Trends: Skunk Works - ' Skunk Works '

Skunk Works

Bell Canada isn't the only company setting up testing labs to prove the business potential of proposed tech projects. Firms from Yellow Corp., a Kansas-based trucking firm, to Ford Motor Co. and Webcor Builders, a West Coast building construction company, also are starting to manage risk this way.

In a recent CIO Insight survey about IT spending, 25 percent of computer and communications manufacturers and 21 percent of government IT executives said they are starting to require more pilots before committing to new IT spending. "These labs do two things," says Wharton professor Eric Clemons. "They defer the big investment until you know more, and they hide the project from the guys who would stop it."

Unlike old-fashioned "skunk works" idea labs, which author Tom Peters in his book, In Search of Excellence, defined as "a band of eight or 10 zealots off in the corner," these new IT proving grounds tend to be led by CIOs and other IT executives, and staffed by business people as well as technologists.

Another key difference? Rather than expect geeky tech brainiacs to dream up fresh ideas that may or may not result in any solid business initiatives, these new IT labs demand business results—and most within 90 days or less.

Yellow Technologies President Lynn Caddell, also the CIO of Yellow Corp., says the labs bring increased discipline to IT projects. She says that 95 percent of projects done at Yellow's lab come in on time and at or under budget. "The movement to create these new labs has just begun across industries, and should improve IT-business alignment and the CIO's ability to support the business," Caddell says. The majority of the projects go forward, but not all of them. At Yellow, about 20 percent of them don't pass final muster, Caddell says.

But the new IT labs can present challenges of their own. Jerry Luftman, a professor with the Stevens Institute of Technology, says they can be tough to manage. First, he says, "if these labs are started and run by the systems (IT) folks, they're more likely to fail. These labs must be run by both IT and business to succeed." Second, he says, there's a risk of having too many projects in the pipeline at once, and advises companies to limit each project to strict, 90- to 120-day timeframes.