Disaster ahead? NSW Govt unveils massive SaaS ERP consolidation


news The New South Wales State Government has unveiled plans for a massive technology-led project to consolidate a number of different enterprise resource planning systems onto just two new platforms, in a style of project which has historically led to cost blow-outs and extended project delays for similar initiatives accross Australian State Governments.

The state is currently engaged in consolidating shared services platforms in different ‘clusters’ across its operations — for example, in the education or health areas. One of the largest clusters which it plans to consolidate back office services for is the ‘justice’ cluster, which includes not only mega-departments such as NSW Police but also Fire and Rescue NSW, the Department of the Attorney-General and Justice, Legal Aid NSW and more.

All up the Justice cluster within the NSW Government represents some 40,000 employees and over 90,000 volunteers. Many of those staff are represented by unions and provide critical services to the NSW community.

In one of several tender documents released this month, the state revealed that one of the major bodies of technology-related work under its widespread shared services consolidation program was to be a huge ERP consolidation program. Currently, the Justice cluster has a variety of ERP platforms in use, including Mincom, SUN financials and Aurion HR, as well as two SAP ECC 6.0 implementations.

According to the documents, the state is planning to consolidate those disparate systems down to just two new ERP systems which wlil provide services to all of the clusters.

“The [Justice Shared Corporate Services] Reform Business Case (Business Case) describes an outcome in which there is a standard ERP solution for the cluster based on SAP and involving no more than two physical instances. It is envisaged that agencies will transition onto the new Justice Cluster ERP system in stages,” the tender documents state. “The Business Case assumed that the Cluster ERP solution would be built up from the existing mature SAP ECC6.0 implementations in the cluster, reconfigured to meet new business processes and service delivery requirements.”

The NSW Government has placed a number of conditions on its future ERP platforms. They must be based on SAP’s ECC 6.0 platform; they must support HR conditions in the NSW public sector; they must be hosted in a NSW Government datacentre; they must support business processes that are aligned to “whole of government process standards for HR, finance and procurement”; and they must be able to integrate “seamlessly” with other on-premise and cloud solutions used by NSW (both SAP and non-SAP).

A wide range of functionality is required to be built into the new systems, ranging from finance (including purchasing, general ledge, inventory, contract and funds management aspects) to human resources (including personnel administration, payroll, scheduling and rostering, workers’ compensation and so on) to asset management, employee and manager self-service, reporting and more, making the planned systems much more than just payroll and employee management rollouts. They will be comprehensive as opposed to single-faceted.

There are also other complicating factors which will affect the new deployments. For example, beyond the wider consolidation project, the tender documents revealed that the Justice cluster has an “immediate need” to transition between 600 to 900 staff onto a new platform for HR and payroll functionality, with a go-live date listed of 30 June 2014 — in just eight months’ time.

Historically, this kind of payroll systems consolidation project is precisely the style of project which has a high risk factor for suffering timeframe and cost blowouts within Australian state governments.

A similar payroll consolidation program at Queensland Health, for example, also using SAP’s software, went bad to the tune of $1.2 billion and was the subject of numerous audits and investigations that eventually resulted in prime contractor IBM being blacklisted from bidding for any contracts with the state. A number of other major ERP consolidation projects have gone off the rails (although not quite to the same extent as the Queensland Health project) in most other states in Australia.

In NSW itself, a landmark report into the management of the NSW Public Sector commissioned by the state’s new Coalition Government and published in August 2012 described how dozens of overlapping and competing systems and services providers had created “chaos” when it came to the state’s current IT shared services paradigm.

Some states are attempting to mitigate some of the issues by shifting their own internal business processes along more standardised global lines, and then adopting standard software as a service (SaaS)-based platforms, in an effort to avoid the complex and risky customisation work that has led on-premises projects off the tracks.

In particular, the NSW Department of Trade & Investment has claimed initial success in just such a project using SAP’s Business ByDesign SaaS package, claiming in August that so far the project has been delivered “on time and on budget”, but with a large chunk of the work still to go. However, SAP has also made modifications to its future strategy for Business ByDesign, leading the department to query the German vendor on the effort.

In its own tender documents, the Justice cluster represented by the Department of the Attorney-General and Justice noted it, too, would attempt to go down the cloud computing path with its new SAP implementation: “Recent interest by NSW Government in cloud and managed service offerings has, however, introduced a potentially new opportunity for the Justice Cluster and a decision has been made to explore the maturity and applicability of SaaS offerings specifically in the SAP ECC6.0 domain.”

I’m going to go out on a limb here and say this is a very risky project, and that the New South Wales Government should consider looking at more piecemeal initiatives before bringing everything (including 40,000 staff) onto just two SAP ECC 6.0 instances. This is precisely the kind of project we’ve been seeing experience massive problems in every state in Australia over the past half-decade.


  1. If done well there is potential for massive savings in such consolidation. Especially around payroll and HR services which by their very nature are quite codified. The Queensland Health project is a great example of how not to do a SAP project. Datacom has been paying most of Victoria Health for over a decade on a SAP platform on a single platform costing a couple of orders of magnitide less than where QLD has ended up.

    • Hi Mark

      did I talk to you on your stand at SSON in Melb in May?
      I was fascinated to understand how you differentaited Datacom from Northgate, given you were both based [I think] on SAP HRMIS.
      If I’ve got the right man, I’d be keen to re-engage with you.

      regs Noel

      • Hi Noel

        Yes, quite possibly. I was at SSON. Give me a call 0413 732536.

        Regards Mark

  2. The element always assumed is that, in this case, the NSW Govt departments have the capability to define, manage and deliver this scale of project. They, along with every other government in the world, don’t. So, if you try to do something you are not equipped to do, you will fail. The bigger the project, the greater the fall. NSW voters need to watch their wallets, this is going to be a costly exercise.

  3. “… a massive technology-led project…..”

    Says it all.

    It;’s the tail wagging the dog and most such projects fail, because they will not have worked through all the business and information problems they will have created.

    It’s called hubris.

  4. Hah! This will be interesting …

    Historically the prognosis for such a project would be diplomatically described as “poor” … and more prosaically as “@#$^%&%%$#!! diabolical”.

    In an alternate universe, however, we are learning from cloud services that it is perfectly possible to for many diverse customers to share a single SaaS application quite happily. The ingredients for success are: (1) that the solution already exists as-a-service and is proven to work. (2) That the solution was engineered from the outset to be a multi-tenant shared service. (3) That the solution is sufficiently configurable to be able to satisfy diverse needs without creating a customisation hairball. (4) That customers choose the service more-or-less voluntarily (i.e. having decided for themselves that it is good). (5) That the organisation providing the service is serious about providing a world-class shared service and has invested in the structures, processes, people and technology required to do this. (6) That the service provider has a sustainable basis for funding iterative service innovation. (7) That customers understand the value proposition inherent in behaving as intelligent consumers of an arms-length shared service.

    The cloud services model works because these ingredients are either there from the get-go (“cloudy is as cloudy does”) or rapidly evolve as customers adapt their behaviors to align with the cloud services model (“It takes two to tango”).

    The problem for a project to consolidate existing ERP systems across multiple agencies is that none of these ingredients are present at the get-go and they seldom evolve over time. Empirical evidence suggests that such a project is simply a triumph of hope over experience.

    The better, lower risk, model is to find (or incentive the building of) the best possible enterprise-grade as-a-service starting point and to create incentives for agencies to decide for themselves that they will migrate into the new as-a-service model. Move into a nice new apartment building rather than try to renovate the tired old building … with its rotten stumps, dodgy wiring and hidden soil contamination.

    Old-school ICT strategies that seek to consolidate, rationalise, and standardise are now recognised as wasteful from many perspectives, but most particularly because of their impact on organisational learning in agencies during periods of technology disruption. These projects take many years to fail, all the while acting as a brake on ICT-enabled transformation projects within agencies. As the pace of technology change quickens, the risk profile of large scale renovation projects increases – they take too long, bog everyone down and are too inflexible. Their opportunity cost is measured in the inactivity of agencies while they wait on the platform for a train that never arrives.

    Looking forward, the key change required is for agencies to accelerate the cycle time of organisational learning by buying services that work rather than attempting to pull off a low budget renovation of existing structures, processes, people, infrastructure and systems that don’t.

    • Very insightful. Please keep up these comments, in fact start writing for delimiter officially!

    • hmmm…
      I know a guy who is renovating his house, I wonder if he would agree with the proposition that it is better to ‘…Move into a nice new apartment building rather than try to renovate the tired old building … with its rotten stumps, dodgy wiring and hidden soil contamination…’?

      Experience to date with all things cloudy in government shows that the cloud implementations that are working are those where the thing being delivered is essentially a green-fields capability, not where legacy is being overtaken by the magic promised in the cloud.

      So, if you are in government and are going to try to move things to the cloud, pick those things that you haven’t already (typically ‘over’) invested in, at least until some experience is gained in how to make the cloud work in the government context.

      That said, I agree, this will not work as it is being sold.
      I give it three years, an Ombudsman or Auditor review, or a change of government, whichever comes first.

      If the ‘…two new ERP systems which will provide services to all of the clusters…’ were going to be provided/serviced by an external provider who understood ‘customer’ and was subject to market forces [read: competition], then the plan might have a chance. As it is being proposed, it is just a slight variation on the shared services model that has failed everywhere else in Australia. And by ‘fail’, I mean it has failed in EVERY jurisdiction, and at great cost to the public. C’mon people, learn from the past!

      “Those who cannot remember the past are condemned to repeat it”
      – Santayana

    • Steve,

      What evidence do you have that government has learnt anything? Or indeed, that government has the capacity to learn – ie the much talked of learning organisation but little seen. Large global brand name consultancies make their living on the business model that they dont.

      Whats the driver for such changes? budget cuts? Lack of money? Don’t think so as there is always money for big projects because big end of town drives the need for them!!!

      At an executive level in Government and like many Board rooms it begins and ends with the language and culture of we have appointed the person(often male) to take responsiblity for IT and systems implementation and that person is the CIO. Hence executive has in their eyes discharged its responsiblity. And or then the CIO appoints a multitude of stakeholder committees /Program /Project Boards that diffray the responsilbity and disperse accountability such that no one person is held responsilble.

      Thats how it works. Can you provide the evidence that there is a substantial change not just your suggestion that there needs to be organisational learning.

      Chris Argyris’ work is seminal in this space – I recall buried deep in one of his books the concession that for real organisational learning to occur one has to in large measure banish “politics” or the politicisation of core undiscussable issues. Perhaps that why we hear so much and never really see the thing called the learning organisation.

      Love to share some war stories even about personal experiences on the very projects that are mentioned but that would be career limiting. Guess thats the point!!!

      • Hmmm … yes sure … it is not easy. There are pockets of hope emerging however. Vic DSDBI’s use of Salesforce for its Gobal Engagement Management System and for Grants Administration. Federal DHS’s ExpressPlus apps. NSW Trade & Investment’s use of SAP’s SaaS ERP platform Business By Design as the kernal of a new-age shared service. The key thing is to ponder how things could be better and to look for harbingers of the future. It is all about leadership and vision … but yes … it is difficult.

    • Are we certain that internal system consolidation projects are predestined for failure, and that *aaS solutions are the way of the future for some government systems?

      QLD Health’s payroll failure (delightful case study as it may be) pointed to a failure of process, rather than a misconceived goal of a consolidated system. Ignorance of project governance, poor communication, poor requirements gathering and even underhanded tender processes were major contributing factors to the failure. Is it fair to say that governments are incapable of addressing these shortcomings and actually delivering projects, now that we have the benefit of high visibility examples of the ramifications if they don’t?

      Conversely, SAP’s decision to back away for Business By Design and the precarious position this could potentially leave the NSW Finance & Treasury department illustrates how much more beholden you are to a service provider when you’re not just licensing their software for use on-premise, but licensing access to the system as well.

      Cloud provides wonderful opportunities for commodity systems to be delivered more efficiently (email being a prime example) – but are governmental CRM, ERP and HRIS processes sufficiently standards-based for it too to be considered a commodity system? And more worryingly, are we so pessimistic about any government’s ability to deliver change that we’re ready to immediately discount the possibility that with public scrutiny and burnt fingers, they can change?

      • H Brad … good question! No … we are not certain at all! There are pointers to the future however …

        My view is that we need to puzzle over the tradeoffs between internal consolidation and external SaaS on two levels. The first, and most contentious, is to actually go out and buy a cloud service – such as the SAP Business By Design scenario. The second is to puzzle over the new logic of cloud services and why it seems to be working.

        [BTW don’t panic about SAP’s commitment to SaaS ERP … Business By Design will migrate into HANA as SAP’s preferred cloud services platform].

        As I explained above, there are two elements to cloud logic: (1) “Cloudy is as cloudy does” – start with a service that is known to work as a multi-tenant shared service. (2) “It takes two to tango” – be prepared to change the way the agency specifies requirements and procures an ICT-enabled business solution to consume a more standardised service offering.

        So, if one was to seek to ‘make a go’ of ERP consolidation using cloud logic one would need (in my view) to be confident about both the suitability of the starting point (i.e. the ability of the existing ERP platforms to support multi-tanancy and to be configured to support diverse agency needs without too much customisation) and the ability to achieve a culture change in sharing behaviors in the agencies (i.e the ability of the agencies to be come intelligent consumers of a standardised shared service).

        The paradox here is that folks are critical of SaaS solution because they worry that SaaS = ‘one size fits all’ … but the truth is that SaaS solutions are explicitly engineered to support diverse user requirements whereas on-premise ERP solutions are customised to support the highly specific user requirements of one agency (only). Given that customised on-premise solutions are now an expensive luxury that few agencies can afford we need to bite the bullet on the tradeoffs inherent in shared-solutions one way or the other. The paradox is that the worst starting point for a shared-solution is a customised dedicated one because both technology and mindsets need to be shifted. At least with a SaaS solution the technolgy is already shareable.

        Of course it is true that the culture change part is the hardest – executives need to learn how to become intelligent consumers of standardised shared services by making tradeoffs between what they ‘want’ and what they can sustainably afford.

        My view is that in-house consolidation projects double-up the risks because they must deal with BOTH transforming the ICT solution and its infrastructure/service provision environment AND transforming the business processes and culture of the agencies involved to become ‘happy’ with the new consolidated shared service.

        At least a cloud services based approach has a better starting point (i.e. the SaaS solution already exists) and so the major focus of the project is on the business process changes.

        One way I often try to explain this is to use the metaphor of a communist or socialist economy. The ‘comrades’ all agree to form a collective and pool their meager resources to make a car for the people. 5 years later they are driving around in smoky old Trabants because this was the best technology they could conceive of within their closed system of thinking and constrained resources.

        A better way forward is the create a more open economy which takes advantage of the best cars available anywhere in the world – find one that works well (maybe a Nissan Leaf) and create incentives for its wider use. There is no way to ‘pimp’ a Trabant into a Leaf – they are totally different generations of technology.

        Internal ERP consolidation is, in effect, seeking to create a closed internal socialist economy where agencies seek to renovate ageing legacy technology to create a shiny new shared vehicle ‘on the cheap’. This can, in theory work … but it is time consuming and very risky.

        Better just to go out and buy a next generation vehicle on the market and focus all the effort on learning how to drive it well.

  5. I agree with Steve Hodgkinson’s learned opinion and second the notion that perhaps some new world education and training is required to bring the current technology thinking to the 21st Century inside NSW Government.

    I am sure similar statements and conversations were had with the advent of modern network computing in a world predominantly serviced with mainframes.

    In my opinion, the largest barriers to anything happening quickly in any organisation is the ability to cope with and understand change, then the second biggest challenge is accepting it irrespective of the technological advances and wiz bang functionality.

    We need a customer focused approach which put citizens at the center of any state government services model where focus is put into delivering services into communities in the most effective and efficient manner. This obviously can be driven by technology.

  6. SaaS as a solution is a hard ask here. Strategic alignment of ICT with org strategy is the only way to deliver fully realised benefits. SaaS doesn’t necessarily allow that to happen given the limitations on customization. Will be interesting to see how Cloud providers and Brokers attempt to resolve providing solutions that enable individual organisations long term strategic goals, in a vanilla world of service offerings.

  7. SH is right but I’d like to scratch a couple of assumptions:

    Anyone that ever worked in or with any government (my experience says that the world over …) understands that from the mindset within, there are two unarguable and immutable ‘facts’, ‘we’ (education, health, justice or whoever ..) are different with unique requirements, and that ‘we, because we’ve always done it that way’ must use this as a baseline if we are to consider even cosmetic change. I think its almost disingenuous and certainly a waste of time and potentially millions of $$ to even talk about the possibility of any potential change without first challenging and proving as absolutely incorrect the ‘immutable’ facts, as is more than evidenced by Qld Health’s odd attempt to make a sow’s ear from a silk purse!

    My point here is that cost, complexity, risk and culture are all threatened if the tail is allowed to wag the dog, that change is demanded of the technology to fit antiquated and inefficient process (which invariably actually hinder service provision) is quite simply an out-dated and daft proposition. The R&D departments of Cloud SaaS providers will never build a payroll or HR system that has the capacity to cater for the mind boggling complexities of a nurse or a policeman in Australia. Why should they when each state is different and each jurisdiction has an award system cobbled together over decades to avoid discussing or addressing the real issue which is base pay rates?

    So, where a situation is as politically threatening as pay there probably is no current answer, but basic day to day departmental accounting process, document/knowledge management, BI/Analytics reporting platforms, service desks, contact centres …. come on! If for no other reason than the immeasurable benefits of mobility with its anyone, anywhere, anytime, anything capacity and the speed at which relatively inexpensive apps can be built and maintained to fill holes in COTS solutions, its a lofty ambition worth supporting. That the project has got risk writ large from one end to the other is without question but that said and if it does go wrong, I’d bet the AG and Ombudsman reports will both suggest it was not the technology but the application of the technology that caused it.

    • Somehow we need to work towards being able to have a more explicit conversations with senior executives about the benefit/cost/risk tradeoffs between traditional ICT project methods and more agile SaaS-based approaches. The key choice is between, on the one hand, ‘perfect’ under-delivered late and over-budget (if at all) and, on the other hand, ‘good-enough’ delivered on-time and on-budget and then iteratively evolving. The only way to make sense of this is through hands-on experience. Seeing is believing. What we need is leaders in the public sector who have had some hands-on experience of the new art of the possible. This will just take time …

      • SH interesting issue with SaaS, if we take the NIST defininition, then iteratively evolving is not necessary on the agenda. Unless you are talking core product evolution which is simply vanilla functionality – ie enabling only common benefit. Surely the question for SaaS is how the Provider can actually deliver individual evolution to the customer that is able to aligned the SaaS service given the changing individual corporate strategic imperative of the customer. Otherwise SaaS will only ever be able to deliver vanilla capabilitity such as email and other non-competitive functionality.

    • I am reminded of the following summary of the NHS Program:

      “Poor and mal-performing projects close themselves off and avoid scrutiny; good projects are open to the challenge of feedback and have a habit of becoming better projects with more enduring outcomes as a consequence”. NHS IT Review 2007

      Anyone familar with QH Payroll, National ehealth PEHC programs will know that a key characteristic has been that they were closed off to scrutiny.

  8. It appears as though this project may well fall into the same trap that has seen other shared services projects fail. Namely, the belief that any ERP can automatically be applied to a shared services environment. Tax payers are surely getting tired of reading about failed government ERP projects. The reality is that ‘traditional’ ERP software systems that found favour in government 10 or 20 years’ ago simply do not lend themselves to this complex and demanding operational model. As Steve from Ovum rightfully points out, the system must be suitably configurable to fit diverse needs. System customisations must be ruthlessly avoided, instead relying on standardised processes. The ability to support on-going change is also critical, and this must be achieved without the need to march in an army of overpaid consultants for months at a time. It is somewhat surprising to witness severe difficulties, or indeed complete failure to successfully deliver shared services entirely, because of the failings of the chosen ERP platform. Surely the lesson is to analyse needs carefully, and choose wisely. Don’t assume that existing or traditional ERP software will work in a shared services environment. This approach has been the backbone of highly successful shared services deployments in countries such as the UK and Sweden. Thankfully, and to Steve’s point, these examples do not exist in an alternate universe. You just have to know where to look.
    Peter Simons, UNIT4

Comments are closed.