Building A Better Skunk Works

Rod Adkins couldn't understand why his career was taking such a sudden and devastating blow. Up until then, he was the epitome of a hotshot executive at IBM. In a culture where your status is determined largely by how many people you have working under you and how much revenue they produce, Adkins was a star. He ran the Unix computing division, a vital, thriving business with 35,000 employees and $4 billion in sales. And then, one day five years ago, the brass summarily stripped him of all that status and power. They reassigned him to a business that didn't produce any revenue. It didn't even really exist yet. Even its name -- "pervasive computing" -- seemed suspiciously abstract and weird. And now he had no one reporting to him. It looked as though this might be the company's way of letting a senior executive know that he was no longer wanted there.

"Geez," Adkins said. "What do I tell my mom?"

"He thought he was fired," recalls J. Bruce Harreld, IBM's senior vice president for strategy, who was one of the higher-ups behind Adkins's new assignment. IBM's then president, Samuel J. Palmisano, had to talk to Adkins personally and reassure him of his importance to the company. If anything, Adkins had actually been promoted, but that was hard to grasp because the new culture that Palmisano and Harreld were trying to instill was such a shocking departure from the company's long-entrenched ways. Five years ago, IBM was just beginning a radical shift in its attitude toward creating and nurturing new businesses. It used to be that IBM, like many large companies, came up with lots of promising ideas but had a hard time commercializing them. One of the many reasons was that its most talented and experienced executives, such as Adkins, took care of the big old established businesses that accounted for today's sales and profits, not the risky new efforts that represented the company's growth and future.

What Palmisano wanted Adkins to do was actually a lot harder than running a multibillion-dollar operation. It was creating one totally from scratch. And that's what Adkins quickly accomplished. Three years later, in 2003, his new venture, which applies wireless technology to extend computing beyond the home and office to such places as the car (with voice-command navigation systems), had already revved up to annual sales of $2.4 billion.

Adkins was a guinea pig for developing what IBM calls "emerging-business opportunities," or EBOs. The mission is to find areas that are entirely new to IBM and can grow into profitable billion-dollar-plus businesses in five to seven years. So far the program has been an extraordinary success. It's a lesson in something that's often a challenge for big organizations. "Breakthrough innovation has been very difficult for large, established companies," says Gina O'Connor, an associate professor at Rensselaer Polytechnic Institute and coauthor of Radical Innovation: How Mature Companies Can Outsmart Upstarts. "Projects get shut down before they come to fruition because senior management loses patience and they don't have appropriate metrics."

Harreld, who runs the program, promised the board that EBOs alone would produce two percentage points of growth for IBM. Given Big Blue's awesome size, that ain't hay: It means about $2 billion of new revenue every year. The actual results have wildly surpassed all expectations. Since the program's inception in 2000, IBM has launched 25 EBOs. Three failed and were closed down, but the remaining 22 now produce annual revenue of $15 billion, a figure that's growing at more than 40% a year.

But the impact reaches further than today's results. These internal startups are beginning to influence IBM's culture. "Through EBOs, IBM has become more of a learning organization," says Caroline Kovac, who built a new $1 billion, 1,000-employee business in computing for "life sciences" clients, such as pharmaceutical and biotech companies. "We've become more willing to experiment, more willing to accept failure, learn from it, and move on. It's more a part of our culture and our official processes. Now being an EBO leader is a really desirable job at IBM." Harreld says he doesn't even have to recruit EBO leaders anymore: "Today I have my peers coming to me and offering to run these."

The program began in September 1999, when Lou Gerstner Jr., who was then IBM's chairman and CEO, was working at home on a Sunday. Reading a monthly report, Gerstner found a line, buried deep, saying that pressures in the current quarter had forced a business unit to cut costs by discontinuing its efforts in a promising new area. Gerstner, a temperamental type, was incensed. How often did this happen?