Centralization vs. innovation?

I've worked for:

• a large library automation consortium
• a county library IT department
• a state library

Recently, I've been working with committees for:

• a library automation consortium
• a statewide portal for state agencies/egovernment

Something that's ALWAYS bothered me, that I mull a lot, is the power of centralized vs. decentralized when building systems. When I worked for a library automation consortium, I was convinced that increasing the amount of centralization would enhance services. When I began working for a county system that ran its own automation platform, I began to notice how much more quickly we could react to the changes in technology than could a large consortium.

As Web2.0 became the trend, I noticed how some of the smallest, least assuming libraries managed to extend services or create more sophisticated web experiences than the big libraries. I wondered if the innovation happened closest to the ground, where the people building the systems were directly in touch with the experiences of the end users, good or bad.

I know that the architecture of the web is an enshrinement of the principle that decentralization ensures progress, innovation, and sustainability. Should a search engine create a better algorithm, the other search engines will try to improve on it. Should a node go out, that's fine, the net routes around it. There is no single point of failure. With Web2.0 the number of communication channels have increased. News often travels most quickly online via social networking sites - Twitter & Facebook, for example. So this argues for the power of a seemingly organic method of growing systems.

Governments build things in a more official, hierarchical, and centralized way. As a result, innovation is often stifled.

On the other hand, without centralization, we are constantly reinventing the wheel and duplicating efforts, are we not? Aren't there economies of scale in having everyone pool their money and bargain with the big vendors for the best possible setup that would cover all of the members of a project (whether we're talking about groups of governmental agencies or library consortia)? Wouldn't it be great if there were standards that users of multiple libraries or governmental agencies could count on when dealing with those entities?

An additional foil to a more tightly regulated, centralized system being the answer for library or governmental systems is the concept of monopolies of power. If the system is fully centralized, the power lies all in one place. There are no competitors. There is no ability to seek alternatives to the Dept. of Revenue Services, for example. In time, the relationships between the agency or library system and the customer grow strained. The customer has no options, no way out if the agency or library system isn't giving them satisfaction. It creates a very standardized experience, but not a user-centered experience.

But maybe that's part of the issue. If we could just center on the users' needs, maybe we would know how to build systems and services more effectively. The greatest successes happen where the maker's vision intersects with the users' needs & desires. That's where both libraries and governmental agencies want to go.

I'm open to your responses to this question of centralization vs. decentralization of systems and services. I've been wanting to post it for a long time and every time I envisioned writing about the issue, I was much more eloquent than this. But it's time to stop worrying about that and just get the discussion going.