Australian State launches IBM probe IBM's dealings with the Australian state of Queensland will be the subject of an inquiry with the powers of a Royal Commission, after an SAP-based payroll system implemented in part by big blue failed so badly returning it to working order is expected to cost $AUD245m, with the whole project costing $AUD1.25bn. Queensland is …
This post has been deleted by its author
This post has been deleted by its author
I've worked on a public-private contract in the past, and can attest that this kind of flop is often the Government's fault. There are always a bunch of different departments and offices who all think that they're the only ones in charge of the project and don't speak to each other, and they all change requirements frequently, with no notice, on whim. Then when the project sinks there's usually an unspoken agreement that the private contractor will take the blame, in return for repeat business.
Random downvotes are part of life on el reg, I ignore all the votes these days :-)
You are largely correct on the changes of scope part but there is a special reality reserved just for private\public contracts where all normal laws and common sense is ignored. Having worked for a group of private companies on the railways with varying degrees of government control \ intervention I can say I'm not hugely shocked it went tits up. A change of scope or similar in the real world is normally a sensible request and easily accomodated. When you are working with politicians and civil servants all trying to get their own personal victories it's enough to make you cry. It very quickly became evident that they hadn't a clue what they wanted or needed. Politicians wanted quick and cheap wins, civil servants had a more logical long term plan but that went awry as they changed priorities and staff. One example (by this point we were on cost plus and had given up the will to live o question anything), it was decided we needed safe cess walkways alongside the tracks which isn't hugely difficult but isn't quick either in some cases. You don't just turn up, dump gravel and nip off for a butty. You need temporary consents for access and often permanent consents if you don't have enough space and then grading and possible relocation of cables / pipes and the work has to be done alongside a working high speed rail line, frequently with overhead power. Basically a long term project. Somebody had an idea, why don't you start from the middle and work out towards stations. In some ways it made sense but anybody with experience of government contracts knows how its heading. Priorities changed, other budgets overran and last I saw it there were large, nice paths all over the place that couldn't be used as they never made it all the way to a station because work got stopped and diverted elsewhere. We aren't talking small money either. I left for pastures saner and they could have finished them at some point, but please don't assume that large government contracts are subject to the same realities as elsewhere.
"can attest that this kind of flop is often the Government's fault"
In this case IBM pushed ahead with the payroll system, despite it not being tested fully, to avoid having to make late payments on its contract.
So every other taxpayer in Queensland now has to pay for IBM's clusterf***. Hopefully Newman can get some of our cash back
I get US $1.35 Billion.
It's 1 state in Australia and it's population (if you believe Wikipedia) is c4.6 million
That's close to a US$300 a charge for every man women and child in the state FFS.
Icon expresses my surprise unless that includes all the over/under payments.
If that's the bill to fix the software alone then someone is having a very large pole up one of their orifices.
My thoughts exactly. I am sure the days of knocking up a payroll system in cobol over the course of a couple of months is long gone, but I would have been shocked to find it cost that much to begin with, let alone to repair it. If that was the cost of a national payroll system, including hardware, maintenance and support for say 10 years, maybe.
Are they sure they haven't screwed up some calculations? I've seen some pretty impressive 'mis'quotes over the years, usually due to a basic formula error. One poor sod ordered what they thought was 1 ream of paper \ 500 sheets, of course a large truck duly turned up with 500 reams.
Why ? About time to bring back COBOL and fixed prices with agreed stages and results with external measurements. Somewhere I have seen a Visual COBOL even, so even if someone cringes at the thought of server with PC/whatever platform GUI, it can be done, even if one has to deprogram Java victims into a real language.
Because it isn't 'cool' (apparently). Theres nothing wrong with cobol, I know a country council that even a few years back used cobol for it's benefits and payroll systems. It worked and worked well, year and and year out. It has probably been replaced already with something written in RoR. Of course Turbo Pascal would have been a far superior choice ;-) but they had a few cobol coders on staff so it got used. I doubt the cost ran into the billions.
Briefly put:
I have only ever had a quick glance at SAP and this is my opinion and *my* opinion only: SAP tries to be everything to everyone and ends up being a mess of configuration which are sometimes counter-productive. The fact you need so many Consultant to configure it for *any* site should ring alarm bells in any CTO's head.
Oh, and once you start down the dark path, forever will it dominate your destiny. Getting the data back out is even worse than shoe-horning it in.
YMMV, of course. I'm sure there have been easy, successful SAP installations out there. Somewhere.
I worked with SAP systems for five years; I think that the product design is out dated, processes are bureaucratic and not easy to learn; and the implementation process is generally very cumbersome. But if it is correctly configured and managed it does actually work.
Unfortunately, IBM have a record of promising the customer that they will manage and deliver a working system, but then buy in the cheapest people that they can find to work on the project (whilst charging the client an arm and a leg). These contractors often have limited experience or knowledge of SAP or its implementation; the result is inevitable.
As for the costs; I tried to find out what would be involved in that figure of $AU1.2 billion but there don't seem to be any hard figures. If it includes new hardware, a complete redesign, transfer of data, internal staff costs etc, then I would expect it to be around a tenth of that for the size of operation. It could be that they are asking for compensation for all the trouble caused; good luck on that!
SAP will probably stay at arms length from this (as they do with most other failures). The Queensland government has a contract with IBM so they have to sue them. If they try to get at SAP, their lawyers will simply show that the software works elesewhere and it is down to the implementation. IBM will probably try to off load blame onto the contractors, staff, managers; everyone they can (and yes I've been at the receiving end of something like this)
I feel sorry for the residents of Queensland; they are the ones that will end up paying for this cockup.
"As for the costs; I tried to find out what would be involved in that figure of $AU1.2 billion but there don't seem to be any hard figures."
Its $250m (ish) to fix it now, its cost us poor taxpayers $1.25b since it was introduced.
I know everyone's politicians are oxygen thieves, but you gotta admit in this case the Anna Bligh government did something really special.
That lot could have burnt water they were so incompetent. Thankfully they got their comeuppance from the voters. Next year its Gillard's turn to see just what the voters think of her "leadership" - it aint gonna be pretty, bring popcorn :-)
There is no real technical excuse for screwing this one up. I know some people hate SAP, but payroll is SAP bread and butter and has been a stable product for 15 years and SAP payroll consultants are two a dozen.
This sounds like a complete cock up at the requirements gathering stage and potentially a data migration cock up, project management and scope management the likely overall cause.
This KPMG report from may gives the breakdown of the 1.2 billion - http://delimiter.com.au/wp-content/uploads/2012/06/KPMG_audit.pdf
• $1,008.0m relates to payroll operations that has and will continue to ensure that QH staff are paid on a fortnightly basis, and
• $245.5m relates to fixing the key issues and undertaking a systems analysis to determine the longer term solution for the payroll system.
The majority of the key issues appear to be legacy operational and business process practices - fairly typical of government projects I'd say.
AFAIK raw SAP is like a swiss army knife (Keep in mind SAP coined the term "ERP").
BTW SAP got their start doing mainframe payroll and accounts systems capable of coping with multinational companies IE mult site/curency/country/legislation.
LIke most big packages it's got core native commands and the rest is built out of it's internal language called ABAP. It also seems to have lots of commands with lots of options (some potentially quite dangerous if you don't understand their implications).
However SAP have also released pre-configured versions for various industrial sectors. In which case it should be an extraction/translation/load problem from the existing system (for any significant sized system running for any length of time this is not trivial.
I'm not sure if state healthcare payroll is one of these versions.
SAP would say that they have studied an industrial sector, identified best practice strengths and configured a version of SAP that when implemented in your business will purge your system of its weaknesses.
So its a win if your company sees the world SAP's way or can be configured to.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
