In the past we’ve said how much we enjoy StumbleUpon. It’s a great way to find new sites, and it’s a definitely a great way to pass the time by when you’re bored or have a few minutes to spare. We’ve even used it to find new tips and sites to share here on CyberNet. It has always worked like a charm, so we’ve never really put much thought into how it works. Recently VentureBeat got in touch with the co-founder of StumbleUpon who then explained all about how StumbleUpon works and it’s actually pretty interesting. If you want the complete scoop, definitely checkout the VentureBeat article, otherwise we’ll be summing it up and throwing in a few tidbits of info we found on the StumbleUpon site for those of you who want the quick version.




So how is it that StumbleUpon works? There are three main “engines” that keep StumbleUpon running:

  1. Classification Engine
  2. Clustering Engine
  3. Recommendation engine

Each of these engines work together to provide users with a “channel-surfing” experience on the web that, as they say, “automates the word-of-mouth” referral of peer-approved websites and simplifies web navigation.”

Below is an image that SU put together which shows how the process works (click to enlarge):

stumbleupon technology.png

When somebody gives a thumbs-up to a page for the first time, the engines get to work. It’s sent to both the Classification Engine and then the Clustering Engine. The Classification Engine either looks at the tags and the topic that the submitter chose for the page or if that information wasn’t provided, the engine will decide how the page should be classified. The Clustering Engine is working hard too by keeping a watchful eye on on which sites are getting votes so that they know which sites are quality and should have more traffic sent their way. The Recommendation Engine is important too because this is what keeps track of the specific topics that interest users, socially endorsed pages, and peer endorsed pages.

Between the three engines, the process of collecting and sharing information with millions of people is automated and then allows for new content discovery.

Any StumbleUpon fans out there?