The network effect says that a person that's connected to five family members is not as (measurably) valuable as someone connected to fifty people.

If you were Google, which would you rather sell to advertisers, correlations between groups of five people of fifty?

Hmm, is this backed by data?

I mean that is it known that in current social networks a life-time value of a person that with 5 strong relationship ties is less valuable over time than a person who has 50 very very weak ties (like most of my Google+ followers, people who I don't know)

My conjecture on top of the general value of a more connected network.

What is Facebook's motivation for connecting people? Or Google's? Or LinkedIn's.

A newspaper sold millions of individual pairs of eyeballs to advertisers. I would think the large numbers of correlations available in a largely connected network would be much more valuable than unconnectable individuals.

But stronger ties could be more "sticky", i.e. it's harder to leave Facebook than Twitter because people (in general) have stronger ties to people they are connected in Facebook than in Twitter. Thus, life-time value of a stronger tie for the service provider is larger. This seems common sense, but I'm interested in if there's any kind of data what the ratio is. Inside a single service, of course, as values are not directly comparable over the services due to different business models)

> If you were Google, which would you rather sell to advertisers, correlations between groups of five people of fifty?

I would rather sell a network that people want to use over one that they don't.

But what if you don't have that network?