Let's face a few facts. Almost all of the work of an IT organization is highly structured, disciplined and driven by well-controlled processes. IT organizations are put together around control points such as project offices. We measure performance and manage with the goal of improving that performance, and we are slowly but surely architecting a well-designed, comprehensive systems base. This is what we in IT do, and do well.

And none of it leads readily to innovation.

Charles Handy, a British management theorist, has talked about four kinds of organizational units. There's the unit that's built around networking to get deals done -- lots of quick conversations, calls and e-mails. There's the unit that's built around processes -- that's all about controls. There's the unit that's built around projects. Finally, there's the unit that's built around individual contributors, each of whom is valued for his unique skills.

Handy went on to say that this model extends down to individuals and up even to national cultures. In other words, each of us has a personal style. It's typically composed of a dominant and a secondary style -- in other words, we can flex and "fit in" to some extent. But we're most at home in our dominant style, and we're most likely to stay in an organization whose style matches our dominant one. (Fifty years of separation between the infrastructure staff and the development staff actually has its roots in this matter of self-selection by style.)

For innovation, you need the individual contributors and the networkers. Ideally, you'd have a group of people, some of whom are dominantly individual contributors and secondarily networkers, and the rest of whom are the reverse. But IT organizations are built on the building blocks of processes and projects. No wonder innovation is difficult, if the very people you need aren't around and the values of the organization are stacked against them being there.