Danger, Will Robinson: Commission of Audit wants Centrelink core IT systems outsourced


blog Yours truly hasn’t yet had the chance to comb through the recommendations contained in the Abbott administration’s Commission of Audit report released this afternoon; that will take the better part of a week. However one notable item which has already been picked up by technology media outlet iTNews this afternoon is that the report includes some rather … drastic recommendations for Centrelink’s extremely complex and high maintenance core IT systems. The site reports (we really recommend you click here for the full article, as it contains a great amount more detail):

“Following Treasurer Joe Hockey’s comments about the need to modernise Centrelink’s 30 year-old Model 204 database management system, the CoA has endorsed the upgrade and set the expected cost between $1.2 billion and $1.5 billion. But it went even further, suggesting the Government test whether all or part of the DHS’ payments system (ISIS) … could be outsourced”

To say this statement is controversial is putting it mildly. Over the past decade, Centrelink’s IT systems have, at the very least, been stable, functional and secure, even if the welfare agency hasn’t at times performed at an ideal level in terms of how fast it can deal with the enquiries of welfare recipients. Outsourcing this kind of platform to the private sector is risky in the extreme.

Would a major Australian bank outsource its core banking platform? No. Would the ATO outsource its core taxation tracking systems? No. And yet that is what the Commission of Audit is recommending. Disturbing — to say the least. One suspects former Centrelink CIO John Wadeson — who kept Centrelink’s IT ship-shape for almost a decade — would be more than a little surprised by this one.

I’ll have more to say about this in coming days as I take a closer look at the recommendations in the report.


  1. The private sector can deliver an IT system much more effectively and at a cheaper cost than the government could. The government is not in the IT industry, it’s their job to provide a service, not be involved in IT. There’s a reason why major companies everywhere are adopting a cloud approach to their IT systems, it’s simply not a core part of their business.

    The mentality of government should be how can we provide the service, how can business help us do it.

      • Your concern is very much a fear mongering exercise.

        Just because banks choose to house all their IT internally, doesn’t mean that might be more effectively run elsewhere with similar security. There are already Platform as a Service and Software as a Service providers out there performing such a service.

        • Maybe the private sector could create this system cheaper, but it would be full of bugs, lots of shortcuts taken, cost and time over-runs, scope creep, and in the end would be an inferior system, constantly requiring “ongoing” maintenance just to keep it running, and full of security holes (because that isn’t important, in a cost cutting exercise).

          This sort of system should never be trusted to the “lowest bidder”, there’s far too much involved that you can’t place a (cheap) dollar value on.

        • Yes David it could be outsourced to the private sector who could do it better, potentially using Cloud resources (PAAS, IAAS, etc)

          But it wont be better. Why? Because the government will tender it to a big consultancy who will partner up with a big vendor, who will flog their products not because it’s the best fit for the job, but because that’s how the vendor makes money. The outcome will be a mess of buck passing vendor(s), horribly integrated systems and will cost more money than just leaving it where it is. Want proof? Look at Ultranet in Victoria.

      • Interesting article and i’m seeing both David’s and Renai’s POV’s here.

        David, I completely agree with you – it is absolutely true that you could outsource operations at a cheaper rate while still ensuring security, compliance and provide better outcomes for the customer. There’s a reason the top 500 companies in the world outsource extensively – it’s because it makes economic and operational sense to do so.

        Renai, I completely agree with you, though probably with some extension to your comments being that the reason it (almost certainly) won’t work is that outsourcing is a fundamentally different way of operating – your governance and contract management skills need to be world class AND you need to have a deep level of executive sponsorship and engagement.

        I won’t comment on the first components – I haven’t been inside government for a number of years, however let’s be optimistic and say that both governance and contract management skills are excellent.

        Then we move onto executive sponsorship and engagement.

        Unfortunately politicians and political parties do neither of these well if at all – agencies are expected to make changes on whims, implement poorly crafted and considered decisions which inevitably means that the ‘system’ that has been developed (by system I mean people, process and technology together) is bastardised to the nth degree leaving us with a hodgepodge of nightmare operational stacks that realistically and appropriately cost lots of money to manage

        Add to that a healthy dose of knowledge of outsourcers that government departments will inevitably behave like this (because they really have little choice) and they set themselves up for long, tedious and endlessly changing sets of deliverables to delivery – and they extract a premium for it.

        Sad really, all that wasted opportunity.


    • show me one “private enterprise” outsourced IT project that a single government department has had done that hasn’t gone years overdue and billions overbudget.

    • > it’s simply not a core part of their business

      I would say that efficient and competent IT is absolutely a core part of their business, if not *the* core part. I find it utterly baffling me that anyone would think otherwise. What is centrelink without the database? A bunch of public servants sitting in offices twiddling their thumbs!

      • Look at what the customers want, they want the service. How the government gets there should be the most effective use of taxpayer funds. If it’s external providers, a cloud platform, they should do it.

        All the backend systems and staff that manage it to get to that result shouldn’t be the government’s responsibility.

        • Yes, lock them into using an external service for their IT. I am sure they will keep costs very low and maximise their profit. They will also have the government over a barrel as it will be extremely difficult to then just shift the service to another provider, so they can after a few years maximise profits by charging like wounded bulls.

          • They’ve haven’t even got to the bit where they accept the recommendation, let alone are “locked in” to anything.

            Everyone should be pleased if we are getting the service for a cheaper cost to the taxpayer. I’m not too worried if the company is making a profit in the process. It will just go to show how inefficient the government was in delivering it.

          • That’s your argument as to why they won’t price gouge? That they haven’t done it yet? I am not saying “make a profit”, of course they need to. I said price gouge, cost more than it does now because the government doesn’t have a choice but to pay them.

        • David,

          You are barking up the wrong tree and the more you do the more you are looking like an idiot. The point being made by Renai and others is that government has a horrific track record in specifying and project managing such endeavours. It is often said that one has to know your business more to manage an outsource deal than you would otherwise need to if you kept it in house. Cuts to government often hit this capability more than any other. So outsourcing after large slash and burns often result in more project failures not less – the only thing is that the details are buried in commercial in confidence contracts and the failures a hidden from public view. I would suggest the more you run the lines you have been in this post the more you look a fool – noting the significant outcry made recently re ehealth and NBN on this site.

          • It’s nice that we’ve gotten to the part of the debate where we insult each other isn’t it?

            Nowhere in his article does he mention past failures of IT outsourcing. We are in furious agreement that people are required to be accountable on the government side too, not discounting that. Just not a whole department.

            The mentality of business when it’s not their core function is to get someone in to deliver the function for them. The government should start acting this way too.

          • sorry didnt mean to insult – however you are insulting my intelligence and the intelligence of the readers of this site when you are running with such a narrow argument I was simply calling you on that.

          • (i’m hoping that Renai deletes these last few comments, including this one as they’re really not adding any value)

            For the record he’s certainly not insulting my intelligence – please be cautious when you as an individual feel insulted does not equal ‘the readers of this site’

            Oh, and if you want my POV you can scroll up – hopefully it’s a little more balanced than your own narrow minded position on the topic (which ironically I tend to agree with)

            David is right though, a mindset change in government when it comes to repeatable service delivery (which is predominantly what this is) would come at benefit for the taxpayer and therefore the broader economy.


          • “It’s nice that we’ve gotten to the part of the debate where we insult each other isn’t it?”

            In the interests of accurate reportage, I should observe that you previously wrote:

            “Your concern is very much a fear mongering exercise ” which, at the very least, implies that Renai is being misleading, if not worse.

            I think the debate would be helped if you, and others here, would step back from your dogmatic “government should/should not outsource everything non-core”views, and try to see that, at least sometimes, there are valid views other than your own.

    • “The private sector can deliver an IT system much more effectively and at a cheaper cost than the government could.”

      Nice marketing spiel.

      For a private company to provide the same service, and make a profit to return to shareholders, will, by it’s very nature, mean they actually cannot do it cheaper…whether the system implemented is more efficient depends entirely on the quality of those doing the implementation regardless of that implementation being public or private.

    • “The private sector can deliver an IT system much more effectively and at a cheaper cost than the government could. ”

      Look – the Tea Party is here.

    • David, you must have no idea on complex and complicated IT systems, such as our social security system. Would you chance your personal and private information to Microsoft or to any other IT company.

      You have no idea on how complicated our Social Security legislation is do you? We have the most complex and complicated legislation in the world.

      So, you reckon we outsource our tax systems, defence systems and our Customs/Immigration and border protection systems as well?

      Going on your misguided logic, governments should revert to pen and paper based systems seeing Governments should not be in IT!

      Private does not mean better or cheaper. Just ask Victorians on their electricity bills and the regular ‘brown outs’.

      Also, when was the last time the Australian government or any ‘Western’ government for that fact, went bust? Never. How many private companies, big and small have gone bust over the past 20 years?

    • David, I am in agreement. Business COULD do it better. Can you name one major IT project in Australia where this has successfully occurred though? No scope creep? No cost/time over-runs? No zero-day big bugs? And which services would not hold the system to hostage when they are able (not explicitly pay us or else, but more, give us xxx and we can do that)?

      The problem is govt changes too often for long range projects to be PROPERLY done. And when it changes, scope, ideas etc change and the new govt holds everything up “Look what went wrong?” when they should be looking how things can be fixed better. Would probably be better long run if govt made IT part of the core platform and then cut it loose like the RBA somewhat.

  2. Seriously David? Big IT outsources make money by getting bums on seats, NOT by implementing projects efficiently. This isn’t a cloud vs on premise debate. IBM and Accenture would love to get their hands on that contract and milk taxpayers for all they’re worth.

    • Spot on. They have a ‘body shop’ model. in fact they have every incentive to string out the project and make it so they are always needed. We should hold these people commercially accountable for delivery. If they fail they pay.

  3. The private sector does not work in the same fashion as the public sector; there are often ideological and entrenched decision making systems that can lead to endemic delays, cost overruns and eventual failures to deliver agree outcomes.

    History is replete with many an example.

    That is not to say it cannot be done, rather it is frankly never as simple or as academic as people would like to suggest.

    It’s not a case of whether a government can outsource a departments’ infrastructure and application needs, it’s that without good governance it’s an absolute recipe for disaster.

    Every time such a report is commissioned, the same basic recommendations are delivered; it’s an expedient way to considerably reduce system and workforce costs.

  4. The article is incorrect. Both Westpac and CBA outsource their core banking platforms and the vast majority of their IT. However anyone who thinks it was done for cost savings is a fool. It’s more expensive and a poorer service. It was done in Westpac’s case as both a power play (the business perceived IT as having too much power) and to reduce union power. I assume CBA had similar drivers (I wasn’t there during that period).

    I’ve worked for both IT outsourcers and outsourced companies, as well as done outsourcing transition. I am very confident that outsourcing Centrelink’s IT will end in tears.

  5. There are 2 distinct (potential) aspects of this proposed outsourcing: The INFRASTRUCTURE and the APPLICATION(s).

    “Spearheading the service delivery reform agenda will be funding to the tune of $373.6 million over four years to integrate the ICT infrastructure of Medicare Australia, Centrelink and the Child Support Agency. Some $295.4 million of this, including $205.3 million in capital, will be met from within the existing resources of the Department of Human Services.”

    These 3 agencies (plus the Dept of Veterans” Affairs) now run on the DHS INFRASTRUCTURE. Medicare, Child Support and DVA were insourced with significant effort: it’s outside this arena to discuss the financial and operational benefits or the lessons learned.

    Each AGENCY has their own APPLICATIONS which may or may not be suitable for outsourced development. As per other comments, there have been some painful experiences with this in other government entities.

    Perhaps we should be specific about which aspect we are discussing.

  6. It would be a complex system to replicate and migrate to. Cutting code using humans is ridiculously expensive (even with cheap Indians, no sorry, cheap Cambodians, or whoever they are exploiting now) and prone to massive failure. We need to get rid of this ridiculous approach to building complex applications.

    There are a number of companies that have developed toolsets that enable complete highly robust applications to be generated automatically from a highly detailed ‘engineering design’.

    The big boys want to continue to use the current failed body shop and cottage industry application development approach as it locks in massive costs (big margins) and it perpetuates the myth that this is only way of doing it.

    Centrelink should break the project down into three stages – the first to solicit a global response to an RFI as to all of the methods of building the replacement and put machine generated approaches up against the body shop clowns. This may result in a shortlist and part build of a component of the solution to test the approach. The next stage after selecting the ‘approach’ would be the build the application and finally the commissioned application may be handed back to Centrelink and run by a specialist hosting service.

    PS – if anyone whinges about you can’t build applications without human coders get real. I have seen it done. The applications are very robust and all you maintain is a single specification file and not millions of line of dubious code (written offshore). A change in the application is a change in the design and regeneration over a number of hours.

    • Hmmm, interesting.

      At the end of the day it will be a SAP build. I’d bank on that.

      • I am convinced given what I have seen with some of these new technologies government could build its own Cloud based cross government back office service for a fraction of the money they throw away each year trying to make ill designed and flawed systems try and work. How does trying to shoehorn a manufacturing system help government? Much of the functionality is irrelevant and the rest inaccessible due to the limited architecture.

        In fact you could build a single cloud service for all governments with some vision and the elimination of egos.

        The Scottish Government has saved over a Billion Pounds (audited) by using a single cloud service for all of its public sector procurement. We don’t need to buy tens and tens of poorly designed and costly systems to support what is a single generic function.

  7. I worked for large IT outsources in both commercial and government. They were invariably a failure. If you outsource then you have to maintain a group of people to ensure that you are getting your moneys worth. You might as well do it yourself. The outsource people are not working to enhance your business they are only working to enhance the return of the outsource.

    • I must admit, I know of very few successful outsources. Even using 3rd party software solutions can be a big gamble. You lose control and if there is a problem there is very little you can do about it if that 3rd party doesn’t respond quickly to the issue.

  8. OMG large visions of IBM and the Queensland Health payroll debacle for me. i dont remember what the original budgeted figure was supposed to be but i do remember the final figure was many many times larger than that.

    there were massive project management issues there; and that was just payroll, not all the various strands that Centrelink are supposed to look after. which is to say, that was a much ‘simpler’ job than this, on the face of it. does anyone (other than David) seriously think this will end well?

    • Was it not the case that when the project continued to fail, everything was so complicated and confusing that the government could not pursue its remedies in court to recover damages?

  9. In principle there is no difference between a group of developers in Centrelink and a group in IBM or whoever. If managed well with good specifications the result should be the same. It will cost the same number of hours of work. The main difference is that IBM or whoever will be desperate to make a profit. So, there is no logic to outsourcing if the aim is the save money. Outsourcing only makes sense if there is some set of skills and knowledge held by the outsourcing provider that Centrelink do not have. I can’t believe that the IT people in Centrelink are so bad that they can’t match any prospective private sector provider in skills, knowledge and experience. I do strongly believe that whether the task is done in-house or outsourced it will fail if the management don’t understand the needs, their own specifications and how IT works. They will then manage it to a teary death.

    The CoA proposal smacks of the attitude that caused the mass of IT outsourcing in 1998-2005. Remember how badly that went ! In the Department of Finance there was the promise that it would save 45%. Yes, it was cheaper (for a while) but so was the service. Products cost 30% more than retail off-the-shelf. The provider had more, slower, bureaucratic processes than the department. Nothing got done quickly when needed. Nothing was done unless it was specified precisely in the contract. The contract didn’t cover data and knowledge management. The long term result was the massive loss of data and data analysis capability. Data was stored without thought to access – some data permissions were set which prevented all access – only the IT security staff could access it if they knew it was there. A capability review in 2012 recently found that those problems persist.

    The CoA comments elsewhere are spot-on. Look at this (p234 of Phase One): “There are three key changes needed: ….. a change in the mindset of government agencies from collecting data for filing to collecting data for use;”.

    So true, and current practices are the result of agency management thinking that when IT is outsourced they don’t have to think about how it should be used. Outsourcing will not solve the attitude and cultural problems. It will enable the poor managers to outsource responsibility for knowing their business and getting it right.

    I might sound like a bitter IT professional but, I’m not. I was an IT user frustrated by years of not achieving what was possible because of the outsourcing arrangement and by the increased costs in staff time and effort to do simple things that were not done or were done poorly by the provider.

    A last thought. The ATO did outsource development of eTax (to DWS). I don’t know who was responsible but, everyone can see how poor the application/UI design is and can assess how poor the value for money was in the ATO paying $5.2m for the Mac OS X version (created by staff who didn’t know basic OS X development). DWS have benefited from extreme lock-in. How can that be avoided by Centrelink if they have to outsource the ISIS re-development ?

    • “How can that be avoided by Centrelink if they have to outsource the ISIS re-development?”

      I guess there are theoretical ways to do this. If the redevelopment is all SAP based as expected, then they build in-house expertise in SAP (DHS would have some already anyway) and then they take over the maintenance of the developed product.

      In practice I don’t know how well this would work out.

      Unfortunately this whole debate is basically an ideological one. There is the irrational side which states that small government is always the end goal and that private efforts are always superior. Then there is the rational side which states that private and public efforts are by their nature different and that any attempt to compare them apples to apples is nonsense.

      I have to say, as a member of the public, that I would not be happy for any of my DHS information to be stored or handled by any third party (except as required by mail-handling contractors etc).

      • Yes, the outsourcing of a decade ago was ideologically driven and the same people still may be pushing that barrow.

        Have limited knowledge about SAP but, I believe it is primarily intended for private organisations and their accounting requirements. I have seen material about SAP’s use for social security systems in other countries. But, it does look like manipulation of an existing private sector model for a totally different purpose. The ISIS replacement will not be an accounting, CRM or ERP system. I fear it will be a square peg in the round hole kind of development. It can be done but, it’s ugly and costly.

        It reminds me of the CBMS redevelopment underway in Finance. They chose SAP which seemed reasonable as the budget numbers are pulled together from accounting transactions. Yet, they’ve had to build at least one custom module and alter most of the rest. It will cost quite a lot which, makes the economics of COTS less clear.

        Anyway, SAP seems to be the most popular accounting pack in the APS so, with the staff cuts coming along there should be a lot of experienced people available to work up ISIS.

        Contrary to my earlier comment, maybe lock-in might be avoided if Centrelink take a true project management role rather than outsourcing it all ?

        • Just to attenuate my comment on CBMS, there is this advice on the Finance web site:

          “Of the 290 [CBMS] Functional Requirements, 77 require some custom development or enhancement. These developments have been built in alignment with SAP Custom Development Programme (CDP). Some of these developments will be put back into the SAP ‘Core Product’ and offered to future clients.”

          I wonder how good the cost equation would have been without the role up of custom developments back into SAP for resale ?

  10. Yep, time for Haliburton to step in.

    Maybe they will add some of their special $1,000 toilet seats to each contract.

    Now that is REAL private enterprise.

  11. Completely outsource the system. I understand most enterprise tender the system design and support that is quite a standard. But to remove every single in house engineer and support person to make sure the vendor is delivering as promised what are they thinking. That’s like outsourcing the driving of your car to India. The problem of that way is loss of control over how your business operates. A business’s IT systems determines how a business operates. Reminds me of the family guy episode with the TNG cast with Patrick Stewart slamming Whil Wheaton head into the window saying “You’ll get nothing and like it”. That’s how contractors operate.

Comments are closed.