Peer Innovation and Production

The co-authors of Wikinomics explain how old-school companies like IBM can create value by embracing open-source models

by Don Tapscott and Anthony D. Williams

When Linus Torvalds first posted a fledgling version of Linux to an obscure software bulletin board in 1991, no one would have predicted that open-source software would be much more than a short-lived hacker experiment. Even after Linux became a serious global movement, large software companies denigrated it, arguing that open-source code would never offer the completeness or reliability of proprietary software.

One large software company saw things differently. IBM (IBM) recognized early that open source represented a new mode of software production that could potentially upset the balance of power in the industry. Today, IBM provides a quintessential example of how a large, mature company with a tradition deeply rooted in proprietary development can embrace openness and self-organization as catalysts for reinvention.

While many have heard about IBM's foray into open source, the details of the company's journey are not well known. The story offers an abundance of lessons, not just for software firms, but for all companies seeking to harness peer production for growth, innovation, and profit.

IBM's Open-Source Journey

IBM was an unlikely candidate to become a champion of open-source software. After all, we're talking about Big Blue—the company that became an industry powerhouse by building and selling proprietary everything. But by the mid-1990s, many of IBM's Web server and operating system businesses were failing. IBM's Web server product Domino, for example, had less than 1% of the market. It's fair to say that IBM did not have much to lose.

And increasingly, the company was being pinned between low-end hardware vendors, particularly Dell (DELL), and operating system vendors Microsoft (MSFT) (Windows) and Sun Microsystem (SUNW) (Solaris). Linux adoption, meanwhile, was growing quickly. Customers were increasingly asking about running Linux on IBM hardware, and the company was finding that new hires from universities were fluent in Linux and supported open source.

Linux offered solutions. It was a scalable operating system that would work well on small servers and could grow to handle heavier tasks. Because it was free, customers could try it easily. These advantages would help shift the locus of differentiation from operating systems to services and solutions, IBM's sweet spot.

Nevertheless, there was considerable nervousness in the senior management ranks. "At the time we had great concerns," recalls IBM strategist Joel Cawley. "Would the open-source community reject us? Will there be hostility to IBM? Will we face new legal issues that affect our ability to develop software?"

Communication Strategies

IBM eventually decided to dive in, donating large volumes of proprietary software code and establishing teams to help the Apache (Web server) and Linux (operating systems) open-source communities. Difficult as it was, the strategic shift was just the first of many changes IBM had to make. Ultimately, the decision affected everything from internal communication protocols to resource allocations to the competitive landscape of the industry.

Working in these communities proved difficult at first. Open-source software communities run on instantaneous, transparent back-and-forth communication and rapid product iterations. By comparison, internal company communications, attentive to internal sensitivities, are frequently slow and measured. IBM worked hard at getting the culture and processes right, starting with the way IBM communicated with Linux developers.

"When we were responding slowly with canned answers we weren't fast enough or transparent enough," said Daniel Frye, head of IBM's open-system development group. "It was not a level of technical exchange that was attractive to Linux developers." So Frye told his team: "I'm unplugging you from the network.