IT failures show lessons have not been learned

The Inland Revenue's disastrous Tax Credits system suffered from many of the problems identified by the Parliamentary Office of Science and Technology (Post) report on government technology failures.

Post's job is to provide Parliament with objective analysis of relevant science and technology issues.

The Government IT Projects report published this week outlines the reasons for failures such as the Tax Credits system and the high-profile collapse of the Magistrates' Courts Libra system. It also looks at what measures have been put in place to avoid repeating the same mistakes, and how effective they are.

The Inland Revenue's failure to follow the official monitoring process run by Whitehall buying arm the Office of Government Commerce (OGC) is symptomatic of a wider problem identified in the report.

'The government has introduced a number of initiatives aimed at increasing the success rate of IT projects,' it says.

'However, it is too soon to say how effective these will be and it remains difficult to ensure guidance is followed by all departments and lessons learned from previous project failures.'

The OGC's Gateway Review process is possibly the most far-reaching attempt to address the problem.

At key milestones - or 'gateways' - in the implementation an independent review team from OGC examines the project and reports back to the Senior Responsible Owner (SRO) in the department running the development. The aim is to ensure the project is ready to move on to the next stage.

Reviews are graded red, amber or green. Red means remedial action is needed, amber means recommendations must be acted upon before the next level review, and green means the project is on target to succeed.

The six stages begin with Gateway Zero - a strategic assessment of the overall business case - and close with a benefits evaluation at Gateway Five.

The process is widely thought to be effective and while there have been a number of high-profile project failures since 2001, the OGC maintains these were all programmes begun before the process was instituted.

According to the Post report, the OGC considers Gateway Zero to have had particular impact.

'It is compulsory and expected at the start of a programme and is recommended for high-risk projects,' it says.

The OGC clearly emphasises that all 'big bang' projects with a single go-live date are automatically classified as high risk - the tax credits system falls into this category, but it seems the Inland Revenue did not follow the process to the letter.

The Post report acknowledges there is some uncertainty whether the Gateway process is mandatory.

'It is possible for political pressures to override Gateway recommendations - so, for example, projects may not be cancelled if they are of high political priority, even if given several red lights,' says the report.

Liberal Democrat IT spokesman Richard Allan says this is one of the big problems.

'The underlying issue is that politicians want changes that just can't be delivered in time,' he said.

'The whole Gateway Review process came out of a need for third-party scrutiny. I hope the Public Accounts Committee and the National Audit Office will take a robust view of those departments that have not take advantage of those processes which are designed to help,' he added.

Other issues highlighted by the Post report include the benefits of breaking projects down into smaller parts and the need to provide time and resources for training users of the system.

Post board member Professor Jim Norton says the government still isn't putting enough effort into the people side of technology projects.

He says the rule of thumb should be half the budget for a 'business change' project goes on issues such as training, explaining why the project is necessary, and budgeting for a fall in productivity while staff become used to the new system.

'The government is only paying lip service to the human dimension - it is getting better at the mechanics of running projects in terms of the IT, but not in terms of the people who use the IT,' said Norton.

'The interface between people and IT is often the problem, but the technology will be blamed.'