that, even away from the airport, Lothian Buses are using hoardings next to the tram track to advertise competing bus routes. “Every few minutes” being considerably more frequent than the tram:

Advertising the 22
Advertising the 22.

Oh, and before your inner transport economist exclaims, “that means the value of time is precisely £6.42 per hour!” (Benjamin Franklin has a lot to answer for), consider that the tram is both slower and more expensive than the airport bus. Or that if the tram conveys any passengers whatsoever from the airport to the city, it has surely broken rational economic logic. Could it be that money, and indeed time, are secondary factors in transport mode choice?

The Track Factor

Outside the depot I happened upon a pair of Scandinavian tram enthusiasts. At least I shall call them that, for their knowledge was far broader than the average ‘spotter. They struggled with one word they could not translate – “the track factor”. The track factor is the ability to see a public transport service when the temporal element of the service is not yet there (or, has already departed – but one has to be quite pessimistic to have any use of a transport service that has already departed, so we’ll stick with the future). We see the track, we implicitly know there will be trams. We see the road, we don’t necessarily know there will be buses.

I suspect the track factor is more biased to perception than those statements suggest. Public transport users are reasonably adept at identifying services which aren’t physically there yet. A bus stop, even just a bus stop flag, is normally enough to spark an awareness. The track factor is more relevant to non-users, even people that have no intention of ever using a service themselves, but who gain tangible perception benefits – such as (house) property prices, or attractiveness of inward investment and employment in their locality.

Such a track factor is a less appealing concept in the context of Edinburgh, because the city gained just such a thing 20 years ago when it introduced Greenways: A network of continuous bus lanes, painted bright green – much the same permanent visual indicator of route as tram lines. Greenways lacked specific integration with services – Lothian Buses continued to operate exactly the same services it always had, with exactly the same quality of vehicle. But the core idea is not so new. Perhaps in an augmented reality future, we may merely glance at the street to see an overlay of the bus services that run there – a fundamentally cheaper method than laying tram tracks or resurfacing the highway. Or perhaps this all misses the point of trams entirely:

My classic explanation for Edinburgh’s trams is that “we like trains”. Specifically people outside of Edinburgh, but whom the city might like to attract, “like trains”. For historic and geological reasons (a mix of sustained service quality and hills never giving railways a chance) bus use has remained socially acceptable for a large proportion of Edinburgh society. Hence the native enthusiasm for trams was never overwhelming, even before the troubled construction began. And by the time it began again opinions tended towards the negative. I eventually lost track of all the holes.

The late Colin Eastman (briefly head of transport strategy in Birmingham) expressed this well when he grumbled that Manchester’s tram always appeared in publicity for that city, while Birmingham’s tram (which stopped at the edge of the city centre, pending a city centre extension which never came) remained unseen. In the fierce rivalry between Britain’s second cities, Manchester was winning on perception, and the political high command was not happy about it. In transport terms, both tram systems made relatively insignificant contributions to the commercial viability of their respective cities, with traditional bus and heavy rail conveying far more passengers.

And this is why the Edinburgh Trams are now so interesting: They have absolutely no unique selling point, nothing they are best at, devoid of value, ostensibly useless – except that which they are: Being a tram. One can get to the airport faster or cheaper, probably both, and based on the first day, more reliably and comfortably by using bus services. Or driving. Or frankly, walking, if time really isn’t money. If you like trams, especially slow corner-y trams, that’s great. But if you have any other preference, the tram is hard to recommend.

However, the decision has now been revealed to be perception based, with little to do with fares or journey times or frankly anything that ever went into the formal calculations that justified building the Edinburgh trams. This is no longer necessarily just about trams as a glittering units of carriage, or track factors. Perception can be easily influenced by irrational stuff like marketing, group behaviour, even trivial changes like the position of the tram terminal compared to the bus stand.

Sure, it is possible that enough people just like trams to make the Edinburgh Tram project both a perception and commercial success. If not, we could see some interesting, and potentially painful commercial decisions being taken in the coming years. But ultimately, if the tram transpires to be nothing more than a pawn in Edinburgh Airport’s battle to become the dominant international Scottish airport, or a token to allow the Royal Bank of Scotland to become an even more powerful player in global finance, does it really matter if anyone actually uses the thing?


The lack of redundancy in tram systems is a recurrent theme, and by design. Trams suit highly stable local environments, which is invariably not the environment they are deployed into. Our environments – not least social, political, economic – change, but when we perceive that change to be excessive, we slow down time by evoking stability. Tram schemes are perfect for that. The concept has history, the very tram tracks firmly rooted to the land. But of course the stability we evoke to slow change, remains in constant tension with the process of change. Perhaps it is just as inevitable that these sorts of projects will be pursued, as it is that forces in Silicon Valley be determined to reform taxi operations.

But there is something far more contradictory about Edinburgh’s tram project. While seeking stability, it was justified by optimism. Two opposing trajectories of time, simultaneous in result. And for Edinburgh’s local government, a quite destabilising optimism. A facet of the fact that the further one looks to the future, the more unstable one’s assessment can become. The contradiction is inherent, irresolvable. For if one moderates the grand schemes, such that they take months, not years, one greatly limits the scope for snowballing optimism bias. But without such grand and lengthy schemes, one will be constantly bombarded by change, not least changes of the minor-but-potentially-important variety, perhaps the hardest to manage.

Yet to admit that one is merely a poorly timed juggler of balls one barely understands, does not meet with social expectation of responsibility. This essay can be read as another rant about the infernal tram, this particular one is written by someone who is somewhat complicit, albeit not exactly holding the Lead Piping in the Study alongside the recently deceased Mrs. White, who may or may not have just had a bit too much to drink and could yet be fine in the morning. Fingers crossed. For unlike the neat fantasies of Cluedo, there is no definable perpetrator, because the act is lost in utter complexity. It’s not that I don’t know, it’s that no one can know, yet we persist in that fantasy that someone can, and build our decision making processes and chains of accountability upon it. Such seems a rather optimism strategy.

To write a postscript is, I fear, equally self-defeating. There is an inherent now-ness to our contradiction, which has no place for endings, merely temporal pauses. Sleep on it. For a couple of years. There is more here, which has been festering over the years since I last wrote as myself. Perhaps I should learn to juggle.

1 thought on “Edinburgh Trams: Perception vs Performance”

  1. I couldn’t agree more, Tim (well, apart from the philosophical bit at the end which I haven’t yet worked out!). Your experience seems to mirror my own about 2 months later, when the “every 8-10 minutes” from the Airport turned into a 18-minute interval between trams to the city centre, while the driver tried repeatedly to reset the “passenger information system” (destination screens to us). Not that anybody bothered to explain this … perhaps they left out an internal PA system to help save some money? Definitely back to Airlink next time!

    Meanwhile, keep up the good work.

Comments are closed.