IBM and Accenture are sniping at one another in public over just who should take the blame - and the fall - for the $AUD1bn blowout of a project to provide the Australian State of Queensland's Department of Health with a new payroll system. The project kicked off in 2007 with a budget of just over $6m. It's now expected to cost …
who blew up the money???
after reading the article 3 times (I admit I'm not a native English speaker) - who blew up the money? IBM or Accenture?
"unsuccessful bidder Accenture"
So, that would be IBM's fault.
But, over a billion dollar for a payroll system? Words fail me
Re: "unsuccessful bidder Accenture"
"But, over a billion dollar for a payroll system? Words fail me"
Gets better when you consider that Queensland Health have around 78,000 emplooyees. So the IT system is supposed to cost AUD 16,000 per employee.
Personally I'd only expect a big bucket of fail if I hired a huge global mega corp to implement a payroll system, aided by a fat cat management consultancy. Did the berks at Queensland Health think when they let this contract: that a couple of multi billion dollar, high margin global corporations were really looking to provide a low cost, good value payroll solution for a regional public sector health provider in the middle of nowhere?
Re: "Ledswinger: unsuccessful bidder Accenture"
I protest; vigorously: Queensland is NOT in the middle of nowhere; it's well to the right of it, both geographically and politically. And I'll also endorse the Health System there: good care in Princess Alexandra hospital 55 years ago means I'm still well, physically......
On the one hand, I'm glad it's just not the British public sector that can screw up IT so badly.
On the other hand, my commiserations to the Queensland tax payers for having to ultimately shoulder the burden of this almighty screw-up.
Given the two parties involved, it'll be very hard to work out who is the more incompetent of the pair and therefore who should take the lion's share of the blame.
Does anyone know of a situation where the presence of outside consultants ever made things go better?
There are three parties to this dispute, and I am sure that at least IBM will argue that project costs exploded because the requirements the customer wanted implementing were not the requirements it tendered for. That's how IBM (and others) make their money: the knowledge that the customer will want something else.
Fixed price == fixed spec.
> That's how IBM (and others) make their money
To be fair, that's how everyone makes their money when tendering for fixed-price work. The basic problem is two-fold. First, the customer never really knows what they want - nor have worked out the conflicts between what they are asking for and what they want to do (that would take far too long). Second is that no project exists in isolation: so as (development) time moves on, the requirements: regulatory, business and technical and systems that the proposed project is meant to interface with will change, and therefore the fixed-price spec. will have to be changed to accommodate these.
The basic issue is that F.P. tendering is a flawed process. It was originally meant to put a limit on the huge, open-ended overruns of government projects based on cost-plus (where the company who won the contract passed on their costs and an agreed "profit" margin) principles.
It is possible, if unlikely, that any contractor could deliver what the spec asked for, for the agreed fixed price. However, it's almost certain that this wouldn't perform the required tasks for the reasons above. The result would then be a system, delivered to specificaton, that had to be either tossed and re-done or battered into shape with a secondary project - that would have all the organisational drawbacks that the original one had. And since only one tenderer would have sufficient knowledge, skills and background to complete the work, the tendering process would fail as there would only be a single viable applicant.
So until clients can work out what the hell they are actually trying to do (a rare breed, indeed), keep all the interfaces to the proposed system frozen until it's final delivery AND making sure there are no omissions, errors, inconsistencies or ambiguity in the design documents it looks like we're stuck with the present system - or something like it. Still: at least it keeps the lawyers happy.
Yes, change orders are where the big margins are always made. A good bid leaves as little money on the table as possible by guesstimating how many unforeseen, unspecified future requirements will recapture higher margins after the low balled deal is already done. Regardless of the industry, this game is a common one. However, when there's collusion in the process (whether its between suppliers or even between procurement personnel and suppliers), the opportunity for massively fraudulent gains are significant.
Re: Fixed price == fixed spec.
Many (many) years ago I worked on projects where there was an initial 'cost plus' phase leading to a detailed spec that formed the basis for a fixed price implementation. At least the initial phase had to be a joint process with appropriate customer staff. Changes to the FP phase had to be signed off in blood (almost) after a separate negotiation each time over allocation of any extra costs.
I know, a ludicrously old fashioned idea, from the days when we had software companies rather than management consultancies.
Even AU$6m sounds a pretty healthy budget given that you can buy payroll software off the shelf for around $99. But a 200x overrun - that's pushing $300 a head for the entire Queensland populace, men women and children. What were they doing, buying everyone there a new iPad to access the system?
"But a 200x overrun "
Probably the cost of running it on IBM OSs / hardware instead of on Wintel....
You do realise that things like payroll and other batch processing systems are the natural choice for a mainframe system?
This is just another example of governments being bad at managing IT projects.
it's SAP, not an Excel worksheet with formulas in two columns
Both fail equally hard.
I guessed it must be SAP. In which case I withdraw my comment, they got away relatively lightly ...
So 7 *decades* after people first started running payroll and accounts apps
They are still screwing them up.
Seriously double entry book keeping was developed around 1500 in Venice.
It's not exactly new technology is it?
Re: So 7 *decades* after people first started running payroll and accounts apps
The *Requirements* have become a bit baroque though.
And if you interface with State, all the rules that politicians can come up with will cover you in complexity vomit.
It would have been cheaper...
...to hire a bunch of bookkeepers to do it manually.
Similar thing happened in West Oz
Both instances down to multi-national IT firms ripping off taxpayers.
I worked on an outsourced project once. It was funny that we were instructed not to fix any bugs during the warranty period, just do data fixes. Some of the users figured this out and learnt to withhold some test cases so they could prove 'fix' did not work. On the day that they secured the maintenance contract, then we could fix code bugs. Anyway they got away with it. The 'blue- chip' consultancy charged top dollar for all their people, even though most were recent graduates paid modest salaries.
Problem is that public sector prefer to shift the risk onto outsourced companies who will do so at a price.
How hard can it be? Knowing how these things go, I wouldn't be too surprised if the person in charge of the requirements went on maternity leave, leaving behind a replacement with an entirely different concept of what a payroll is*.
* Yes, that kind of thing does happen, in my experience, at least when dealing with government and healthcare.
Don't sue IBM...
I thought by now it would be common knowledge that the real reason you don't risk litigation, as a government entity, with IBM is because that company by now might as well be a "law firm with an attached tech department".
They have lawyers orders of magnitude better than the piddly third-stringers a state government could afford. It'd be a massacre and IBM would probably end up *owning* the entire population of Queensland as indentured serfs.
From $6,000,000 to $1,250,000,000? I'm dumbfounded. If they actually pay that bill, I'd be even more dumbfounded.
- How much did NSA pay to put a backdoor in RSA crypto? Try $10m – report
- Geek's Guide to Britain Mosquitoes, Comets and Vampires: The de Havilland Museum
- Google fined by Spanish data watchdog over audacious privacy tweak
- Fanbois, prepare to lose your sh*t as BRUSSELS KILLS IPHONE dock