Kevin Rose Presents
Behind the scenes at Digg.com.
How news breaks at Digg. 600,000 registered editors looking for news and digging them.
Rumsfield news broke a record. Saw it hit the AP and 3 or 4 minutes later it hit the Digg web page. Google News did not pick it up for another 20 some minutes.
How can we further enable users to break news?
33% of the Rumsfield diggs occurred in Swarm. Stories appear as a little bubble. As more users digg a story, the bubble grows. Rumsfield swarm grew to a gigantic bubble. Multiple stories attach themselves to each other.
They are working to give these tools to web publishers. Will be launching a Flash toolkit to let individuals do this on their own.
Gaining
Users are attempting to gain the system. Users try to get their stories noticed. Digg does things to prevent this but still let the community contribute. They are letting the users control the spam. Also, they are looking at user behavior as stories are "dugg". Looking at patterns of healthy stories: referrer, upcoming stories, 90% no referrer detection. Kevin displayed a chart of a story's life with strange behavior compared to a baseline. All the same IP address, same user digging the same story, etc.
Another effort: They are detecting prescient users: users who can forecast trends.
Another version coming out next month.
How news breaks at Digg. 600,000 registered editors looking for news and digging them.
Rumsfield news broke a record. Saw it hit the AP and 3 or 4 minutes later it hit the Digg web page. Google News did not pick it up for another 20 some minutes.
How can we further enable users to break news?
33% of the Rumsfield diggs occurred in Swarm. Stories appear as a little bubble. As more users digg a story, the bubble grows. Rumsfield swarm grew to a gigantic bubble. Multiple stories attach themselves to each other.
They are working to give these tools to web publishers. Will be launching a Flash toolkit to let individuals do this on their own.
Gaining
Users are attempting to gain the system. Users try to get their stories noticed. Digg does things to prevent this but still let the community contribute. They are letting the users control the spam. Also, they are looking at user behavior as stories are "dugg". Looking at patterns of healthy stories: referrer, upcoming stories, 90% no referrer detection. Kevin displayed a chart of a story's life with strange behavior compared to a baseline. All the same IP address, same user digging the same story, etc.
Another effort: They are detecting prescient users: users who can forecast trends.
Another version coming out next month.
posted by Mitchell Starnes | 2:19 PM
0 Comments:
Post a Comment
<< Home