As the web continues to get older, more and more data is being stored across millions of different sources. Brands like Amazon and iTunes have been using this data to help create personalized recommendations for years. The classic "people who bought this also bought..." has become a staple of many sites - mainly because proper recommendation tools can generate a dramatic lift in sales and add value to the users themselves. It's quite a simple idea but one that takes a powerful algorithm to get it right. Netflix, as another example, has spent millions trying to perfect their recommendation engine because they know the powerful results associated with it.
What interests me isn't just the data and common-links that can be associated with different products, but they way that we can begin to connect with people we've never met via their data trail (and Social Graphs).
I recently posted about how eHarmony and other dating sites are using recommendation tools for people - fill out an extensive profile and they will provide you with your matches. While this process is has worked for some people, I find it very rigorous and demanding for a user (even if they are a bit desperate to find a partner). I do think, though, that with tools like Facebook connect being implemented across thousands of sites, friend recommendation engines could start to become more and more popular.
What will truly be interesting, though, is how you go about contacting someone who has been recommended to you by a system. It's not like your at a bar and you strike up a conversation, you've got to send a message to your recommended friend and hope that they respond.
What do you think? Do you use the engine on Facebook right now to add new friends or do you feel like you have way to many as it is?