6 Secrets to Managing IT Projects

Despite a glaring need for specialist project managers, universities tend to embark on major IT implementations without their services. CT shares 6 ways to make the best of a less-than-ideal situation.

• By Dian Schaffhauser
• 06/20/13

Nobody should be surprised when new IT projects don't go to plan. According to the latest McKinsey research, the average large IT project overshoots its budget by 45 percent, takes 7 percent longer than expected to complete, and produces only 44 percent of its anticipated value. And that's the industry average. Higher education has further hobbled itself by being slow to recognize the need for dedicated project-management specialists. According to Educause's "Top 10 IT Issues, 2012," while the role of project manager has certainly been around for a while, it "may still be new to some IT organizations."

"The concept of project management is still evolving throughout university culture," notes Christina Griffin, director of the Division of IT Project Management Office for The George Washington University (DC). Nevertheless, she says, "It has come a long way in the last couple of years." Two years ago, for example, the project-management constituent group at Educause 2011 drew about 40 participants; the latest gathering had a crowd more than twice that size, with people "literally falling out of the room."

To ensure that major IT projects run smoothly, your best bet is to hire dedicated PM specialists to join that same throng. But if your pleas fall on deaf ears--if the powers that be moan about hiring freezes and general poverty--don't throw in the towel. Even without PM professionals on hand, your team can beat the odds of failure by following these six well-honed PM practices.

1) Do Pre-Project Planning
Requirements gathering is the job of figuring out what users really need. GWU decided to formalize its approach to this work by adopting standards and best practices from the International Institute of Business Analysis. Bill Koffenberger, director of service management at GWU, says the university has tapped "specialized staff" for the effort, specifically business analysts. They use "a series of approaches, checklists, and templates"; do a lot of interviewing; and perform "small-group brainstorming" in an effort to identify the stakeholders and "exactly what these projects are supposed to accomplish."

Although this due diligence can delay the start of projects, it also prevents IT from committing itself to new initiatives without proper consideration. Before the new process went into place, for example, campus constituents would simply show up with money and a shopping list, says Koffenberger. But many of them didn't understand that their implementation requests were often very complex, involving interfaces to ERP systems or serious security issues. Now, IT is learning to take a step back, notes Koffenberger: "Is that something we already have the capability for in-house, or are there [other vendors] that better meet our needs?"

Despite the additional hoops, campus constituents are cooperating with GWU's new process, in part because it helps them create a stronger business case. And, in these lean budget times, a compelling business case is a must since projects are competing for limited IT resources. Unlike in the business world where the key driver is usually profit, says Koffenberger, the primary drivers in higher education today tend to be cost savings and efficiencies, along with student satisfaction and higher standards of learning.

GWU's planning process is paying dividends. Recently, the university began a project to install a risk-management information system. A presentation by a vendor left the IT team impressed with both the product's low cost and ease of implementation. By going through the formal project-planning process, though, additional requirements surfaced, with GWU's leadership asking questions about potential long-term uses for the system. "We changed gears and went a different route," explains Koffenberger. "We took a step back and created better requirements."

2) Teach PM Skills to Staff
Like every college IT shop, GWU's has a long list of projects--probably about a hundred. "We can't possibly manage each and every one of them," laments Griffin. With a view to working through the list faster, the school has started assigning management of some projects to the departments that ultimately will benefit. To avoid a potential PM train wreck, GWU has adopted a training curriculum, including a game that helps people learn how to "manage their own projects."

The Bridge Game, a creation of Bill Geraats, an "experiential learning specialist" in Canada, provides a kit of components that allows multiple teams to create two-dimensional, table-sized bridges using strips of plastic and various joiners. The goal is to introduce the basic concepts of project management and show participants what can go wrong. Hazards include wayward "client representatives" who are reluctant to reveal too much detail (thereby introducing bureaucracy); important-sounding--yet faux--change requests that have "fallen out" of the fax machine, offering a taste of what happens when requirements aren't entirely understood or governance isn't followed; and disruptions in leadership and staffing that mimic the realities of resource management and scheduling.

The game fills twin roles at GWU: It's used by the business school to introduce students to the program, and the IT and finance departments are deploying it as a training tool. While students are resilient in the face of the game's obstacles, staffers "tend to react a little more negatively," says Adam Donaldson, program manager in the business management and analysis group for the university's Finance Department. "They tend to shut down more. They don't bounce back." In fact, one staff participant left the room and refused to play anymore. "That was unfortunate--we didn't want to turn people off," notes Donaldson. "But it is a good lesson to take away. Learners with different styles react differently."

Even for employees who won't assume the mantle of a project manager, there is a point to the pain, explains Donaldson. Since all staff members participate in project teams at some point in their careers, it's helpful for them to learn the precepts of project management so they can understand why project managers do what they do.

3) Invest in Discovery
With any new project, it's difficult to know at the outset exactly what you need or what obstacles you will encounter. While this may be irksome for in-house projects, it can be ruinous when vendors are involved. Vendor contracts usually spell out a specific set of services, and any add-ons or changes are likely to incur hefty fees.

Potentially expensive unknowns were the problem facing CUNYfirst, a project that kicked off in 2007 to replace The City University of New York's systems for finance, human resources, and student administration. With 25 institutions spread around New York City and systems that have been in place for as long as 35 years, CUNYfirst is one of the most complex initiatives ever undertaken by a college system.

CUNY chose Oracle to provide the software--PeopleSoft--and handle the integration work. "We wanted one-stop shopping, so to speak," says Suman Taneja, executive director for CUNYfirst. The university system negotiated a fixed-price, performance-based contract, meaning any departures from the agreed-upon requirements could become pricey. To stay within budget, CUNY could afford a minimum of change orders.

To avoid unintended surprises, CUNY asked Oracle to participate in an extensive discovery process before the contract was signed. For four-plus months, teams from both organizations worked their way through a list of 3,500 requirements.

The process gave both sides a chance to understand each requirement in depth, and helped CUNY avoid major fees during the implementation phase. But the joint project resulted in another benefit as well: The time together gave each team a chance to understand the other's work culture, and gave CUNY an opportunity to understand Oracle's implementation methodology.

4) Create Project Lifecycles
Whether in business or education, projects tend to begin with a bang and then lose headway as team members get reassigned, responsibilities shift, and fires break out elsewhere. As a result, projects often fail to deliver on their potential. To combat this, GWU has worked hard to introduce more discipline into its PM process. Starting in the 2010-2011 academic year, a working group was formed to develop a more formal PM lifecycle. The group comprised reps from the project-management office, the IT business-process management team, finance-system support, and the business-management and analysis group. The goal was to create a framework for standardizing methods, documentation, templates, terminology, and approval processes across projects.

Previously, "project management was typically an IT thing," recalls Griffin. "We had our processes in IT--and that's what we followed for IT projects. But it became evident in some of our projects with the businesses that we needed more stakeholder feedback." While projects like these may have a technology component, most of them also have business aspects--such as user training, stakeholder communication, and buy-in--that should be led by the individual departments.

To ensure that stakeholders have adequate oversight of projects, GWU decided to pursue a phase-gate model. Phase gates are checkpoints during the life of a project where everybody stops to ensure that the work is still on track. The phase-gate reviewers--the business owners--must give their approval before a project moves onto the next milestone. This prevents a project from heading to completion, for example, only for the team to learn that a key requirement has been overlooked, forcing IT "to go back to square one," says Griffin.

For GWU's major projects, phase gates are positioned at several points: after development of the project idea; after completion of the proposal; after the project plan, analysis, and design are finished; after the development is wrapped up; post-testing; post-launch; and at project closure.

"The gate review was one key to make sure our leadership was aware," says Donaldson, noting that previously "there was some unjust finger-pointing, with people saying, 'Well, I thought IT was going to do this,' and IT saying, 'No, the business should have been doing this.'" In Donaldson's view, the lifecycle approach ensures that everybody is "on the same page--that it's all clearly spelled out, so that we know who is in charge of what."

5) Adhere to a Governance Structure
A clear command structure is an absolute must. Nothing can derail a project faster than when contributors don't have a clear understanding of who's making the decisions and why.

To tackle its massive implementation of PeopleSoft, CUNYfirst is using a collaborative governance structure. A core team of about 35 people focuses on project issues that cut across all campuses--business processes, functional configurations, training, change management, and technical problems. A program-management office handles issues, risks, schedules, budgets, and resources. To handle issues at individual colleges, though, each school has its own CUNYfirst team, headed by a "campus project executive." A project-management liaison coordinates with the core team and manages the plan for his college.

As Taneja sees it, the core team provides "the methodology, the guidance, and the central development, but the campus team has the responsibility of doing these things on their campus."

Before the launch of any new applications at a school, CUNYfirst leaders sit down with college leaders to examine the risks and decide whether the campus is ready.

What CUNYfirst is most eager to avoid are customization requests. "Our target was zero," says Taneja of such requests. "That was obviously impossible." In some cases, changes are necessary because the colleges follow a different academic calendar or serve a different type of student. In other cases, he notes, PeopleSoft doesn't possess the required function, forcing CUNYfirst to create a custom solution.

If the change "can be done quickly, doesn't cost anything, and doesn't have a schedule impact, typically we just get it done and move on," notes Taneja. But a substantive change--something costing tens of thousands of dollars--initiates a cascade of reviews. First, the change order is evaluated by a steering committee. If it passes muster, the CUNY CIO scrutinizes it. Eventually, it lands on the desk of the CUNYfirst project sponsor, the executive vice chancellor. "He's the person where the buck stops," says Taneja.

6) Measure the Outcome
During long-term projects, it's easy to forget to measure success. By the time the project has lifted off the ground, won buy-in from stakeholders, and been implemented, those leading the charge have frequently moved onto new initiatives. Nobody's left to determine what worked, what didn't, and what's worth repeating. That's a big mistake.

Experienced project leaders know that monitoring the success of a project has multiple payoffs. By broadcasting new successes--even if they're only incremental gains--a project manager can maintain momentum and keep attention focused on the initiative. And remember: Even though the head of the parade may have finished marching, those in the rear may just be appearing in front of the spectators.

Second, notes Taneja, the organization can learn valuable best practices as projects unfold, generating a continuous cycle of improvement. Within CUNYfirst, for example, the PeopleSoft training evolved with each new implementation. Based on experience garnered during earlier installs, users now get a "replica" of the PeopleSoft system a few months before the real thing is launched at their college. "We call it the simulated training environment," explains Taneja. "It has their data, their security, their configuration--it looks like their system. The feedback we were getting was that that is the best way to learn--on the job. We have evolved, and we have seen people ramping up quickly."

Although CUNYfirst is still in the middle of its systemwide implementation, benefits are already accruing. A new course planner, for example, spells out exactly which classes students need for a given degree and tells them whether they've fulfilled the requirements for a course. Another major objective was the elimination of Social Security numbers within the systems, something that has been achieved at schools that have launched the new solution.

Finally, the new applications provide the foundation for data-driven decisions. As Taneja notes, "I'm starting to see some of our provosts using this data to look at decisions about classes--about students--so it's starting to seep into student success."

While Taneja acknowledges that these benefits are qualitative in nature, he does highlight one major quantitative benefit: The PeopleSoft software replaces old systems that required massive amounts of pricey maintenance. That cost goes away the moment the new programs come online.