FAST IT PROJECTS CAN BLOW THE BUDGET–AND STILL COME IN LATE. PROJECT MANAGEMENT BEST PRACTICES CAN HELP MEET EXPECTATIONS FOR QUICK DELIVERY.
BY MALCOLM WHEATLEY
AS BRITONS WOKE UP TO the advantages of Internet-based banking in the late 1990s, staid British banks were slow to realize the opportunity. Sensing blood, a number of second-tier financial institutions saw a chance to leapfrog the majors with Internet banks of their own. It was a classic first-mover advantage scenario: The quickest off the mark could execute a land-grab and lock in the keenest customers, leaving slower institutions with the crumbs.
Certainly, the faster strategic projects can be implemented, the sooner the benefits can be banked, but achieving greater speed can eat up resources, resulting in budgets and timescales finely balanced between risk, reward—and ruin. For example, when Abbey National, a U.K. savings and loan institution, launched its Cahoot Internet bank in June 2000—developed at a cost of approximately $400 million—the fledgling bank's Web site crashed on its first day, unable to accept applications for accounts for most of its first two days of operation, and crashed again on its fourth day.
Halifax—another savings and loan business with banking ambitions—thought it would learn from Cahoot's troubles. Spotting likely technical troubles, it rolled back the opening of its own Internet bank, Intelligent Finance (IF), from July 2000 to September of that year. The delay was painful, as the company conceded in its annual report to stockholders, but was necessary to avoid any repetition of the Cahoot experience.
Project acceleration—perhaps to recover from project slippage or to trump a competitor's market advantage— can prove perilously expensive, resulting in unsatisfied customers and blown budgets.
Checks and Balances
➔ Quick-start IT projects call for careful judgment. The perspective at a project or portfolio level may not be the same as the view from the overall business level, warns Bob Woodruff, special assistant to the CEO at project planning specialists Robbins-Gioia, Alexandria, Va., USA. “It's not just the cost in terms of blowing the budget, it's the opportunity cost of the slippage,” he says. “Exceeding the budget by 10 percent may not be significant compared to the lost revenue foregone through being late.”
Getting the basics right can help, of course. IT projects are no different from other projects in terms of the need for solid disciplines such as resource planning and forecasting—especially when attempting a fast implementation. “Budget, budget accurately, and budget with sufficient granularity,” says Steve Hoddinott of Maconomy, a Copenhagen, Denmark-based provider of project and portfolio management solutions. “If you are going to need someone with Oracle database administration experience in six month's time, know it now, and not in six month's time.”
PROJECT AND PROGRAM MANAGERS—AND CIOS— OFTEN DON’T GET ENOUGH WARNING OF PROBLEMS UNTIL IT’S TOO LATE.
David Hurwitz, Chief Marketing Officer, Niku Corp., Redwood City, Calif., USA
➔ Rushing an IT project can be risky and costly, so the pros and cons should be considered carefully.
➔ Basic project management techniques can help, but there also are some IT-specific best practices that can compress a schedule.
➔ Many companies developing software are ignoring best practices and turning projects into “ventures into the unknown.”
➔ Outsourcing companies in India have embraced the impetus for process improvement and speed.
What's more, Mr. Hoddinott recommends project managers apply that same granularity to skills definition; for example, if you need someone with version 9 expertise, and the resources being offered only have version 8 expertise, then that discovery will impact just how quickly your team moves forward. At speed, he says, such minor bumps in the road can quickly send a project crashing off course.
Acting decisively to mobilize those resources also is vital. When insurance company Prudential, London, U.K., established a project in October 2002 to locate an 850-person call-center in Mumbai, India, using “virtual desktop” software from VmWare, Palo Alto, Calif., USA, program manager Simon Matthews was given just six months to complete the task. A director of infrastructure services was assigned the role of IT sponsor, making sure the projects didn't hit any resourcing roadblocks— perfectly possible with a core team of just six people, supplemented by 25 to 30 additional staffers at times of peak demand or when vendor resources were applied.
“We had a small group of empowered people who were keen to be a part of the project because it was ‘sexy,’” Mr. Matthews says. “The director made the right call about getting the right people appointed.” Mobilization wasn't an issue because the team did a good job of “selling” the project to potential resource units in terms of the skills and potential career kudos. The first “test” call arrived in India in May 2003, and the project went live in July 2003.
➔Other tried-and-tested project management techniques also work well in the context of faster IT projects. “When there's pressure from stakeholders to deliver IT projects sooner, I always advise looking at the project plan to see what concurrences could exist,” says Gina Davidovic, director of project solutions at Toronto, Ontario, Canada-based project management training consultancy Bay3000. “Invest the time to crawl through the project process that can be undertaken concurrently, rather than sequentially.”
Conventional wisdom typically leaves customer feedback until quite late, when definitions are fixed and code generated. One outcome is costly rewrites—which act as a brake—when users see something that won't work or that they don't like. But even when rewrites aren't necessary, involving stakeholders earlier is a best practice when time is tight, Ms. Davidovic says. “Thanks to that early involvement, they hit the ground running, with higher levels of understanding, and require less time for training and sign-off,” she says.
Staying on top of things, too, can help sustain speed. “One cannot over-emphasize the need for tactical control and visibility,” adds David Hurwitz, chief marketing officer at Niku Corp., Redwood City, Calif., USA, developers of the Workbench project planning tool. “Project and program managers—and CIOs—often don't get enough warning of problems until it's too late. The problem might be at a subcontractor or with some technical resource— and if you don't know about it, you can't recover from it.”
VERBAL AND PICTORIAL VIEWS CAN BE JUST AS EFFECTIVE IF DONE WELL AND ARE INVARIABLY QUICKER TO UPDATE, ASSIMILATE AND COMMUNICATE.
Nuala MacDermott, Partner, Business Transformation Group, London, U.K.
Expect the Unexpected
For Nuala MacDermott, a partner in the Business Transformation Group at London, U.K.-based PA Consulting Group—who worked on the successful launch of Egg, another British Internet bank—doing IT projects faster without blowing the budget has become something of a passion. “To do things faster, do them differently” is Ms. MacDermott's personal mantra. The Egg project, she points out, yielded a number of best practices, many of which she now routinely applies when time is tight.
➔Too many IT projects, she points out, set their own pace, thanks to the way that they are structured and monitored. Doing things differently, she argues—through such best practices as slimmer IT project governance models, agile development methodologies and “storyboarding” definition techniques—provides a much-needed, if radical, reform of the way that IT projects are run. “Written documents often are not needed,” she says. “Verbal and pictorial views can be just as effective if done well and are invariably quicker to update, assimilate and communicate. A picture is easy to change: a two-inch-thick document isn't.”
According to Ms. MacDermott, faster IT projects also call for a greater adoption of formal software development methodologies. A number of proven iterative define-and-code techniques exist, she points out, that aim to deliver agility and speed by breaking projects up into much smaller steps than is usual. These phases then can be blended with more traditional formal software development methodologies, such as the reliability-centered CMMi (Capability Maturity Model Integration), developed by Carnegie-Mellon University's Software Engineering Institute in the early 1990s on behalf of the U.S. Department of Defense.
A modest proposal, one might imagine. Yet the size of the leap that this requires can be gauged from the relatively low levels of take-up of even the relatively conventional CMMi approach, some 15 years after its launch. “Some organizations still are operating as though CMMi was never developed,” says Arnold Ruskin, PMP, Ph.D., a partner in Claremont Consulting Group of La Cañada, Calif., USA, and a lecturer in project management at the University of California, Los Angeles.
EXCEEDING THE BUDGET BY 10 PERCENT MAY NOT BE SIGNIFICANT COMPARED TO THE LOST REVENUE FOREGONE THROUGH BEING LATE.
Bob Woodruff, Special Assistant to the CEO, Robbins-Gioia, Alexandria, Va., USA
The result is to transform every new software project into more of a venture into the unknown than it needs to be, he points out. “It's like taking your car to an auto mechanic who has never opened a carburetor before and is learning the job while practicing on your car,” Dr. Ruskin says.
No wonder, then, that veteran software implementers advise doing as little code development as possible. “If you write your own code, experience shows that you're almost certainly going to go over budget and probably won't come close to delivering the functionality you need,” says Daniel Garvin, managing principal at Concentric Solutions LLC of Greenville, S.C., USA.
Especially when time is of the essence, he urges, off-the-shelf packages—or even several best-of-breed packages, stitched together—offer a much higher chance of delivering the vast majority of the project requirements. In Mr. Garvin's eyes, hitting the top 80 percent of the requirements on time and on budget beats aiming for 100 percent with “homebrew” code and falling flat on your face.
The Speed of Outsourcing
➔An increasing number of software experts point to the growing maturity of outsourcing companies as a means for compressing project schedules, particularly the Indian software industry. Among the countries now routinely associated with offshoring, India arguably offers a unique mix of high-quality talent, low wages—and skill and experience in applying formal software development methodologies. “Indians have seriously embraced formal software development methodologies, thus boosting the lines of error-free code created per hour,” says Marc Hebert, executive vice-president of marketing for Fremont, Calif., USA-based software company Sierra Atlantic, which has an 800-employee software development center in Hyderabad, India. “You just can't underestimate the impact it has had on their productivity.”
Indian software services company Wipro Technologies is firmly convinced of the value of formal software development methodologies as it bids for work from Western companies. Wipro has just received CMMi level-five certification, says Laxman Badiga, the company's chief executive of talent transformation—a title roughly equivalent to head of resourcing. He claims that Wipro is the first Indian company to be certified to CMMi's highest level, but given the region's current focus on consistency and the drive for speed, other companies may not be far behind. PM
Malcolm Wheatley is a U.K.-based freelance writer who has written for CIO, MSI and CSO magazines.
PM NETWORK | JUNE 2005 | WWW.PMI.ORG