Brainnovate

For the longest time, I have been a fan of Digg, but never a big time user. I first signed up almost two years ago, and since that time, I have Dugg a few things here and there. But I never really got into it until the last couple weeks.

I am not sure what happened to be honest. My increased interest coincided with the launch of their new “recommendation engine” technology.  Maybe the release of this tool lit some inner spark, because I watched the videos, and immediately went on the site to play. Before I knew it, I was logging hundreds of Diggs, and reading article after article on how best to use (and benefit) from Digg.

Actually, I have always been a fan of “optimization” algorithms.  Over at Vertster we use something called the Taguchi Method to accomplish some pretty cool multivariate testing, so now you know why.

Anyways, the Digg recommendation engine is particularly interesting because on the surface it seems to give you more of a chance at getting noticed if you suffer from low Digg authority (ie you haven’t been doing it for long.)  However, in reality, it adds another layer of “black box” to the site- what we call it at Vertster when you have an unknown algorithm and all you can see are the inputs and outputs but have no idea what goes on in the middle.

On the surface it appears to give Digg another chance at separating the wheat from the chaff so to speak.  By analyzing a number of signals, the algorithm could display certain submissions more frequently than others.  It also adds a fertile new area that should earn more pageviews than the old standard upcoming page, which was nearly impossible to get noticed on without help from friends.

So what are these signals likely to be?  Well I am sure there are some very smart Diggers who are probably trying to decode this right now… but it will be tough because its likely to be a moving target.   Some of the factors I would expect to be included and weighted heavily would be the Digg authority of the submitter, and also that of the browser.

Of course the contextual matching engine is also at work- the part of the software that tries to show you stuff you are likely to be interested in.  But also, I bet there are other things like time of day, day of week, and current events.  These “moving targets” will make it difficult to decode the exact algorithm and should improve its effectiveness at showing sticky content.