Published on the 03/02/2017 | Written by Anthony Caruana

Over the last few months the federal government has been embarrassed by a number of major technology glitches…

It all started with #censusfail in August 2016. Then, before the ink was dry on the blame-shifting, sorry, inquiry and report, the Australian Tax Office and several other connected services suffered a catastrophic failure, blamed (unsurprisingly) on a vendor following a hardware failure.

Late last year, a “data-matching system” (with error rates conservatively estimated at 20 percent, not a very good one it seems) was responsible for delivering debt-collection notices to thousands of social security payment recipients. These notices often demanded payments of thousands of dollars from some of the most vulnerable people in our society.

This week, it’s the Child Support Agency under the pump, with rumours surfacing that payments administered though this body were being delayed around Christmas. It seems that a system upgrade, costing a measly $104 million has been delayed and botched. In case you didn’t know, this is the agency charged with distributing about $3billion in payments to single parents.

Been there, done that
Now, some context. Having personally worked in IT for most of the last 20 years, problems are inevitable. I started doing IT training, worked on the help desk answering support calls at the crack of dawn (my personal favourite call was from a fellow in a warehouse whose opening gambit was “My computer’s f@#$*ed”. When I asked for more specific information he said “Very f@#$*ed.”) I’ve worked on major projects that drive parts of the national electricity market, managed IT departments delivering services to thousands of users and even the occasional office move and fitout.

Anyone who has worked in the IT industry knows that cost over-runs, delays and poorly delivered functionality are a regular occurrence in technology projects. But organisations that so consistently deliver failure are a rarity. If this string of project failures had been churned out by a commercial entity, it is certain that the shareholders would have the entire board and management team dismissed.

What we are seeing in the government is the combination of two factors. Both are indicative of poor leadership.

Wherefore art thou, governance
Firstly, there is a catastrophic failure of good project governance. When one project goes off the rails, it’s possible that some unforeseen circumstance comes to pass. But this string of failures tells us the project management – which is ultimately the responsibility of government – did not adequately plan for problems and was ill-equipped to deal with issues.

While it’s possible to outsource the project management tasks, you can never outsource responsibility.

The second issue is a more strategic one. Why are so many new systems being deployed at one time? And why deploy an algorithm-based system, such as the data-matching program for debt recovery, on the eve of Christmas, when people are not only less likely to be able to pay, but when timelines for fixing issues are crunched because of looming holidays?

After the massive public failure associated with the 2016 census, it was clear that project governance was an issue. The sure sign of this is that no-one within the government or the Australian Bureau of Statistics was prepared to stand up and say it was their project and there was a problem. It took mere hours before the government shifted blame to a meek IBM which sucked up the bad press, knowing it would continue to work on more lucrative projects in other parts of the government.

The technology PM?
Despite having the most technically adroit prime minister of modern times – a huge contrast from the previous incumbent – we see technology failure after failure.

And let’s not get started on the complete policy and technical failure of the NBN.

The finger can be pointed, equally, at state governments, starting with Victoria and the Myki ticketing system – a ‘smart’ electricity metering which has delivered very little benefit to two million households – or the Ultranet debacle in the public education system.

But the federal government ought to be the standard bearer for getting it right. It has the most resources for hiring the best people, the greatest scale when it comes to buying power, and the potential to do the most good.

But it seems to consistently fail at the first hurdle – effective project governance and management. And it lacks the will to see this, preferring to point the finger at docile vendors unwilling to endanger their ability to bid for future work.