Sunday, February 13, 2011

The Myki Fiasco

The Baillieu State Government is making a decision soon whether to axe the trouble-plaqued $1.4 billion Myki system (Ref 1, 1A). This system has gained such a notoriety that it is being used as a synonym for any problematic project, particularly those that are failing, overdue and over budget. For example, the HealthSMART program that is designed to link computer systems in Victorian hospitals is being referred to as the Myki of the health system (Ref 2).

I am from Singapore and have used both Myki and EZ-Link, Singapore's contactless smart card system so I am able to compare between the two systems. Unlike Myki, EZ-Link was rolled out in 2001 as planned, without major glitches and its smooth implementation was taken for granted, like most other things in Singapore that are run with clock-like efficiency on a daily basis.

Myki is built and operated by the consortium Kamco (Keane Australia Micropayment Consortium), which is wholly owned by Keane Inc, a US-based IT firm. The EZ-Link card is based on Sony FeliCa smartcard technology, which is widely used in Japan and also by Hong Kong's Octopus card, the world's first contactless smart card for mass transit payment (launched in 1997) and the most successful and mature one (Ref 3). Both the Octopus and EZ-Link systems were designed by ERG (now Vix ERG), an Australian-based technology company which has designed, supplied and operated automated fare collection systems for a large number of cities around the world. Ironically, ERG which has created Melbourne's Metcard, had repeatedly offered to upgrade Metcard to a smart card system for under $100 million since the current Metcard validating equipment has built-in support for a contactless validating mechanism (the yellow circles on the front of current Metcard machines) (Ref 4, 5). The main contract was eventually awarded to its competitor, Kamco (Ref 6, 7). The Myki, Octopus and EZ-Link systems are compared in the table below.


MykiOctopusEZ-Link
WhereMelbourne/ VictoriaHongKong/Macau/ShenzhenSingapore
Previous SystemMetcard & VLine
(stored value paper tickets)
Common Stored Value Ticket (flexible magnetic plastic card)TransitLink (flexible magnetic plastic farecard)
Start Date of ProjectAU$494 million contract awarded to Kamco on 12/7/20051994US$78 million awarded to ERG-Motorola alliance on 5/4/99 (Ref 8)
Scheduled Completion DateMar 200719972002
Date Introduced12/12/2008 (Regional Vic) 29/12/2009 (Metro Melb)Sep 199713/4/2002
Cost$1.4 billion (current estimate)$100 million
initial building cost (Ref 9)
US$78 million
contract
(Ref 8)
Transactions per day1 million
(Ref 9a)
11 million4 million
Current UsageTrain, Tram, BusAll public transport (train, tram, subway, bus, taxi, ferry), more than 1000 merchants, convenience stores, supermarkets, fast-food restaurants, parking meters, carparks, photo booths, pay phones, service stations, vending machines, access control cards in buildings and school administrative functions.
Train, Light Rail, Bus, Taxi.

Used as supplementary ID & concession card and for micropayment in McDonald, some food courts, carparks, schools to mark student attendance and to pay for canteen food.

Other DevelopmentsFate to be decided next week, whether Myki is to be retained, modified or abandoned.Octopus Rewards launched on 6/11/2005.

Apart from the card, mobile phone covers, wrist watches, pocket watches, key chains and wristbands embedded with the Octopus chip are available, rechargeable and function exactly like the card.


Trial on mobiles with an embedded EZ-Link card in 2007.

Replaced by CEPAS EZ-Link card on 1/10/2009 that can be used for contactless e-purse applications.

NETS FlashPay CEPAS card was launched on 9/9/2009 to compete with EZ-Link card.

The Myki debacle is not without tell-tale signs. In fact, the seed of the problems to come was sown in the inception stage. According to a leaked Auditor-General's report in 2008, the Transport Ticketing Authority officers selected US company Keane to write key elements of the tender brief in 2005 (Ref 10, 11). This is the same company that later led the consortium Kamco to participate in the tender and win the contract to build the Myki system.

In an apparent conflict of interest, the then Chief Executive of the Transport Ticketing Authority, Vivian Miners was found to be an executive of one company in the Kamco consortium and hold shares in another (Ref 11a). The leaked audit report also said: "Keane had no corporate experience in developing, implementing and operating a ticketing system … Keane has barely demonstrated adequate capacity." In May 2008, the then Transport Minister Lynne Kosky said that "the one mistake the government did make was to actually accept the expert advice that we could put the new ticketing system in place much quicker than any other jurisdiction". By then, one of the experts blamed by Ms Kosky, Vivian Miners, the highest paid public servant in Victoria, had quit his $545,000-a-year position just hours before his scheduled appearance at a parliamentary committee to answer questions about the Myki tender (Ref 12, 13).

Apparently, Kamco has no qualms in extolling its achievements. Johanna Waldon, the Communications Advisor at Kamco, had added to an article in Wikipedia on 20 Jan 2011, the following : "The card readers, electronic gates and other equipment that are part of the myki ticketing system have demonstrated exceptional performance in high-volume transit and vandal-prone sites around the world.....The new smartcard puts Victoria among an elite group of states around the world that are transforming and future-proofing public transport." (Ref 14). What she has omitted is the reliability of the Myki software system and how Myki has performed in Melbourne, which is really what matters rather than what its individual hardware components have performed in other parts of the world. This table shows a large number of cities already using contactless smart cards for public transportation so it is far-fetched to say that Melbourne/Victoria is among an elite group. (Please note that this Wikipedia article has a complete makeover and resembles nothing like its previous version when I read it again on 23/02/2011).

In November 2009, the Transport Ticketing Authority explained why it could not opt for "off-the-shelf" smart card systems through this statement on its website: "while hardware components can be bought off the shelf - and Myki uses some - software must be designed to best meet our state's individual fare structure. This meant maintaining many existing Metcard and regional fare features, while introducing many new myki features - an excruciatingly complicated feat." (Ref 15) It is clear now that the then government had made a critical judgement error in choosing the inexperienced Keane rather than a company with a long proven track record in building successful smart card ticketing systems. Singapore's EZ-Link system is actually no less sophisticated than Myki. Whereas metropolitan Melbourne has only two fare zones (Zone 1 and 2), the train and bus fares in Singapore are computed based on the distance travelled. Buses will have to depend on reliable GPS technologies to accurately compute the fare between the boarding and alighting bus stops. The passenger volume is also far higher in Singapore than in Melbourne, which means that the reader needs to be very sensitive to ensure speedy passenger boarding and disembarkation. I can easily validate the EZ-Link card without taking it out of my wallet and the balance is clearly displayed unlike the Myki machine, in which I find difficulty reading the display from a short distance or at an angle. The Myki machine appears flimsy and less robust compared to the Singapore's one. Apart from public transport ticketing, EZ-Link cards are increasing used in more and more non-transit electronic money applications.

What really perplex me is why the contract was written in such a way that was heavily leaned towards the interest of the successful bidder. I thought that the customer (the State Government) would have the upper hands in dictating the terms to its best benefits. Shouldn't the tenderer adhere to the payment it has agreed so that any blown-out costs should be borne by the bidder rather than by the Government? Doesn't the contract include penalty for delays, unsatisfactory product, failure to meet performance targets and an exit option if things go wrong? It is unfathomable why the State Government should have landed itself in a situation in which it faces hefty, drawn-out legal suits if it chooses to walk away from this project (Ref 16). A protection clause should have been included in the contract to take care of such an "unlikely" scenario.

I really hope that the current and future governments will not repeat the same mistake when negotiating contracts for large-scale government projects. The terms should be in our best interests and should take into account all possible situations. Ultimately, it is the taxpayers who will foot the bill and suffer the most if something goes awry.


22/02/2011 Update: Vix ERG had once again offered to extend the Metcard system with a smartcard capability that can run side by side with Metcard for 3-4 years. Vix Technology chief executive Steve Gallagher said that the equipment can be rolled out to regional Victoria within weeks and to the greater Melbourne metropolitan area within 6 months, with the smartcard function available within 12-18 months. He estimated up to $100 million can be saved by discontinuing Myki but this cost does take into account any legal action from Kamco (Ref 17).

21/06/2011 Update: Premier Ted Baillieu announced that Myki would be kept and modified, with improvements to its capacity and capability (Ref 17a).


Response to a Reader's Comment on 28 Feb 2011 that my discussion is one-sided against Myki

I have provided hyperlinked reference(s) against each statement in the post so that readers can click to check the source of the information or to read further details, including ERG's T-Card failure in Sydney which is often mentioned in newspapers reports such as the recent article on ERG's offer to extend smartcard capability to the existing Metcard. Please note that I have never suggested anywhere in the post that ERG is better or should have been chosen. I would have no knowledge or insights of the professional skills of the respective companies. The post merely states what had been widely reported in the media, that is, ERG had offered to do a smartcard version for the State Government. Since you have mentioned about ERG, I will add the following stuff I read about ERG which may or may not further interest readers:
  • According to its Wikipedia entry, ERG did have a lot of international experience. But even experienced companies, such as Toyota and Boeing, do run into problems.
  • ERG withdrew from an initial bid for Melbourne's smartcard system to join the Kamco consortium instead.
  • Vivian Miner's partner and former wife both worked for ERG and Headstrong, which was part of the Kamco consortium (Ref 18).
  • As recent as 2010, EZ-Link of Singapore had collaborated with Vix ERG on the Next Generation CEPAS Fare Validator (Ref 19).
What I feel is that getting a well-known company is not sufficient in itself to guarantee the success of a project. The managerial role of the Government is equally important.

On the "cost" comment, "initial building cost" was clearly written against the $100 million so that the readers are aware of it. I had also clearly written AU$494 million contract awarded to Kamco on 12/7/2005 and US$78 million awarded to ERG-Motorola alliance on 5/4/99. These are the only figures that I can find from the internet and newspapers reports. By the same yardstick, median house price will not be meaningful as it does not take into account factors such as land size, building condition, workmanship, age and size, proximity to public transport, amenities, etc. But people still love to see the price tags so I am providing the figures but with a qualification and linked references so that readers will not be misled.

The systems in Singapore and Hong Kong are considered highly-successful systems, despite issues that viewers can further read from the hyperlinked Wikipedia articles on EZ_Link and Octopus. When I select these two systems for comparing against Myki, I hope that readers can also open their mindsets that confining the use of the contactless smartcard system to public transport ticketing is not exploiting its maximum potential. Extending its usage to e-purse applications opens a whole new range of possibilities and opportunities that will serve to enhance its value, cost-effectiveness and consumer benefits. I firmly believe this is the direction that our government should pursue and not just contend with a smartcard with only transit functions.

22 comments:

Anonymous said...

Congratulations on providing an excellent insight into a seriously flawed procurement process. It's vitally important that projects are specified correctly for both parties and sadly this appears to be missing here.

Wills said...

Silly government decided to develop a brand new system rather than pick an existing package that has been proven overseas and customise it to our need. That and picking an inexperienced company which outsourced the development and testing to India, little wonder the whole thing is a mess. I hope the government can the whole thing, why throw more good money after bad one? They should have canned it long time ago before it became this bad.

Anonymous said...

Myki is a joke! It is a very typical failure of an IT project.

I don't understand why the state government did not act quickly to dump this incompetent Kamco.
What a shame!

I don't know how this company can fix the following problems if the decision is to keep their flawed system:

1) scanner cannot read a card if the card is not positioned at a certain angle on the reader
2) top-up on the internet won't be available in 24 hours

Anthony Ang said...

It is beyond my comprehension too. Imagine going to an in-store or online shop and you are told to return 24 hours later, as this amount of time is needed to process your payment. Why is on-the-spot transaction so difficult? If your credit card is rejected, you will be told immediately. I am not aware of any other Point-Of-Sale systems that aren't capable of processing your payment immediately.

However, it will be too costly to dump this system. Hope that the State Government is able to force improvements from Kamco (e.g. upgrading expertise of its IT professionals) with any additional costs to be borne by Kamco as it is contractually obliged to deliver what it has promised in its bid.

Anonymous said...

Do you know how hard it is for a private enterprise to deal with government bodies? The company I current work for is current in contract with the government on a project that is almost as large as Myki. My role is to lead and manage the software development side of aspects for this project.

Government is generally uncooperative especially uncooperative in Australia. In additon the high turnover of decision makers in government departments has made configuration management a hell.

The launch of our project was scheduled to be last year, however the government consistently changed their requirements of the system on a weekly basis which was almost impossible to keep up even given our budget.

Recently we spent hundreds of thousands of dollars in setting up internal training programs, all government bodies were notified of the timings of these training sessions months in advance and it was confirmed a month ago that we had the green light to proceed with the training. The commencement date for training was yesterday however NO ONE showed up to class yesterday and we got "just in time" responses from the government bodies that they were not ready.

The Government is obligated to compensate us for these delays, and so far the compensation has been in the millions and it will keep growing until they get their act together.

Anonymous said...

You have obviously put a lot of effort into this research. What a sad state of affairs! I get so annoyed when I see dozens of people on the trams brandishing their myki cards and avoiding paying. Whenever the inspectors come along they just ignore anyone who shows a myki card. I won't take that approach. I just buy my reliable metcard that works all the time.

You really should contact Neil Mitchell with this info as the public needs to know. You never know he might give you some free publicity for your blog.

Well done. Keep up the good work.

Anonymous said...

Anonymous said...Thursday, February 17, 2011 11:59:00 AM

While most would agree the Myki project failure was caused by the government's mismanagement partially, I see no point in blaming the government for technical incompetence.

Did the government request to design the scanner (the basic touch on/off hardware) in such a way that one must take the card out from the wallet and physically touch it on the scanner in order to make it read the card?

Did the government request to make internet top-up unavailable in 24 hours?

If the project had finished on time and could work perfectly like the other smart cards in the world, there would be no complaints.

Anonymous said...

I see the department has put out a press release today saying they have processed 1M fares in a day and they seem ecstatic about the achievement! At a cost of more than 14 times comparable systems that process many times more transactions a day I see little to be excited about. Why should government departments be allowed to get away with this "Yes Minister" stuff. Come on Ted. You said you would get rid of it. Time to deliver on your promise.

Anonymous said...

You are all aware that the $1.4 billion figure that is continually quoted includes the cost of keeping Metcard operational alongside Myki, and the operational cost of Myki for 10 years?

Costs that are apparently not included in the figures for the other systems? (eg $100 million initial building cost.)

Anonymous said...

Funny that your "research" hasn't uncovered the fact that ERG were behind Sydney's failed T-Card venture. Or that Octopus has also had issues, namely deducting money from bank accounts without crediting their cards: http://www.thestandard.com.hk/news_detail.asp?pp_cat=11&art_id=39322&sid=12484549&con_type=1
And yes, ERG offered to upgrade Metcard to Metcard Xpress smartcards for $100m, cards so smart that they could only store passes, not stored value, cost $25 each, and if they were blocked (which happened frequently in testing), had to be sent to Switzerland to be unblocked!

Your article, while it does raise valid points is extremely one sided. As the previous comment advises, you are comparing a full 10 year contract figure for myki, with only initial building costs for EZ-Link and Octopus - not at all a factual or relevant comparison.

It's worthwhile noting that another smartcard system, London's Oyster, has taken 7 years to get to where it is now. And Sydney, on their second go at getting a smartcard will be licensing Oyster at a cost of $1.2b, for a much smaller implementation than myki, which is the largest implementation by area in the world.

Have a look at the Parliamentary Select Committee reports to see some real facts and figures:
http://www.parliament.vic.gov.au/images/stories/documents/council/Select_Committees/Trains/Presentations/Myki_Pres1.pdf
http://www.parliament.vic.gov.au/images/stories/SCTS_myki_22_July_2010_Corrected.pdf

Also the Tourism and Transport Forum's position paper on smartcard ticketing, with a case study on Oyster also dispels many misconceptions.

http://www.ttf.org.au/DisplayFile.aspx?FileID=896

Anonymous said...

Where did the "$100m" Oyster initial build cost come from? Myki is supposedly $1.4b including 10 years running costs (on top of the 3-4 year installation time, total 14 years?); Oyster was £1.2b for a total of 17 years including installation time. This £1.2b included £150m for capital equipment, and was to be paid after installation at £100m per year. The cost of development and installation is not given. (http://www.railwaygazette.com/nc/news/single-view/view/prestige-brings-smart-cards-to-london.html).

The Oyster system is only slightly larger than Myki - and cost considerably more than Myki. The original contractor was removed a few tears ago by the then London Mayor, the reason given being the ongoing costs. (http://www.mayorwatch.co.uk/mayor-cancels-oyster-contract/20081281)

Anthony Ang said...

There is no mention of the Oyster system in this post. The $100m initial building cost refers to Hong Kong's Octopus System. You must have mixed up the two systems, probably both begin with the letter "O". The table is wide enough to accommodate at most 2 other systems for comparing against Myki. I had chosen EZ-Link because I came from Singapore and Octopus since it is the first contactless smarcard system in the world. Also Metro from Hong Kong is now the train operator in Melbourne so it will be of relevance to Melbourne readers.

Anthony Ang said...

I have given a detailed response at the end of "The Myki Fiasco" post to a reader's comment on 28 Feb 2011 that my discussion is one-sided against Myki. This is too lengthy to go into the "Comment" section.

Anonymous said...

You are correct. The reference to Oyster came from the main topic of your reference #9 from your text, and it is this reference that states that Oyster cost less than Myki. Oyster is a good comparison as the sizes of the two systems are similar.
Regarding Hong Kong (Octopus) and Singapore (EZ-Link), the station numbers are about 84 and 78 respectively, compared to about 300 Melbourne Metro and V/line stations covered by Myki. Regarding buses, there are many more Hong Kong and Singapore buses than the 3000 or so Victorian Myki bus fleet, but this is partially balanced by the Melbourne tram fleet (very different from the gated single entry Hong Kong trams). The latest bus costing information I can find is from the University of California, Berkely, “Evaluating the Costs and Benefits of Transit Smart Cards” (UCLA_SM_CostBenefit_071308_Fin.pdf, Page 8, free download), which reports that buses will now cost between $35m - $47m per 1000 buses (from recent US bus fleet conversions). Imagine what Octopus or EZ-Link would cost today, with 6000 plus buses to be converted!
The more I delve into Myki build and running costs, the more they appear to be about comparable with the larger overseas cities using smart cards. The time to build the systems also appears to be as could be expected.

Regards

Anthony Ang said...

Good on you if you think that Myki deserves a pat. There are two ways of looking at the same thing, whether a glass is half full or half empty. I prefer to benchmark against a better-rated system rather than a poorer one so that one can aspire to make improvements to the system. I am not in favour of contentment or even settling for the second best, less a much complained system.

Anonymous said...

Some more reading for you. Independent audits found that the myki tender process wasn't bungled or corrupted: http://www.theage.com.au/opinion/society-and-culture/in-defence-of-myki--my-dinner-invitations-be-damned-20100506-ugsp.html

Also, the myki contract allows for future e-purse options, they were put on hold as obviously the first priority is getting myki rolled out on public transport. Which keeps being delayed.

Anonymous said...

I really can't believe anyone other than a vested interest would advocate Myki. The original tender called for a 150ms response time. At best it is 600ms and in most cases around 1 second. Why then can't the Government cancel the contract? Are there skeletons buried in the department's closet that they do not want exhumed?

Ted, you said you would scrap Myki, now is the time to stand up to the "YES Minister" departments and show some leadership!

Anonymous said...

Ted never said he would scrap myki. A rather expensive review was conducted by Deloitte, however didn't give the Libs the answer they wanted, so now they are reviewing the review and dragging it out as long as possible.

Anonymous said...

I have seen some figures from a reliable source that indicate that less than 20% of transactions involve a Myki card. If it is slow with this level of transactions what would it be like if Metcard was turned off?

How can anyone advocate the retention of Myki given not only its technical failure but also its lack of acceptance by the public!

Anonymous said...

This project shows all about too many wrong people in public offices, there is no hope to get it right with the same people in the office.

woowoowoo said...

umm - I wonder is the anonymous pro myki voice many people, or is just one dominating the conversation. and I wonder if it's the same person (@kiwiguy72) who jumps to the defence every time I criticise the system on twitter?

Anonymous said...

No issues to hear that myki is a waste of my taxpayers money. Purely from an analysts point of view your data doesn't stack up. You're comparing city based ticketing systems with a tiny geographic footprint to rolling out a state wide system where Victoria is the size of the UK. The analysis is great if you're comparing apples with apples. Victoria is the size of the UK so the analysis would be interesting to project the costs to roll out and operate a system to the whole of the UK for 10 years. It would be a fare wack of pounds :).

Post a Comment

Advertisements are not welcome and will not be published.