A month ago, I suggested that with investment capital scarce, social media sites will be forced to close their doors and, as a consequence, users will lose huge amounts of social and cultural capital.
This week demonstrates another, highly efficient path to datapocalypse: data loss. In the most spectacular failure to date, ma.gnolia.com has all but declared the loss of its social bookmarks permanent.
That was only one user losing his data, so it wasn’t a big deal—although it demonstrated the dangers of relying on social software—but backups seem to be a tremendous problem for these services. They effectively double the cost of storage, so corporations seek to cut corners. On the other hand, since we backup religiously at home (you do, don’t you?), we expect that the online services we’d use do as well.
If ma.gnolia can’t recover its users’s data, I don’t see how they can come back from such a loss and even if they do, it seems like their run is over.
But I also predict that in the near future we’ll see a high profile failure or closure that will dwarf the loss that the users of ma.gnolia experienced. I still wonder, for example, how Facebook can survive given the trouble it seems to be having paying its bills. "It’s all about eyeballs" is something that the doe-eyed MBAs used to say a decade ago, but isn’t going to hold any credence in the more mature network culture of our day.
Maybe, though, this will give a boost to develop a counter-cloud, like the one that I proposed earlier this year, in which social media functions would be spread across a multitude of Web sites running common platforms (think Drupal for example). Take a look at this article by Brian Suda, for example, that explains how to carry social networking relationships between sites. It’s a great start, but it needs to be made easy to do for the non-technically savvy user and it would have to be possible to scatter data across sites, with redundancy built into the system to make it work. It may sound impossible, but so in principle, so does bittorrent, but it still works. Like the move from Napster to Bittorrent, this would be a move from centralized to decentralized systems, dispersing control and responsibility into a DIY counter-cloud. Something for Drupal version 10 maybe?
I suspect that some version of my modest proposal is going to be enacted in the next decade. Trusting one site for anything is a big problem, as the users of ma.gnolia found out.