A while back someone, probably a PR guy, came up with the idea that there were a class of people called influentials. From what I can remember, these folks had larger social networks than everyone else and/or people actually trusted them more. So the general idea was to find these guys and then somehow do advertising through them. I think Google just came out with some program to find these guys. The thing is, people don't believe in the theory anymore. The folks with the large social networks don't appear to be anymore influential than anyone else.
But perhaps all is not lost. Watch Clay Shirky talk about the 80/20 and think about influential model again. Maybe the size of the social network is less important than the sustained output of the person. I bought the kind of mp3 player that I did because I listen to linux podcasters. One actually records his podcast on it. Does it make sense from an advertising standpoint to work with a guy like Dave? I don't know, but you'd sure need to have a good product.
Of course, we could shuck the direct route, ignore social networks, output, and come up with a theory that includes a strange attractor component. Seriously cool wackiness, especially when applied to human interaction.
But then, I don't think there necessarily should be only one anyway. If you can get two or more people whose opinion I think relevant to actually like your product, you've a much bigger chance of getting a sale than if you plaster a billion dollars worth of ads all over the place.
So maybe it's not about one influential either. Maybe it's about the truth, and the fact that several people will verify each other's account when a product's really good. Makes the advertiser's job a little harder, doesn't it?