Building A Better Skunk Works

As the head of strategy, Harreld looked into the issue and "found a similar pattern across the board," he says, which he documented with 22 case studies. IBM had plenty of new ideas -- its famous research labs won thousands of patents a year -- but it had a remarkably hard time turning those ideas into businesses. The company had produced many crucial inventions, such as the relational database and the router, then watched while others, such as Oracle and Cisco, built huge companies around them.

What were the root causes? Harreld asked. In a classic defensive posture, IBM rewarded short-term results and served current markets, and was constitutionally reluctant to devote management attention, resources, time, or talent to rolling the dice. "Everything was based on the current period, not on the future," says David C. Dobson, an IBM strategic planner who's currently Harreld's deputy. Neither Gerstner nor Harreld spent much time on new businesses, and they didn't tap their "A-team" of executives to run them. "We were relegating this to the most inexperienced people," Harreld says. "We were not putting the best and brightest talent on this." They realized they had to risk taking people like Adkins off the mainstay businesses and putting them on projects that didn't fit neatly into any specific existing divisions but looked like potential billion-dollar opportunities.

Intriguingly, the best new ideas for emerging businesses came mostly from talking to customers, or outsiders such as venture capitalists, rather than IBM's own researchers, who focused on advanced technologies that couldn't pay off in a five- to seven-year horizon. "I'm not interested in new technologies," Harreld says. "I'm interested in building businesses. Often I find the ideas coming out of research are really good technical ideas, but they need lots of work to make into a business."

After Harreld considers hundreds of suggestions, picks an idea, and selects the EBO's leader, that executive goes off, sometimes with no more than one colleague, to start the business. Harreld meets one-on-one with the leader for three or four hours a month. And that's usually when the cultural hiccups begin. At first, the process is confusing and difficult for the transplanted executive, who's still stuck in the mind-set of running an established division. "It takes me and them four months at least to stop the crap," Harreld explains. "In an established business it's all about keeping things under control. These guys are so buttoned up. You bring them into a new business area, and it's almost hilarious." When Adkins embarked on his new venture, Harreld re-calls, he showed up at his first few monthly meetings insisting that there were no problems. "Rod came from a culture where the senior managers feel they're expected to know all the answers to all the questions and deal with the issues themselves," explains Gary Cohen, who helped Harreld run the EBO program from its inception until last year. "You understand a mature business because it has a level of predictability. But with an EBO, there's a lot you don't know, and you have to discover, learn, and adjust."

The second big cultural bump usually comes when EBO leaders try to empire-build, Harreld says. "Everyone tries to staff up because in a company like this, how do you measure stature? One, revenues. Two, people. Having run and built businesses, I've learned the hard way that you shouldn't staff up before you have clarity."

So EBO leaders begin working alone, or maybe with a single colleague, as Kovac did with the Life Sciences initiative. Harreld gives them a little money, and they try to tap IBM's collective expertise. "We wanted to be Tom Sawyer getting the rest of the company to paint our fence," Kovac says.

IBM's EBO leaders start out by proving the concept behind the venture through small pilots. They try to "go deep" with market experiments involving just a few prominent customers. "Whether it adds up to $1 million or $50 million may not be a good measure," says Cohen. "But it will tell you a lot about whether customers and clients think what you're doing is valuable." If these small pilots meet specific milestones for success, only then does IBM make the decision to pour resources into the project.

Kovac's experience running Life Sciences is a good example of how an EBO actually gets off the ground. Life sciences was an untapped area for IBM's sales force. Granted, for years it had been calling on the big pharmaceutical companies, but only to sell computer hardware and software for back-office functions, such as payroll and accounting, not for the core work of pioneering scientific research. But now that computing power is opening up extraordinary new opportunities -- the Human Genome Project, for example -- IBM saw the chance to sell info-tech tools for cutting-edge scientists. Big Blue also needed to reach out to the numerous small and medium-sized biotechs. In the past, these smaller firms would have been targeted (if at all) by the same sales force that called on small and midsized businesses such as supermarkets. The salespeople would have no specific knowledge of the needs of medical researchers. But Kovac understood that realm. Her own academic background was in chemistry (she held a postdoctoral fellowship at the University of California at Berkeley before joining IBM Research).