[nrmehta]

Thought-provoking post. To me, one way to determine whether a system will have network effects or anti-network effects is to ascertain how much of its usage is driven by fashion versus utility. Take email as an extreme example of the latter. It's valuable because it's so universal - but it's not fashionable at all. It's a pure utility. So no anti-network effects (perhaps beyond spam but those are less about #s of participants as behavior). I put Facebook in an intermediate category where it's transitioned reasonably well from fashion to a utility, though the folks that looked at it as fashion are now getting more turned off by it. Indeed, the anti-network effect isn't simply about numbers - it's about who is coming into the network and a lost feeling of exclusivity (which honestly sometimes picks on very base human emotions) when the network grows with certain types of people. I think twitter has moved further up the utility value chain than facebook has so I'd posit it's less vulnerable to anti-network effects (not to mention the asymmetric follow model that dalton talks about).