"better data can beat better algorithms" well said.
Scribd is another good case study of how to present data to rank high in searches. Could this be that the view of documents is different to a crawler than one presented to a normal browser?
You can get a pretty good idea by using Links, w3m, or another text-based browser. You can also use Google Cache's 'text-only version' option in the upper right. Here's a random Scribd page:
We do this at TrailBehind too. The normal site is a huge Google map with lots of waypoints. The javascript disabled site, that google bot sees, is a google optimized list of links to pages about each waypoint.
The problem with sites like LinkedIn is that they own your name and can sell it back to you. When you use them they own your contacts as well and will stop you from contacting your peers when you don't pay (LinkedIn competitor Xing already did that to some users).
Before we reach the age of data portability or proper ownership of data you have to limit the power of third parties over your name. You have to use those services or someone else might do for you unofficially but you have to keep your name and contacts to yourself.
So basically you need a good SEO and CRM strategy where you own your name and contacts yourself independently of services like LinkedIn or Facebook.
Scribd is another good case study of how to present data to rank high in searches. Could this be that the view of documents is different to a crawler than one presented to a normal browser?