I joined Facebook in 2004 and Twitter in 2007. Over the years, I built up a lot of connections on those networks, and I get a lot of value out of at least one of them. But with all the friends and followers and followed comes noise. Much as I've tried to tame my network, it feels more like a hydra or the carnivorous plant from Little Shop of Horrors than the calm and orderly information drag net that I thought I was weaving.

Somehow, though, the idea of unfollowing or unfriending everyone and starting fresh seemed impossible to execute. It might take a long time, I told myself. I might hurt feelings, too. And my Twitter network, at least, is still pretty functional, more feral than wild. So I've been stuck, unable to restart my networks, but unable to manage the information coursing through them. I needed a greenfield in which to grow a different network.

This is one reason that I've loved Instagram, the iPhone-only photosharing app. I decided on a whim early on that I'd only connect with people I knew in real life, so the network that sprouted there is small and manageable. I keep up with my personal friends' travels and dinners and that all feels good. It's a niche of my life.

Yesterday on Google+ when a small group of friends were discussing the merits of the system, tech journalist Chris Mims pointed out this is a key advantage of Google's new social service. "Also it's a chance for a restart," he wrote. "I'm ignoring everyone whose name I don't recognize, simple as that. Some types of networks have a value in direct proportion to their selectivity."

I have an intuition that the network effect can sometimes reverse itself when people decide by change of life circumstance or growth or chance that their bonds enable a kind of bondage. The cry, "But my whole life is there!" can cut both ways. Americans love hitting the open road and getting a fresh start, and I wouldn't be surprised if we extended our love for residential mobility to the online domain.

People act as if the negative impacts of leaving a social network are just SO devastating that no one will do it, but people move to new homes constantly and the dislocations that result from moving across town, let alone across the country, are far deeper than switching from Facebook to Google+.

I want to propose a new kind of network effect, a bending back of the value of the network when you have too many connections. Some network connections have a negative effect on the value you perceive, though they may be difficult to pinpoint; and other negative effects emerge from the sheer quantity of connections. Plus, with every social net I build, I'm getting better at doing it; I'm learning.

Here's a sketch of how I've started to think of things, open to revision, and presented merely as illustration, not law*:

socialnetworkeffect.jpg

We're not laying the first national telephone system here. In an age with many social networks -- and the ability to import contacts from many places -- maybe we need to revise what the network effect looks like. New networks do not start at 0 or grow rapidly in value, and older networks do not keep increasing in value as they add users or as users add connections.

*Samuel Arbesman pointed out on Twitter that Metcalfe's Law deals with "overall connectivity not individual ties." I thought of plotting that here -- and would suggest the general shape of the graph might hold -- but I want to emphasize that I think individual's connections are a better measure of how much value you're getting out of the network.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.