To Err is Human, but We Won't Forgive

In "A Losing Gamble with Public Funds: Why Large Public-Sector IT Projects Are More Likely to Fail and Are Harder to Cancel", Payson Hall, consulting project manager for US consultancy Catalysis Group, examines why the IT project game is so much tougher in the public sector than in the private. And while Hall told CIO magazine he did not have CMR in mind when writing the piece, from the outside his analysis of the way the deck is stacked against public sector IT projects from the very beginning seems to fit the project to a T.

And perhaps none more so than Hall's first point, on the public sector approach to risk. Under the sub-heading "To err is human, to forgive is not the policy of this administration", Hall makes the point that while both public and private sector projects seek to avoid failure, only the private sector expects risks to be identified, discussed and managed.

"The public sector paradox is that government's extreme risk aversion increases project risk," Hall writes. "Large IT projects are risky - many fail to achieve their cost, schedule or scope goals - yet up-front discussion of uncertainty and risk is often discouraged in the public sector."

And that, claim its critics, is just the way it was with ICS, the Integrated Cargo System destined to replace existing reporting and processing procedures with one integrated IT system, and intended to significantly enhance Customs' risk management assessment at the border while assisting industry to track cargo movements more efficiently. Those critics allege a culture where discussion of risk - or emerging problems - is not just aggressively discouraged, but actively punished.

"People really matter to processes like this: having the right people, the right skills, the understanding of business, the right understanding of project management, the right understanding of how to foster team cultures," White says. "These things are utterly, utterly critical to any successful IT project, and critically, I can point to probably a dozen people whose names I know directly, who have left Customs for one reason or another mainly around the pressures of the project or feeling unloved related to the project or feeling that there was no future, or indeed that there were risks in their future in being involved in the project. And those people represented the knowledge of the system that went before CMR, so Customs is faced with having to reconstruct its understanding of itself, by trying to bring people back in."

Rocky Road

CMR seeks to modernize the way businesses report the movement of goods across Australia's borders, and embraces major changes to Customs' legislation, business practices and IT systems. ICS is a key component - its core transaction hub.

CMR national director Jenny Peachey says although there were a number of drivers for CMR, the information technology was among the most significant of these.

"We have, in effect, over seven different systems that do what this ICS will do. Some of them were built in the late 70s, and they've survived the Y2K and the GST changes, but only just. Others systems were built through the 1980s; then the last of them came into production in the early 90s, which was Air Cargo Automation," Peachey says. "They were all built in different codes, they all had various levels of documentation, and in fact some of them had none. [Now] we're in a different age of systems development, and of course those systems were becoming very clunky: they were not going to last forever."

But development of ICS has been far from smooth. The project was first conceived of in late 1996 or early 1997, Harrison says, around the time Customs entered into an outsourcing contract with EDS. As Minister for Justice and Customs Chris Ellison told federal parliament in December last year, at the time, Customs believed the project was modest enough not to require a budget measure and that the agency could absorb the expense involved.

In March 1999, Ellison says, Customs estimated its costs for cargo computer programs at $30 million. In the following year, the parliament passed legislation enabling CMR. By March 2001, when the work was put out to tender, the estimate for CMR had risen to $35 million to accommodate the increased complexity and size of the program, and development was placed in the hands of Computer Associates, with IBM and partners developing the associated gateway component - CCF.

Yet late last year Customs chief executive Lionel Woodward and Peachey told a senate committee the total bill for the CMR project was now expected to be about $145 million. Ellison told the parliament the figure included $48 million for the ICS, $47 million for the CCF and $51 million to cover transition costs.

How can public servants get such estimates so wrong? Under the heading "Procurement: Lie to Me", Hall writes: "There may be a time and place for fixed-price/fixed-schedule contracts for massive IT projects. The time is almost never, and the place is nowhere near competent organizations. Solicitation of fixed-price/fixed-schedule bids makes sense only if work is exceptionally well defined, the processes and quality standards are clearly described in advance, and it is expected that any of several competent contractors could provide a comparable work product within a narrow range of schedule and cost variance."

None of these conditions would seem to have applied to CMR.