Kevin Rose resigning from Digg stirs up a lot of fond memories of Digging great content, but also stirs up some painful memories of the worst software roll out we had ever witnessed. There's a lot of speculation and debate about what killed Digg, from pressure to monetize to bickering within Digg's own ruling party, to the concept of power users getting all of the lime light. What killed Digg was none of these things, because most of them still exist the same way they existed two or more years ago. Powers users like Blinker are still all over the front page, and the advertisements are no more intrusive than they had ever been. What killed Digg was entirely software based.
In many ways it should be applauded that Digg tried something so brazen. Digg really tried to capture the web 2.0 feel. If you spend a few minutes visiting the new Digg you'll see lots of ajax-y calls to update information, dynamic things that slide in and out from all over the place, in place comment editing, and other user interface niceties. Couple that with a more Twitter feel where Digg put a bigger emphasis on acquiring followers and following others. All in all, the changes were positive changes that really pushed for usability. It's not the UI that went wrong. Most of the UI changes worked well on day one with the exception of some over looked things like breaking the back button, most of the over looked pitfalls were quickly patched up almost in real time as the community complained.
So, if I'm going on and on about how great the UI changes were, when am I going to say what killed Digg? What killed Digg was a stupid mistake that should have never been made by a responsible development team. The worst part is, what killed Digg was only temporary and what killed Digg is already gone. It wasn't the UI changes, it wasn't the move to imitate Twitter, it wasn't even the automatic RSS spam feeds that were later removed. It was constant crashing and poor performance by a system that was not properly stress tested.
There was lots of finger pointing when it happened. He said that she said. This person made this decision, that other person made that decision. It was the move to use NoSQL, it was the chief engineers, it was Kevin Rose bending over for advertisers. No, yes, I mean maybe, no, it was all of this and none of it. It was a lack of testing. If you visit Digg today you'll see that 99% chance it's going to work and perform well, so it's not the technology stack. The fine folks at Digg, for whatever reason, decided to push out an untested monstrosity of a system live without properly stress testing.
Sure there was a beta people could access, but if you were on Digg at the time, we never heard too much complaining from the beta users. No, the changes were fine.
Digg died because it wrote a whole new code base, pushed it live, and then watched in terror as "Digg broke an axle" for droves of users. You might want to make an argument against that statement, but any other complaints people had were fixable and to this day have mostly been patched. Sure the content on Digg is suffering these days, but that might have more to do with an exodus of its best users more than any system change. The moral is, whether you're a big site or a small site, don't push out software that crashes. Crashing buggy software erodes user confidence. That's why Digg died, from my perspective.