The Challenge: Figuring out which project to tackle first is no easy task for IT executives. That's especially so when they're juggling competing pressures. The trick seems to be is working tightly with the business lines and closely monitoring projects.
Bob Yale, a principle of IT who runs the institutional-application group at The Vanguard Group, says when it comes to prioritizing IT projects, money talks.
"Clearly, cost pressures the industry is facing continue to be big. Everybody is cost focused. The challenge for us is how do you pick the right things to do. It's a constant battle," he says.
The focus is on value, says Yale, and making sure "you're getting the biggest bang for your buck." That means pursuing projects that have the highest client impact or which provide the most effective improvement in operations.
That's unlike the late 1990s, when IT executives had the luxury of focusing on cutting-edge projects like wireless and Web development, where the payback was not so immediate or measurable.
A Team Effort
In figuring out what brings value to the organization, Yale says IT managers have to rely on the business lines to tell them what works. "The IT shop has to be at the table to understand the business," he says. The buy in of senior business leaders is also key to ensuring a project's success, he notes.
Yale says Vanguard meets with the top five to seven business heads every two weeks to review projects and proposals, asking questions like, Will the project bring the right solution to the organization, and are existing projects on track?
Atul Sareen, a vice president in the office of the chief information officer at Ameritrade, says, "Resource allocation - how people should be assigned to the project" is one of the biggest pressure he faces in managing IT projects. "There's too much work. We have come up with a process where we sit down with the business line and prioritize our projects."
That makes it open and transparent, so that the business lines understand how IT is deploying its resources, he explains. "They can see what should be in the priority list and what should not be on the priority list."
Josh Levine, chief technology officer at E*Trade Financial Corp., agrees. "You can't work on everything," he says. That's why, "Partnership with the business line is key. It really has to be hand in hand with the business (lines) deciding what projects you are going to work on in the first place."
"Where things break down is when you don't go hand in hand with the business to figure out how to start a project or where to start," adds Levine.
And technology projects certainly break down sometimes. The Standish Group International Inc., a West Yarmouth, Mass. consulting firm, surveys IT managers every two years about the success of their projects.
The 2002 "CHAOS" findings, released in March of 2003, showed major improvements when it comes to successfully completing IT projects, but the numbers are still alarming.
Standish examined over 13,000 technology projects and found that success rates have increased to 34 percent, up one-third from the last survey and an increase of 100 percent from when they started tracking projects in 1994.
Project failures have declined to 15 percent, down from 31 percent cited in 1994. The remaining 51 percent are "challenged projects," which means they didn't meet initial expectations. However, there are improvements, as more than half of the challenged projects had lower cost overruns than in 2000.
Nonetheless, the survey notes that the average cost overrun is 43 percent of the project, down from a startling 180 percent when it was first launched.
Overall, Standish estimates that the lost-dollar value for U.S. IT projects in 2002 was $38 billion, plus another $17 billion for overruns. That $55 billion total, however, pales in comparison to 1994, when it estimated the total waste at $140 billion out of $250 billion in spending.
Another cause for concern is that time overruns have jumped significantly, increasing to 82 percent from a low of 63 percent in the year 2000. As well, only 52 percent of the required features and functions actually make it to the released product, compared to 67 percent in the year 2000.