Meandering thoughts on social search

On my drive into work on Monday, my mind was filled not with thoughts of Storage Resource Management but rather Social Search.  Google recently made some inroads into this area, but I feel like we’re on the cusp of something revolutionary and nobody is seizing the opportunity to change the game.

Everything I am about to describe is achievable with today’s technology.  And yet it sounds like science fiction.  Here’s the world I want to live in.

Somewhere, an agent acting on my behalf understands my networks.  My networks overlap, are complex, and come from distinct sources, but I believe this is a solvable problem.  Ideally this requires my agent to access my RSS subscriptions, Twitter, Facebook, LinkedIn, and even my email on my behalf – which means I better have ironclad trust in this agent.  With this network aggregation built, it’s easy to locate content which is produced, recommended, or subscribed to by my network.  The aggregation of all this content should be the first tier for my search results, and should set the context for resolving any ambiguities in my searching.

This though is a specific instance of a more general problem.  The importance of a source in recommending content into my search results is not binary.  Within that network I defined earlier, you can imagine a ranking of sources based on frequency of contact, number of networks in which the individual appears, and number of times I follow recommendations from that individual.  And once you have that, there’s no reason you can’t extend it beyond my direct network.  Even if I don’t subscribe to a specific industry blog, if ten of my work contacts do and that blog has a relevant search result, I want to see that search result before some unknown source.

Suddenly my agent is doing a lot more than compiling my network.  You could for example implement this with a Google Page Rank which is unique to each user of the system.  Suddenly SEO gets a lot more (and less) interesting, and Google (or Facebook) needs a ton more storage.

Once I have this in place, there’s no reason my agent can’t tell me when my network is too small, or needs pruning.  It knows people I should be following, blogs I should be reading, and supposed friends who I probably don’t even really know (and who in fact might be poisoning my search results).  It can suggest ways to change all these things, and my responses and actions will help shape its behavior.

We started by talking about social search, but suddenly we’re talking about social management.  I don’t want to just search using this information, I want to browse and explore.  We’re leaving a ton of data on the table.  You have to imagine someone out there is just itching to figure out how to help us use it, and in the process learn some incredibly detailed things about us all (someone has to pay the bills, right?  May as well be laser-focused ads….).