A month ago Digg launched the ‘Digg Recommendation Engine’, on a trial basis, and made it available to some of Digg’s registered users. Now they have found that the Recommendation Engine was very successful in driving up user activity and social interaction.
The Recommendation Engine is a suggestion tool that shows users new content on the site that they may be interested in. The Digg technicians found that it was not an easy task for users to find stories that would be of importance to them, as the ‘Upcoming’ section of Digg now receives an input of over 16,000 new articles on a daily basis.
Through the services of the Recommendation Engine, Digg recommends articles that a particular user may be interested in, based on records of the user’s past Digging activity. In other words, the type of stories that a person has been inclined to Digg in the past are more likely to be shown.
Users are categorised into groups of “Diggers like you”. This enables users to find other like-minded Diggers and then send stories to one another. The list of ‘Diggers like you’ is visible in the right hand column on the page, along with a compatibility rating.
Now that this trial has been going on for a month, Digg has released some stats about the trial, which are very encouraging. They find that Digging activity has gone up by about 40 percent since the launch of the Recommendation Engine.
The Recommendation Engine generates about 54 million recommendations at any given time, with each user receiving nearly 200 recommendations, from an average of 34 people in that group of ‘Diggers like you’.
Digg reports that interaction between friends has also gone up, by nearly 24 percent, and the comments received on the site have gone up by 11 percent.
Based on feedback received by Digg, they will be making some further improvements in the format by including more stories in the Recommendation Engine widget, on the homepage, lengthening the time frame of the Digging window from its current setting of 30 days and making changes in the back-end algorithm to make better recommendations to users.