in Uncategorized

IT failure – the integrated ticketing system

The papers reported this week that the RPA has spent 9.5 million euros on the integrated ticketing system. There’s nothing to show for it. I’m not surprised (and I have been saying this privately for over 3 years).

Here are the problems I saw (three years ago).

– There are no clear requirements for the whole project. Basically this is a project which had no foundation. No one seems to be even able to remember why the system was required in the first place

– The technological basis for the solution (smart cards) seems to have been decided before the requirements and the problem were clearly defined.

– There has been no meaningful consultation about the requirements for this project. Neither the public nor the operators were asked (in a meaningful manner) what they wanted. Everybody involved just seemed to assume that a smartcard system would be the right solution to the problem, even though nobody knew what exactly the problem actually was.

– The transport system is a political football. There is a phony war going on between the government and the unions over the issue. The RPA is stuck in the middle.

As a result, the requirements of the transport system is a rapidly moving target. This makes it basically impossible for the RPA to even figure out even roughly how many companies will actually be involved in integrated ticketing. This makes a big difference to how it is implemented.

– The ‘business rules’ of the system are not clear. It is absolutely impossible to develop a ticketing system without knowing at least roughly what the business rules (which would describe how the funds will be collected and distributed) will be.

– There is too much business risk. No vendor can guarantee the success of the system, because the business issues (i.e., the future structure of the sector) would be completely outside its control. This would appear to be the reason that the tendering processes for the system have collapsed, when the vendors refused to bid and pulled out of the process. This has now happened twice.

– The RPA has almost no maturity as an IT organization and has absolutely no experience with developing major IT systems. I have seen similar situations in my work with another major rail agency (not in Ireland). It is not a pretty situation. You cannot procure IT systems in the same way that you procure railway lines and carriages. The parameters are completely different.

– The system required is technically far beyond anything that has been done before anywhere in the world. It is too complicated and too expensive. Certain implicit requirements (such as the requirement to be able to deal with tens of small bus operators after deregulation) have never been fulfilled by a transport smartcard system anywhere in the world before, and may well be impossible to fulfill.

– There has never been a trial or prototype of integrated ticketing, even though almost 10 million euros has been spent. (There have been trials of smartcards, but they have never been trialled as a means of providing integrated ticketing.)

– There are unrealistic timelines. Realistically, it will take at least 5 years to implement such a complex system (it took longer to implement a much simpler system in London). But the RPA keeps on saying that it will be able to deliver the system within three years. In 2003 it claimed the system would be delivered by 2005, in 2005 it claimed it would be ready for 2007. All of these claims are and were rubbish. There is no way that any of these timelines could possibly have been met.

This project now needs to be stopped, now. It should be frozen until it has meaningful requirements and achievable objectives.

People complain about using consultants, but the RPA needs to get some decent help in to sort this problem out. It simply doesn’t have the skills in-house.

At the same time the promise of a decent integrated ticketing system should not forgotten in all this mess. This type of system is definitely needed. However, we seem to have the wrong approach to the problem.

(I should say that I have no internal information whatsoever about the progress of the project, but I have read all the publicly available information.)

Write a Comment


  1. hi, your article is really good despite your lack of “internal information” 🙂
    Anyway my take on it:

    1. “There are no clear requirements for the whole project”

    Couple of things on that – firstly, Dublin Bus’s equipment is totally out of date, falling apart and has to be replaced, smartcards are the natural, next generation replacement. Secondly, if the Govts agenda is transport integration then a single ticket platform or media is required so that travellers can switch between connected (“integrated”) transports..

    2. The technological basis for the solution (smart cards) seems to have been decided before the requirements and the problem were clearly defined.

    Sounds like it, but what else? No point sticking with unreliable magnetic tickets (equipment maintenance more expensive). As above, smart cards are next, alternatives like mobile phone payments not likely to happen here for years yet!! From a technical design perspective with multiple organisations there is bound to be conflicting requirements. question is, how/have these been resolved – my guess is that they haven’t and is probably one of the reasons that this project is going nowhere

    3. “There has been no meaningful consultation about the requirements for this project”
    Sounds like crap to me. Sunday times article a while back had quoted them as meeting at “workgroups” for years, though obviously not achieving anything. Article said that they were all fighting with each other, obviously not agreeing on the project. Though why they went to tender in this situation I’ve no idea. Also, I remember some blurb from the RPA on focus groups to do with the smartcard, so presumably they actually did ask joe public what they thought (though maybe they only asked one or two people!!).

    4. “The transport system is a political football”
    Absolutely agree with you. Why would you have RPA doing this in the first place after taking it back from CIE!!
    Also has a lot to do with deregulation – with effective all dublin tickets it should not matter who runs/owns the trams, buses or trains/metros.

    5. “There is too much business risk”
    Yep, why should a company take the risk when the main parties involved are fighting – probably meant that they came in with a price that was off the wall, and I’ll bet money that that was the reason the tendering process collapsed.

    6.”There has never been a trial or prototype of integrated ticketing”
    Thats because the project is only for smartcards and not for creating “integrated” tickets. And anyway, “integrated tickets” are only a small part of real “integration”, which to my mind means a stop or station where you can switch easily between bus,train,metro etc., ideally with the one ticket. Project probably never should have gone ahead in the first place without having sorted this first.. Is there even an all dublin ticket at the moment for all transport – bus, luas & irish rail?

    7. Unrealistic timelines – again I think you’re on the money, Oyster still hasn’t got fully sorted, I remember that they had posters telling people to ring in if they were charged the wrong amounts!!

    8. “RPA needs to get some decent help in to sort this problem out”
    Haven’t they done that? Wasn’t the Indo article criticising them for using “foreign consultants”?? If they haven’t I would be very worried, I agree with you, they should have experienced IT people (not railway people) for a project like this.

    Real problem is that nobody has the balls to make the tough decisions. if this integrated ticketing is important for integrated transport there should be someone independent to run this. Giving the project to the RPA was only bound to get up CIEs noses and if deregulation is the real issue then there is no point going ahead with this project without someone to enforce it. Surely to god the Cullen or Transport can get someone to sort this out??

    BTW I use the luas smartcard and I’m happy with it, though it looks naff.. Would like to be able to top it up online though…

  2. Thanks for the kind words! A few points.

    1. Dublin Bus now has new equipment coming on-stream. Replacing ticketing equipment is actually not a really big deal.

    You are right, the requirement is clearly to integrate these systems. However, what is not stated is what or even how many systems should be integrated, how the integrated ticket should be priced and how the integrated revenue should be divvied up.

    2. Technology alternatives. Well, maybe a paper ticket with an honour system would have been a better short-term solution.

    Mag cards can actually be very cheap and maintenance free. Look at NYC’s MTA metrocard.

    I agree that the smartcard is a pretty good technology. You can do a lot with smartcards.

    However, the sophistication of the technology is a double-edged sword. It can tie you up in knots. Just because something is possible using smartcards doesn’t mean it is a good idea. The way things were looking, the smartcard was just going to be used to facilitate complexity rather than to simplify the system.

    3. Consultation: I meant ‘meaningful’ consultation. They did focus groups and called it ‘consultation’. Focus groups are not consultation. Surveys are not consultation. The sampling for these surveys looked very dubious. The briefing for the studies involved showing a video which was not really representative of what is actually going to be delivered. The consultation with the operators I can’t say too much about, but it doesn’t look like it really took the views of operators into account. If it had they would have realised the project was on a hiding-to-nothing and stopped the digger.

    5. As I understand it the tenders collapsed before any pricing was submitted.

    6. You are incorrect in saying that the project is only a smartcard project. The project is called ‘ITS’ for integrated ticketing system’. It is not a smartcard-only project. Anyway, the individual operators all have their own smartcard projects.

    The ticket you mention doesn’t exist, other than if you buy certain rail tickets which allow you access to bus and luas interconnections. I guess it could exist, if there was the political will. You are absolutely correct, having a ticket like this should be the starting point.

    8. Consultants: They obviously haven’t got good advice. Otherwise they would have stopped the digger 3 years ago. The warning signs were all there.

    I absolutely agree that tough decisions are needed, but I can’t see how having an agency other than the RPA would have made things any easier. The real issue is the complex interplay between the unions, the state, CIE management and the private bus companies.

    You can’t really do online top-up for a smartcard, it’s one of the big problems with it. The japanese have a system that allows this though, by incorporating the smartcard into a mobile phone and it is (theoretically) possible to integrate an oyster-style smartcard into a system like this.

  3. The most significant gap from my point of view is that the customers’ need to have a simple but universal system with predictable journey cost appears not to be a consideration.

    The Dutch _national_ strip-ticket system is an example of a system which “just works”. With this, the traveller can use long-distance and municipal bus, tram, metro, and even (within certain limits) the train. Apart from a well-understood penalty for using the train, journey cost is determined by the zones traversed and time taken. A big advantage for the customer is that the cost is not dependent on either the carrier or whether a line-change is needed. The technology involved is primitive (impact printing on card), but allows self-service validation of the tickets, although this facility is not available in all circumstances. I don’t know anything about the business rules, but I certainly like it that the state dare put the citizens it represents before commercial interests.

    A totally different system which also seems “just to work” is the integrated electronic purse system used in Istanbul, known as “Akbil”. This supports a specific tariff for each mode or operator. I’ve used it on bus and metro (really light rail), and understand that it also is accepted on the trams and ferries. The sensor unit mounted by the entry door or turnstile has a display panel to show how much is left in the purse.

    I don’t have much time for the “complex interplay” argument; it’s the state’s job to work for the citizens. Once that role is made a priority, the complexity is less significant.

  4. Nice to read your comments – it makes me feel much better knowing that the NSW, Australia integrated ticketing system is not the only failure in he world ! Reading this article i thought i was reading about our own system !


  • The Train Standing at Platform 8 is… March 15, 2007

    Dublin’s public transport ticketing project is flawed…

    The integrated ticketing project for public transport in Dublin is not only overcomplicated, costly, and hampered by infighting of inter-state agencies – the project ignores an alternative which could be implemented with speed and little cost. Accord…