Now that we the people have taken over the Internet, relying mainly on the free services of social networks, these networks will start mining all of our usage data (how many times did ‘the people’ say “lost” yesterday) to find a new profit engine. “Search” seems to be the current flavor. Google did it and are doing well.

Let’s take Twitter. Their search difference-maker is the natural data of their system; followers and friends. Currently I’ll search for “lost” and I can see who is talking about it. NOW simply mix in the raw follow/friend data and I can see what the most popular folks are saying about “lost” (there’s a NYTimes piece I have to dig up that discussed how influential it is to a person’s own decisions to see what other people prefer). This mix is the beginning of a “reputation rank.” Trick is it takes more that just your friend/follower count to make you legit. It takes an analysis of who friend/follower peeps to see just how legit you are and thus your rank.

The simple answer is to leave it up to the robots to do the math. The problem is you’ll have people game the system like SEO. The answer is to give the users control of the dials. Build the reputation system but let me the user choose the variables. Sure, I can choose to use an out-of-the-box default setting, but then let me switch it up.

I don’t have access to the available data, so let’s pretend…

I want to see who is talking about “lost” BUT I want the results to be by location (Arizona and Maine) AND frequency of updates (user posts 5+ times a day). I get the results, I want more. Next I want to see the results by filtering friend/follow ratio of 80% and higher. Boom, new results. I define the reputation of my fellow users in the way that makes sense to me. Maybe I can save my reputation-recipes to a control panel so I can use them again, maybe I can share them in tweets with my posse?

I’m excited to see how it all goes AND if it makes any money for anyone.