It's no surprise that most big information-technology projects come in over budget. But a new survey shows just how resigned information-technology managers have become to the likelihood of cost overruns, at least in the public sector.
Only 18% of government I.T. executives have full confidence in their project budgets, according to a survey released on Nov. 30 by a company that specializes in doing technology cost estimates and related planning work. Another 54% say they sometimes have confidence, and 28% never do.
Cost overruns and missed deadlines are the bane of I.T., causing out-of-control projects to be killed, cutting off the flow of capital to more promising projects and leaving end users without services or functionality they were expecting. Making matters worse, the delays and hitches that spell trouble for, say, a year-long project often don't become apparent until the project is halfway done.
"Over-optimism reigns, whether it's within the government sector or the private sector," says Anthony DeMarco, president of Price Systems, the Mount Laurel, N.J.-based company that commissioned the survey. Many companies make the mistake of thinking they've done their due diligence when they've submitted their projects to a competitive bid process. The problem, according to DeMarco, is that bidders are often over-optimistic, too.
In the government sector, unrealistic initial cost estimates have undone projects like the A-12 attack jet and set back others, including the International Space Station, on which DeMarco's 85-person company was eventually consulted. Some of these projects have thousands of requirements, and the job of accurately "baselining" those requirements--figuring out the total resources needed to develop, operate and support each requirement--can run to between 3% and 5% of the total project cost.
So, for what size undertaking is a formal baseline methodology advised? In the federal government, the answer is prescribed, to some extent: Any project with an annual budget that exceeds $5 million must have a dedicated project manager. In the private sector, the main determinants are a project's complexity and visibility.
"You can have a smaller program, but if it's the biggest thing that your organization is doing, you'd better give it more oversight," DeMarco says. A failed project, he adds, can be career-ending.
Business Intelligence 
