If necessity is the mother of invention, then a skunk works is her creative big sister. And if you follow the logic of that family tree, necessity is also the mother of skunk works.
James Dallas, former CIO at Atlanta-based Georgia-Pacific, recalls becoming a skunk works devotee after seeing too many promising projects labeled failures before the development team had a chance to get all of the kinks out. "Within a large corporation, whenever you do something, the spotlight is on it," he says. "When it's not successful, people can be too quick to cancel it. With skunk works, you can get the lessons learned in a real roll-up-your-sleeves, hands-on manner that is away from the spotlight."
Skunk works are small research and development groups formed to produce a technology or product. While the goal of a traditional skunk works is to develop a commercial product or technology, the aim of an IT skunk works is to solve an internal business problem.
The skunk works concept dates to 1943, when a group of employees at Lockheed Aircraft Corp. designed and built the first fighter jet, even though the company didn't get a formal contract from the U.S. Army Air Force until the employees were four months into the project. They took the name "Skonk Works" from the "Li'l Abner" comic strip (and changed it to Skunk Works when cartoonist Al Capp complained). Decades later, another skunk-works-bred product made headlines when Apple Computer introduced the Macintosh computer.
IT skunk works projects vary in size and scope, depending on the financial resources and manpower available. On the ambitious end of the spectrum is the Tour Optimization Planning System (TOPS) developed about three years ago by a Georgia-Pacific IT skunk works. TOPS saved the company several million dollars annually by optimizing the routing of its huge dedicated fleet of trucks, says Dallas.
"It allowed us to factor in drivers, trailers, destinations and traffic lanes," he says. "That skunk works lasted over a year. We kept going through iterations. The first couple didn't work, but we knew there was value there, so we didn't stop it."
At Bacon's Information in Chicago, Scott Thompson, senior vice president of IT, implements a more modest form of skunk works when his department needs to analyze new technologies and products. "It's difficult to take on new technology these days because there are so many options and so much complexity regarding how these things fit together," says Thompson. "Skunk works is an appropriate approach to building some expertise in order to be able to evaluate technology."
Out of the Public Eye
A skunk works team usually consists of a handful of employees from the IT department who have demonstrated a knack for taking a fresh look at how technology can serve the organization.
To reduce distractions, many IT managers tuck the skunk works somewhere away from the day-to-day grind.
Keeping the operation under wraps is a key component of a true skunk works. According to Rick Swanborg, president and founder of Icex in Boston and executive in residence at Boston University's School of Management, "Sometimes CIOs have no alternative because some businesses consider anything technology-related as nothing more than a cost center. They say, 'Why are you spending time looking at technology? You just need to control costs.'" Twenty years ago, Swanborg was involved in skunk works that developed early versions of distributed LANs and portable computers.
Executives are more likely to accept and help fund a skunk works if the CIO can show how the project's goals are aligned with corporate objectives.
Experts advise keeping tabs on a skunk works but stopping short of imposing a schedule. "The nature of research is uncertain. It's hard to predict when you'll discover something," says Jeffrey Kaplan, a director in the Washington office of PRTM Management Consultants and author of the book Strategic IT Portfolio Management (PTRM, 2005).
At Georgia-Pacific, Dallas required his skunk works to come back with the results of a hands-on test in the first 60 days. For example, a skunk works team that was developing radio frequency identification technology several years ago had to apply RFID tags and determine what percentage had been successfully read by the sensors. The skunk works squads knew that without that first deliverable, which had to be produced for less than US$50,000, their project could be axed. "You would be amazed by how creative that would make people," says Dallas. "Depending on the technology, they'd go out and partner with a vendor, and the vendor would get creative with them."
When is a skunk works endeavor ready to become a legitimate project? "If you're starting to develop code and a prototype, you're on the fringe of a legitimate project," says Kaplan. But first, be sure that you have a proposed solution that meets a clear need and is viable, given technological and funding limitations.
Funding the Works
Skunk works aren't economically viable for some IT organizations. "I've seen companies that spend 99 percent of their IT budget purely on operating and maintaining their installed base. I would never tell those companies to allow any skunk works," says Kaplan. But he says businesses with large IT budgets and discretionary spending should have some skunk works. "But they should work on the most strategic issues and on solutions where the payoff is in the tens to hundreds [of] times the cost of the investment," he says.
Dallas says he has used dollars on hand for infrastructure investment or some of the contingency funds earmarked for large projects.
In the end, there is no guarantee that a skunk works will deliver a breakthrough. Kaplan advises CIOs to view skunk works as a gamble by telling themselves, "If nothing comes of it, well, nothing ventured and nothing gained." He adds, "If you can't say that, then you shouldn't be doing it."
Do's and Don'ts
Do establish a relationship with vendors whose products are at the heart of your skunk works, so that they will help you resolve performance problems.
Do encourage skunk works staffers to immerse themselves in the business environments in which the product or technology under development would be used.
Do congratulate the team for its efforts, even if its work doesn't culminate in a project.
Don't allow unfettered experimentation. It can run up costs and stall progress.
Don't be in a hurry to put your technology into production. Make sure it's stable and scalable.
Don't wait for a crisis before you establish a skunk works to solve a problem.
Artunian is a freelance writer in Newport Beach, Calif. Contact her at firstname.lastname@example.org.