Sunday, April 15, 2007

Is Digg the Result of Cumulative Advantage?

April 15, 2007 — 05:13 PM PDT — by Pete Cashmore

The NYTimes has a fascinating piece today about how the “rich get richer”, or popular media gets more popular. In other words, things rise to the top not because they are better quality than the alternatives, but because people copy what their friends do: a tiny rise in popularity an early stage can mean massive popularity further down the line.

This has some really obvious applications in social media. Digg is the premier example: its “network of friends” system inevitably results in users Digging what their friends Digg, often blindly. So while quality stories still have a marginally better chance of rising to the top, this “follow the leader” effect means that users are more likely to amplify the decisions of other users than go against them, even if the stories being Dugg aren’t very good. Digg could prevent that by removing Digg counts and friend networks entirely, but that would counteract its own aims: growing as quickly as possible so it can report huge user numbers. To paraphrase the butterfly effect: one 13 year-old in Illinois can decide whether a news story becomes the most popular item of the day, or falls into obscurity.

But the theory has much deeper consequences when it comes to the success or failure of startups themselves. We love to think that there’s some kind of magic formula for the perfect social site, but the results seem far more random: if you rerun history, it could turn out that a whole different set of startups rise to the top. That’s because the first few users influenced the final outcome of those startups, and as soon as one site hits “critical mass”, everybody gravitates towards that site. So imagine a world in which Reddit had a few thousand more influential users than Digg: it may have won in the long term. This theory also tells us that Digg will never hit the mainstream: it is so heavily seeded with geeks that it will continue to attract that demographic and alienate non-geeks.