If it wasn’t already obvious, this week’s report into Australia’s scuttled business registers transformation has left no room for misinterpretation. The case for large-scale technology projects in government no longer holds water.
This realisation is not new, with waterfall projects considered unfashionable long before the troubled Modernising Business Registers overhaul began in 2018. But it is thinking that has struggled to fully take hold in Canberra.
The Albanese government has chalked up three major tech wrecks inherited from the Coalition in less than 12 months, costing taxpayers more than $600 million. These failed projects have delivered very harsh lessons, and may act as the circuit breaker that force a rethink of the way these projects are managed in future.
In an autopsy of the latest blunder, former Service NSW chief executive Damon Rees identified a smorgasbord of issues that have the potential to be deadly to even the most well-scoped IT project, regardless of who is in charge.
It is an assessment that Assistant Treasurer Stephen Jones seems keen to apply elsewhere, stating that the government “will look to apply the lessons … to other digital and IT projects, including future projects”.
And for very good reason. In a new era of austerity, more than $600 million in failed projects – which also include Services Australia’s complex welfare calculator and the Home Affairs’ permissions capability – is not easy to swallow.
Of course, it is always easier for a government to junk the projects of its predecessor than its own, particularly when the decision is being made to stave off further spending that cannot be justified.
Starting with ending the “pre-conditions” that help justify what it describes as to justify “high-cost, high-risk transformation projects”, the report calls on government to take a more proactive approach to asset management and maintenance.
“Degraded digital and ICT infrastructures that have been long neglected produce critical risks which are then used as the catalyst to force wholesale, systems-focused, transformation at the expense of investment in change that drives meaningful value,” the report says.
With less of a need to rip-and-replace, such an approach allows project to be smaller from the get-go, with even large-scale projects able to be chopped up into “small, digestible pieces that independently add value”.
It is messaging that could prove useful for the government, as it looks to justify a program of smaller, more targeted technology uplifts for the coming years of fiscal discipline foreshadowed in this year’s Budget.
But shifting to a more iterative model of delivery in both theory and practice requires policy change, with the report also calling on government to “evolve its approach to investment, mobilisation, capability, and delivery of value”.
While it remains to be seen what form these changes will ultimately take, what is becoming increasingly apparent is that the federal government will introduce a digital government fund, potentially modelled on NSW’s Digital Restart Fund.
The fund would provide the flexibility to fund projects outside of the ordinary budget cycle, with a key element of the NSW’s fund being the requirement that agencies demonstrate wins before additional funding is released.
The government is already contemplating a similar fund as part of its new Data and Digital Government Strategy, but ironically the arrival of any such fund, tentatively named the ‘Digital Readiness Fund’, would likely coincide with the Budget.
Smaller projects also require fewer people to work on them, removing the prospect of “inherent complexity, overhead and productivity drain”. In the case of the MBR program, staffing costs alone were $12 million per month.
Minimising what the review describes as “external hard dependencies”, otherwise known as consultants and contractors, is another key lesson of the review and something the government has already vowed to reduce in its first term.
In the case of the business registers overhaul, Accenture – which was paid almost $200 million for the work over the last five years – was initially brought in for high-level design work under a $3 million contract that subsequently grew to $109.7 million.
The report also questions how far government is willing to go to avoid the stigma of failure, a driver in “exhausting all possible alternatives – rather than ‘biting the bullet’ and recommending project cancellation promptly when required”.
This will no doubt make interesting reading for the Department of Finance, which declared in May that it was too costly to change from SAP to a local alternative for the delayed Parliamentary Expenses Management System (PEMS).
For all agencies, the report stresses that the information on which an investment decision was made can materially change throughout the life of a project, meaning that any cost-benefit analysis should be a living document.
Projects that deliver value sooner and progressively provide better information on the total cost of completion will also allow government to ask these questions earlier, and ultimately avoid costly mistakes.
The government’s willingness to take on the findings of the report and radically shift its assumptions and approach towards projects could be the beginning of a renaissance in government technology projects.
But, as with anything else, talk is cheap, and change is hard, particularly given the interdependencies that stand in the way.
Do you know more? Contact James Riley via Email.
It is certainly true that many large scale government technology projects run over budget, over time and underdeliver. This is a problem for *all* large scale technology programs. Very (very) few run to time and budget unless significant “interpretations” have occurred along the way.
Rather than narrowly present this as a problem for government (which it is of course), we should consider this a problem for the industry.
For many decades, the technology industry has been developing versions of “agile” and “lean” development approaches – it would be good to see these take hold in some of this “too big to fail” (but they still do) projects.