Dec 31 2009

Dollar Wranglers

Here's how to keep hold of the budget reins on that IT project.

Managing the budget for an information technology project can be a bit like herding cattle — especially if the project starts to go over budget.

At a time when agencies must provide Exhibit 300s annually to the Office of Management and Budget to justify and assure funding for their IT projects, it's more important than ever to deploy hands-on management techniques to make sure tasks run smoothly and minimize cost variances. The real trick is to eliminate budget overruns before they cause major problems.

Fortunately, project managers can create metrics to track budgets against missions, strategies and goals. Feds on the front lines and management experts point to numerous prescriptive actions — setting baseline requirements tied to end-user needs, using earned value management, working closely with staff members and monitoring costs at regular intervals — that managers can use to keep a tight rein on projects and avoid budget-busting consequences.

"It's important to make sure you have a solid requirements baseline — that it defines the requirements and what goes into the system," says James Shugars, acting IT director at the Food and Drug Administration's Center for Drug Evaluation and Research. "Spend extra time on the requirements on the front end. It's an inverted pyramid: If you spend a dollar to fix something in the beginning, it saves you a hundred dollars when it comes time to code."

Shugars came onboard at the FDA during the building of the Automated Drug Information Management System. Designed to help evaluate the safety and effectiveness of drugs, ADIMS was off schedule. Shugars' initial work at the agency focused on righting the situation. For starters, he began using earned value management reporting, as required by OMB, to look at trends developing over time and to spot variances. He also set up weekly meetings to make sure that those involved in the project were operating consistently with the FDA's spending plan.

"What's critical is to come up with performance measures. Things we've used before are lines of code and function points," he says. "For a large-scale integration effort, which would make it problematic for lines of code or functionality, we might use screens."

When it comes to giant projects, Shugars also makes an effort to keep the end user involved from early on in the process, ensuring that problems or issues are identified and resolved sooner rather than later. "We get the functionality in front of them as soon as possible — the nice-to-have versus the must-have." This allows project modifications earlier in the process, helping to keep budgets within parameters and minimizing unexpected problems later in the development process.

Fred Thompson, vice president of management and technology for the Council for Excellence in Government in Washington, adds that by doing this the team can keep in touch with users' changing needs. "The first planning on a project should be a group exercise and reality-based, in concert with the end users," he says. "A feedback loop should be built into the project plan; the more opportunities you have for that to occur, the better."

Following Along

Other managers also recommend keeping tabs from early in the development process. "You have to have progress checks as you go," says Douglas Brown, head of the IT Project Management Office for the Securities and Exchange Commission. He came to the SEC in 2004 to lead the agency's initiation of project and portfolio management.

Photo: Zaid Hamid
"If you spend a dollar to fix
something in the beginning, it saves you
a hundred dollars when it comes
time to code," FDA's James Shugars says.

"When we did an agency turnaround a few years ago, we were pushing a whole lot more fixed-price contracting than we used to do," he continues. "Where everybody gets along collegially and then you run out of money and go ask for more. Now we check the progress toward the financial goals."

An agency can apply other metrics, of course, related to cost and efficiency. Brown's agency spends about 15 percent of its IT budget supporting a document-imaging program, creating a giant repository of electronic documents that lets lawyers find what they need in 10 seconds as opposed to sorting through 10 separate boxes of files. "For the last couple of years, we've moved from an 80 projects mentality to two dozen programs, each one working on metrics in the sense of each project manager saying, 'How do I execute my mission? How do I push more data for less money?' " As things evolve further, program metrics and benchmarks will continue to evolve, too, Brown adds.

Keeping an IT project on track also relies on something much more fundamental: having a detailed plan from the start. "The military has this saying, 'Failing to plan is planning to fail,' " Brown says. "If you can't get anyone to describe how a project works, that's not good. If nobody's responsible for it; if you have a risk assessment that nobody's using, or they ignore, those can turn into issues." Many cost overruns, he says, occur on multiyear projects because the technology becomes obsolete and forces an agency to restart its development process. "You have to say, 'For $5 million, what can you provide for the next 18 months?' — as opposed to four years and $40 million."

Indeed, breaking down larger projects into segments with clearly defined start and end points will not only make budgets and metrics more manageable but also provide more value for the end users. "Projects can be broken down with deliverables that have value earlier," Thompson says. "That way, even if a project is canceled a year and a half in, you still have results; you don't arrive four years later and find you have something that doesn't give value to the organization."

Value in the End

Ultimately, budget management is a balancing act. Keeping projects on budget and on track is important, but it's equally important that the project provides value to the end user. "When program managers mainly focus on keeping IT budgets on track, they risk losing sight of the value-generating reasons why projects and programs were chartered," says Randy Englund, a project management executive consultant and author. "This happens when people skip over or assume there is consensus on discussions aimed at achieving clarity about objectives and value drivers."

If an IT project starts to go over budget, he adds, "check to ensure that the values are still present to justify continuing. Needs for the project may have changed, and so it should, too." A thorough understanding throughout the process of the intended accomplishments can help prevent cost overruns and revisions later on.

Although earned value management — a method for tracking a project's milestones against changes and costs — can help keep tabs on the elements of large projects, Thompson says, it is important to build room for adjustment into a schedule early on. "If one goes through a frank discussion of what can go wrong and what might be encountered, and look at those risks, and look at what those might mean in terms of costs, you can get funded in a way that allows you to deal with those contingencies," he says.

EVM can give a project manager insight into trends and the status of a particular project, but ultimately it's a combination of strategies — including regular meetings and updates, putting the product in front of the end user from early in the development process and keeping track of project values and goals — that will determine whether an IT project comes in on time and at the right cost.