Last week, Neil Patel reported that Digg is making it harder to get Dugg. I can certainly vouch for that. An ex-client and friend of ours launched what was (at least in my opinion) a highly Digg-worthy article on the Ten Servers that Changed the World. We didn't do any work on the project, but Corey Donovan from Vibrant (whose intelligence and positive attitude made him a pleasure to work with) asked for our input and, for our part, the SEOmoz crew thought it was a sure-fire Digg winner. Corey went so far as to recruit help from the Computer History Museum in Mountain View to make the piece the best it could be.

Matt wrote the Digg submission piece and it had some initial pickup, gaining around 34 Diggs in the first 24 hours. While this would normally put the piece in front of a lot of Diggers (thanks to both the friend system and the constant appearances on the upcoming page), we were surprised to see that the Diggs slowed down, and 3 days later, there's only 43 Diggs.

Now, the piece still has the opportunity to play at Slashdot, Del.icio.us, Netscape & possibly even Reddit, but the Digg failure certainly made us take note. SEOmoz's usual success rate with pieces we believe have a high liklihood of getting Dugg currently stands at about 70%. We've launched around ~50 efforts in the past six months and had approx. 35 of those stories make the frontpage. Honestly, I don't know how well this stacks up to other marketers who use Digg as a tool. Certainly, if you wanted to go black hat, you'd have a fairly easy time either buying the Diggs, setting up 4 or 5 dozen extra Digg accounts through proxies or leveraging a large network of friends (all tactics that we haven't engaged in).

To help learn from this experience, I thought I'd explore what signals might be in the Digg algorithm and how the Digg operators could be tweaking these:

  • Number of Diggs in a time period
  • Number of independent members voting the story up
  • Number of new members vs. experienced members
  • Number of votes off "upcoming" page
  • Number of votes off "friends" pages
  • Number of votes off the "Dugg" page (through an embedded "Digg this" button)
  • Temporal pattern matching
  • Up votes vs. down votes
  • Number of comments and comments over time
  • Number of Diggs vs. area of content focus
  • Manual review by admin/moderation team

There might be more that I'm unaware of, but these would be the ones I'd operate with. Using these signals, restricting spam and promoting higher quality stories (Digg's two primary goals) could be achieved. Lots of folks complain about the quality of what shows up on Digg, and there's obviously a huge topical bias, but Digg still has the largest, most achievable audience for linkbait, and will remain popular for marketers for the foreseeable future.

How have your successes and failures gone over at Digg? What's your feeling on the tightening of the algo to reduce the number of frontpage stories?

p.s. Full Disclosure - Vibrant Servers was a client of ours up until a few months ago. We may work with them again in the future (fingers crossed, since I'm a big fan).

p.p.s. Cristian Mezei also did a great piece on Digg's algo last week, noting many of the same points I have (his analysis is actually even more in-depth).