There’s been a lot of chatter lately about social recommendations at MetaBroadcast Towers, particularly at the recent #MetaBeerTalks. I remember back in 2005 when I joined Flickr and Last.fm it seemed that open APIs were everywhere, we couldn’t wait to give our data away and The Internet would do some magic and everything would be solved soon. But nearly a decade’s past and while social media has gone mainstream it’s now all about walled gardens where everyone guards their data jealously. The panacea of social recommendations just never came to pass.
Perhaps there was another difference at work here too… Last.fm, Flickr, OKCupid, as well as being from the wild fronteer of the internet were also social sites built around data instead of the other way around: Flickr was the pre-eminent site for uploading, tagging and sharing photos, people went to great lengths to make sure we could scrobble all our tracks to last.fm which gave us back nifty charts and graphs for our trouble, OKCupid had endless questionnaires that offered better and better matches the more questions you answered, and they also do some really interesting analysis on that wealth of data. In each case providing data was a means of self expression, you got tangible results back, and the data came first, the social later.
Compare that to facebook or twitter; you can get interesting stats on your twitter account or your facebook page but it’s set up specifically for advertisers and pretty wishy washy at anything but the macro scale. Far from users getting something for their data these companies try and pretend there’s no data at all, meanwhile we get the uncomfortable feeling that the system is against us as we’re interrupted by promoted tweets or thrown again and again back to the dubious algorithms of “top stories”. Far from being keen to deliver our data and see what fun stuff the system does with it we’re advised to “unlike” everything to stop your feed being a wall of irrelevant adverts.
These current behemoths of online were social first and foremost. Data was reluctantly and surreptitiously tacked-on afterwards and it’s never really been a great fit. The sad thing is they’re probably poisoning the well for others too; were it not for the privacy-based paranoia they foster, streaming companies like Netflix or iPlayer might be free to do great things by adding some old-school data-first social magic to the mix. In the current social atmosphere I’m not sure I’ve seen the system that’s going to solve social recommendation for video, but if you’re already using something that gives you decent recommendations, drop us a tweet or a comment!