Interesting, but I would be more interested in causal relationships (I.e. growth rate or engagement delta). I think it's quite possible that social merely segme

Interesting, but I would be more interested in causal relationships (I.e. growth rate or engagement delta). I think it's quite possible that social merely segments users by their already existing commitment levels: anonymous users have no commitment, social users want to make their commitment painless.