If you want the lightning bolt of innovation to strike again and again, you have to live with the inefficiencies. You’ve got to make a basic philosophical choice that the inefficiencies and disorder are worth the benefits.

It’s impossible to have an organization with all the fire and zeal of decentralization and the complete efficiency of centralized control. Pick decentralization, fully implement it, and live with its difficulties as best you can. If you try to go halfway, it’ll be like having a country shift from driving on the right side of the road to driving on the left side of the road, but only implementing it part way.

Though expensive and usually belated, reorganizations are widely regarded as the only way to realign an organization with its environment. As a report by the Boston Consulting Group put it, “Rapid change requires companies to reorganize faster than ever before.” Good luck with that!

What’s needed are radically new organizational models that downplay formal structure. In a world of relentless change, trade-offs need to be made as close to the front lines as possible. Boundaries must be malleable. Resources, rather than being hoarded, must flow unhindered toward promising opportunities. Interunit coordination must be the product of nimble, self-organizing communities and market-like transactions rather than blanket policies or cumbersome councils. In short, we need organizations that, like the biosphere, the internet, or a vibrant city, are more emergent than engineered.

Third, like nuclear power plants and space rockets, bureaucracies are complex, integrated systems. Every process is connected to every other process. This lack of modularity makes it difficult to change one thing without changing everything. Where do you start? That’s the paradox of change in a bureaucracy: what seems doable isn’t transformational and what’s transformational doesn’t seem doable. The result: an endless succession of tweaks that never succeed in making the organization fundamentally more capable.

You have to integrate that into the company. The processes that you put in place to do that, they have to be very deliberate. It doesn’t happen by itself. What happens [naturally] is entropy. You have to leverage scale in a way that doesn’t disable entrepreneurialism, business ownership. It’s integrative. It’s not centralized. Centralized is a very different thing. This scale work is bringing the leaders of the businesses to work together towards a plan that not only optimizes the company, but in its best form, optimizes their category as well. As we approach a market, for instance, with multiple categories, the chance for success for each of them increases.” Going into a new emerging market with several complementary categories, rather than just one, for instance, can enable cost sharing and increase local influence, thereby increasing the chance of success in the region.

The dilemma is that when the challenges facing an organization are not about repeatable execution, but about innovation or responding to complexity, the idea of breaking things down into well-understood parts is not only unhelpful, it can also be a dangerous trap.