In the wake of speculation that the Australian government may appoint Victor Dominello to ‘something’ significant to do with the digital delivery of government services, it’s worth considering the likelihood of success. Dominello has done good work in NSW. What could possibly go wrong?
Well, first, we’ve been here before.
We all know that the Australian government needs considerable work to bring it up to date with the systems and services – and user experience – we’ve come to expect from the major digital platforms.
Yet ministers seem convinced that resolving the ongoing issues and from their perspective performance of the public service is simply a matter getting the right bloke in – or, depending on who has a parliamentary pass – the right consultancy or vendor.
They are less willing to grapple with the complexity and effort – and risk the uncertain rewards – associated with sorting out the hardware, software, data, and legacy systems that underpin government work.
Those legacy systems are not simply technologies but include legislative requirements; the financial framework and funding models; the extant culture, skills, and practices; and Canberra power structures.
Without addressing those – and without the right remit and ongoing top cover at both ministerial and prime ministerial level – the ‘right bloke’ will likely be hung out to dry. The internal operating climate is unforgiving, the politics can get personal, and the nature of public policy is wicked.
Paul Shetler’s experience is evidence that being a great bloke, having good ideas, and success in other jurisdictions – plus good well-seasoned deputies such as David Hazlehurst, amongst others – is not enough.
Political support for innovation in Canberra is a timid and fickle beast. IT development and delivery rarely goes to plan, is an attractive cash cow for the legion of vendors, consultants and contractor, and a ready target for political attacks, worthy or otherwise.
Nothing has changed that suggests Dominello’s experience would likely differ. His current success stems in good part from his decision-making and political authority within the New South Wales government.
He would not have that inside the federal bureaucracy; indeed, federal bureaucrats are adept at managing political appointees once they are ‘inside the tent’.
Further, the nature of the problem, even within the Human Service portfolio, is considerably more complex – with sensitivities magnified by the Robodebt Royal Commission – than the digitalisation of the NSW public services suite accomplished by Dominello and his people.
Human Services is not a single, independent actor; its data – that is, your data – is shared with, and its operations are dependent on, the ATO and in some cases other departments.
The ATO, Human Services and Home Affairs compete for control of citizen identity, typically leveraging an efficiency-driven budget funding process and separate national security decision-making apparatus.
That’s hardly a promising basis for an enlightened, beneficial, and citizen-centric outcome. And in such a contest, a lone saviour without a clear remit, a strong political base, adequate resources, and great people, will fail.
Last, there is a belief that government services should reflect the speed and innovation of a Silicon Valley startup.
That’s not unreasonable given the plasticity of software and ubiquity of digital systems. But government and the private sector startup scene are different beasts.
Government ICT, data and digitalisation is a serious affair, especially where it touches the lives, livelihoods, and wellbeing of individuals.
It is societal and economic infrastructure. We expect it to be always up, always correct, thoroughly secure, predictable, and – somewhat naively – to have our best interests at heart.
It is managed as a risk-controlled cost centre, like all infrastructure, with funding allocated on an efficiency and risk-minimisation basis, often piecemeal.
And government is coercive in nature. As we’ve seen with Robodebt – and other examples in the United Kingdom, the Netherlands, and the United States – poor decisions over the design and operation of automated systems have real-life consequences.
A public dependent on government systems, whose data is stored in government networks, is not simply an end-user of a Silicon Valley start-up: they are taxpayers and citizens and have a right to expect transparency, accountability, agency, and recourse.
That is hardly the territory of a ‘move fast and break things’ or experiment-based lean startup ethos. Ministers shy away from the spectre let alone the reality of failure, even as they praise ‘fail fast’ attitudes elsewhere.
And when senior leaders acknowledge that sometimes systems fail – such as the COVIDSafe app – there’s rarely an effort to go back and learn from those failures, let alone share those with the public.
It is worth noting that in contrast to the government’s behaviour over Robodebt, the Dutch Cabinet resigned after its tax system wrongly accused 26,000 families of fraud.
Ministers want outcomes, and departments rarely deliver to the tempo and levels of satisfaction ministers, and the public, feel they deserve. (Somewhat ironically, they probably don’t appreciate that many within the Public Service, including within ICT, feel the same way.)
And so ministers tend to turn to individuals to resolve the problems, even though many of the impediments, constraints and incentives are deep systemic behaviours that have evolved over decades – often due to a myriad of earlier ministerial decisions and behaviours.
The rumours of a role for Dominello are the latest indication of ministerial frustration. But not – yet – of substantive and needed change.
