Behavior Adoption on Social Networks: Two Models of Viral Growth
20bits.com
20bits.com
I don't know about "oversimplification," but it is a model. It's not meant to explain everything. :)
One thing I wanted to say is that the viral model does a really good job at modeling user growth over time and the threshold model not so much.
I feel like the threshold model is more useful if used to model an increase in likelyhood of viral conversion based on penetration into their social graph. (i.e. users with more friends active are more likely to convert from any given organic/viral message they receive).
Yeah, I don't have a clear idea of how these two models relate to the things one would typically measure in "the industry." That's one thing I'm trying to figure out because I believe fields like SNA, sociophysics, complex systems, etc. can help us understand how to build better, more interesting, and more sustainable social networks -- but AFAIK there's no communication between the people forging new social networks and the people researching them.
But I do know this: Facebook has focused on density from very early on, which is a key factor in the threshold model. Their strategy has always been depth first, as opposed to many FB apps, which I think are breadth first. That is, they go viral, but they spread far and wide without spending time building sufficient density to become self-sustainable.
There's some formula like engagement = density*activity buried in there that I can't quite pull out.