ATO urged to avoid lone-vendor projects
Summary: Government agencies should avoid picking just one contractor for large information technology projects, the taxation inspector-general has advised as part of his long-awaited report examining the implementation of the Australian Tax Office (ATO)'s troubled Change Program.
Government agencies should avoid picking just one contractor for large information technology projects, the taxation inspector-general has advised as part of his long-awaited report examining the implementation of the Australian Tax Office (ATO)'s troubled Change Program.
The Change Program was an eight-year, $756.7 million overhaul of 180 ATO legacy ICT systems, many of which had been in place since the 1970s. It began in 2004, when IT services company Accenture was tasked with replacing the system with an integrated core processing system. It was completed in mid-2010.
Inspector-General Ali Noroozi was first called upon to investigate the program by then-Assistant Treasurer Nick Sherry on 19 April 2010, following months of complaints about the implementation of the tax returns and refunds component of the program. The upgrade had led to delays in refunds, incorrect dates of lodgement for tax returns and changed dates for payment liabilities for some users. There were also concerns raised that the ATO had communicated badly to the public and taken too long to resolve the issues.
Chief amongst the nine recommendations listed in the report (PDF) handed down by Assistant Treasurer Bill Shorten today was the recommendation that the government require agencies undertaking massive ICT projects to avoid using a single contractor by dividing the projects into modules. This would also mean that changes in government policy would not have too much impact on the project as a whole.
"An integrated system that takes around seven years to deliver by a single contractor increases the overall level of risk. Contract management over such a long duration becomes increasingly difficult for a range of practical, commercial and organisational reasons," Noroozi said. "The likelihood of major government policy initiatives arising during lengthy projects is significant."
"Fixed price contracts for projects over a long duration also increase the risk of non-delivery in part or whole in the event that the contract becomes uneconomical for the contractor. This can pressure the Government agency to become captive to the contract and accept sub-optimal performance as a means to realise some benefits from the contract," he said.
Noroozi also recommended that agencies avoid lengthy projects in general and establish better governance practices, including the appointment of an independent ICT-skilled government representative to the decision-making body of the project.
The government has said that it will consider this recommendation.
In the other eight recommendations, the Inspector-General has advised the ATO to ensure it is more aware of the potential adverse outcomes of a software upgrade before implementing it so it will be adequately prepared to cope with any potential fallout. He has also recommended improved reporting, additional guidelines for ICT projects, better consultation with staff, and implementing the recommendations of a CPT Global report into the outage. All of these recommendations have been agreed to by the ATO.
The ATO has also agreed to publish a report into the costs and the benefits of the Change Program.
The ATO did not agree to the recommendation that the office consult with tax practitioners about a compensation application scheme for outages. The ATO said that it judges each compensation claim on its merits.
In the four months since the disruption caused by the failed upgrade, the ATO took some 17,000 complaints from taxpayers and tax agents.
Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.
Talkback
My point here is not to make Accenture a whipping boy for these failed projects but to highlight that the larger the project the greater the risk of failure, particularly when using a sole vendor. Yes, multi-vendor projects can be difficult to manage but if you tackle the task as multiple mini-projects things tend to go much smoother (and a little failure isn't as bad as a massive project failure).
The larger the project the greater the complexity. Seriously, what Project Manager believes that integrating/updating 170 separate applications into one mega-application, via one vendor has a chance of running to budget and schedule? Probably the one being paid a fortune no matter if it succeeds or not...
Government considers sole sourcing as risk mitigation. It's not, it simply creates a single point of total failure versus multiple points of smaller potential failures.
Governments and similar large organisations would be better off formulating long term strategies for system integration programs, but run as a series of progressive small programs without being locked into a single vendor. This provides room to adjust without a long term big bang committment of tax payer resources.
Chapter 2 gives the game away. The project is not just about replacing the old software and technical systems with new, its more about delivering new and improved IT enabled business services to citizens; this is a business transformation program with an IT project embedded in it and it needed managing that way. Clearly there were technical ICT issues, including a change of strategy half way through which may have pushed up the costs. But long experience shows that big ICT driven government business transformational programs rarely fail because of poor ICT project management. They typically fail because of poor executive leadership, bad management judgment, sloppy people and process change management, ineffective citizen/industry communication and blame shifting on to IT by way of business management abdication of responsibility; or put another way bad business governance. I am surprised the TIG, did not pick this point up and make much, much more of it; there were plenty of hints in the report. It would also be very interesting to know whether the $150 million or so a year of business operational savings forecast in the business case have been achived each year.