The concept here is about creating containers. By offering 20% time, Google gives its employees a limited container within which they have leeway, flexibility, and freedom. This is the hallmark of a decentralized culture. You are not completely inverting your organizational pyramid structure. You are simply figuring out how to give the maximum freedom to specific pockets within the system, while still being able to maintain the integrity of the enterprise. Centralized cultures, on the other hand, put in controls to prevent the chaos they assume would ensue were the controls not in place. They limit who can decide, and they demand records of time spent that map back to job descriptions and task lists (typically with a punishment for noncompliance). Decentralized cultures place limits, too, but they do it at the lowest level possible. They look for ways to create pockets of chaos because they know that's where the power of a decentralized system is unleashed. But they have to be real pockets of chaos. You actually have to let people make the decisions in those contained realms and then be clear about what happens after. Let people spend time on projects they deem important (some of the time). Increase the budget that department has to bring in outside consultants or buy equipment. The challenge is in making the pockets of chaos as big as possible, while still having controls on what emerges out of the chaos. Implementing new practices like these helps create a more decentralized culture.
This "pockets of chaos" idea is absolutely crucial, and I think the Humanize authors do a tremendous service in illustrating it and its potential advantages. Too often, when talking to leaders in heavily centralized organizations, you'll get a lot of push-back on the idea of decentralization. You'll get attacked for advocating for a complete absence of control, for allowing every person throughout the organization to have complete and total autonomy. You'll be dismissed as a crackpot (or worse, as unserious), more focused on individual freedom than on organizational objectives.
In my experience, these arguments are attacking a strawman. They're attacking the idea of anarchy, not of decentralization. Decentralization, at its essence, is a system that diffuses responsibility for experimentation and action throughout the system, rather than keeping those functions tightly constrained within an elite group. To be successful from an organizational perspective, everyone in the system has to understand and support the mission, know what tactics are off-limits, and have control over the resources necessary to try new things.
When leaders argue against decentralization, I think they are more often pushing back against one or all of these requirements. They don't want to communicate the mission because they frankly don't act in support of the mission themselves. It's a similar situation with defining which tactics are off-limits. They have no problem defining what others cannot do, but they don't want their own actions to be placed under similar constraints. They act according to a different set of priorities that derive from their power base rather than from the objectives to which their power is intended to be applied. And with regard to giving others control over resources--they simply don't trust others to use them wisely.
I try not to be too hard on these leaders. I understand where they're coming from. Communicating clearly, living your mission, defining limits, trusting others to make good decisions--these are all hard things to do, even when your motivation is organizational effectiveness. The leader typically has direct accountability for organizational performance, and that makes us hold these things closer to our vests than we probably should.
"Pockets of chaos," however, is a concept that can help us frame the necessary conversations with the people we lead and begin moving in the decentralized direction. Forget about the whole organization if that's too much to bite off. Start with a single project. Something non-threatening to your power base. Sit down with the project leader and hammer out the three essential pieces: (1) What's the objective by which success will be measured? (2) What tactics are off-limits? (3) What resources do they have at their disposal? Then, step out of the way and see what happens.
That kind of chaos can be a very positive thing.