Digg’s Design Dilemma

This past week’s Digg controversy is one in a growing number of incidents that suggest that a small group of users are having an undue influence on the promotion of stories. In response, Digg is changing the way that it handles votes by adding complexity to its ranking algorithm. I think that’s the wrong approach, so here’s another idea: change the actual design of the site…that’s the real problem.

The most recent controversy happened on September 5th, when someone named jesusphreak posted Digg the Rigged?, an in-depth article exposing some of the curious details of recently-popular stories on digg. Many of the stories, jp pointed out, were dugg by members of the Digg Top 30, or the 30 most popular digg members (popular being measured by number of stories submitted that were promoted to the frontpage). The Top 30 includes Digg founder Kevin Rose.

This was not the first time that someone has pointed out this phenomenon. On April 18 of this year Macgyver at ForeverGeek posted Digg Army, which included screenshots of who dugg two recent articles on the site. Each article had the exact same 16 people digging it in the exact same order. Of the first 19, 18 were the same. Included in that list of people was, again, Kevin Rose. ( for an in-depth history see Tony Hung’s excellent: A Brief History of the Digg Controversy)

These incidents, taken together, are more than coincidence…

This past week’s Digg controversy is one in a growing number of incidents that suggest that a small group of users are having an undue influence on the promotion of stories. In response, Digg is changing the way that it handles votes by adding complexity to its ranking algorithm. I think that’s the wrong approach, so here’s another idea: change the actual design of the site…that’s the real problem.

The most recent controversy happened on September 5th, when someone named jesusphreak posted Digg the Rigged?, an in-depth article exposing some of the curious details of recently-popular stories on digg. Many of the stories, jp pointed out, were dugg by members of the Digg Top 30, or the 30 most popular digg members (popular being measured by number of stories submitted that were promoted to the frontpage). The Top 30 includes Digg founder Kevin Rose.

This was not the first time that someone has pointed out this phenomenon. On April 18 of this year Macgyver at ForeverGeek posted Digg Army, which included screenshots of who dugg two recent articles on the site. Each article had the exact same 16 people digging it in the exact same order. Of the first 19, 18 were the same. Included in that list of people was, again, Kevin Rose. ( for an in-depth history see Tony Hung’s excellent: A Brief History of the Digg Controversy)

These incidents, taken together, are more than coincidence. They strongly suggest that Digg is being gamed by a small number of users, artificially inflating the value of stories that might not deserve such attention. This flies in the face of the democratic ideal of the site. And so far, nobody has claimed that the two articles I mentioned are false: Digg exposes most of the voting activity for all to see. A small group of users is definitely having a large effect on popular stories.

But before we get outraged at the corruption of it all, we should give everyone involved the benefit of the doubt and consider how this might have happened without evil influences.

Don’t blame the users

The users of Digg aren’t to blame. They’re simply playing by the rules as outlined by what they can and cannot do on the site. They’ve figured out how to play in the environment they find themselves in, and that’s OK. Jason Calacanis, creator of digg-clone Netscape.com, said in his post One User, One Vote: “The top users earned their spot and they should be reward for their contributions–not penalized.”. I agree with that…it’s not like there were any rules to follow.

Blame the design

Instead of blaming the users, blame the design of the site. From the ranking system, to the friends feature, to the display of content, to the ease with which users vote, the design of Digg.com conspires to make it haven for gaming. Not only is the pile-on digging activity possible on the site, it is actually enhanced and made easier by the very design of it!

Here are the features in question:

  • Rankings list
    If you want people to compete, rank them. This is a big part of the reason why there is gaming on Digg. Getting a higher ranking becomes an incentive to game because if you do then others will notice and you’ll get recognition. (that’s important to people, even in social software) In addition, with the recent offer by afformentioned Calacanis to pay people for this type of work, high rankings may also be a source of income.
  • Friends feature
    The Digg friends feature is the means by which the top users promote stories so quickly and with such synchronicity. In particular, the friend’s history page shows the stories that friends have dugg in reverse-chronological order, so that the newest diggs are at the top of the page. By refreshing this page often, top Digg users (who are all friends in the system) can stay up-to-the-moment with each other. During waking hours, a quick 30 diggs will draw further attention to any story, making a pile-on more likely.
  • Exposing who diggs what
    At the bottom of each dugg entry is a list of people who have dugg it, and serves as the evidence that the two articles above used to expose the issue going on at Digg. This is a perfect example of what in psychology is called “social proof”. Social proof is something that is “proved by society”. In other words, the mere fact that others are doing makes it seem like it is what should be done. We learn that way, by mimicking the actions of others. So, when we see someone else digging something, we would be much more likely to digg it ourselves. Or, to put it another way, we let others influence our decisions and help make them for us.
  • Stories at a distance
    It is very possible to interact on Digg, digging stories and burying others, without actually reading a story. That’s because Digg only shows summaries of posts. If you want to read a post, you actually have to click on them and go to the external site to do so. Many people will make this extra effort. But many people won’t.

    In It’s all a Farce Anyway Tara Hunt recounts an interesting (and scary) conversation with people who game digg. They submit stories and ask their friends to digg them. After a post reaches a certain digg count they say: “people just hit digg if they are remotely interested in the topic.”. This, again, is the effect of social proof, exacerbated because the stories are at a distance and it is extra effort to read them.

  • Ease of voting
    While it takes extra effort to read posts, it takes almost no effort to digg them. This might be backward…digg is essentially making it possible to vote without knowing what you’re voting on. Although the digg feature is amazing, an excellent example of technology that makes our lives easier, it is also in danger of trivializing them.

    This reminds me of a story by Derek Powazek in his book Design for Community, where he makes the point that the harder it is for someone to comment on something, the better the comments are. In other words, people who jump through hoops (or pay attention long enough) to comment are the ones who really care about the subject matter, they’re invested in the story and see value in taking the time to respond.

Other factors

The Digg community is protective. Stories that are about digg get a lot of positive attention there. If you want to get noticed, for example, simply write a post entitled “10 Ways to Get Dugg”. That will get them interested. However, it has also been noted that many Digg community members react strongly to anti-Digg content, often burying it when it reaches the front page of the site, effectively censoring it. This has the unintended effect of making it seem like Digg the service is censoring all non-Digg content (which isn’t necessarily false, either).

Also, people use Digg in many different contexts. I’ve dugg stories myself that I just want to read later…stories that I don’t have time to read right now but that seem valuable to me and I want to be able to find them later. It could be that others do this activity as well, causing votes where none should happen. When you give people tools, they don’t always use them as prescribed.

The result: no independence in voting

The result of all these factors is that Digg breaks the cardinal rule of voting: independence. As outlined in James Surowiecki’s book The Wisdom of Crowds, independence arises when a person makes a decision (votes, diggs) without the direct influence of others, on their own, by making up their own mind. Of course, there will always be influences on that decision…what others have said, where their political party is leaning, their current situation, but in the end they need to have the privacy of their vote. On Digg, no votes are private, and when you make them you can’t help but notice the way others are voting.

If we compare this to how people vote in Presidential elections, we see how different it is. In those, we anonymously vote. The anonymity of the vote is key…once we start exposing who voted on what we’re gaming the system because we are inevitably influenced by others votes. And the ranking of voters just solidifies this…imagine if we could see how others voted over time in Presidential elections…

Digg vs. Del.icio.us

The voting on Digg is in contrast to a site like Del.icio.us, where voting (saving a bookmark) is done more independently, often without having any idea whether or not someone else even viewed it, let alone voted on it. Del.icio.us isn’t immune to gaming, however, as there is a popular list, and it’s very easy to simply copy those bookmarks into your own, driving up the numbers just like on Digg.

So far, however, Del.icio.us seems to be more immune to the outcries of gaming. This may result from a smaller user population, as it is nowhere near the size of Digg. But I think it has more to do with the nature of the tool. On Del.icio.us, the main value is personal, as people use it to store bookmarks that are valuable to them. On Digg, the bookmarking utility is secondary to the voting, in both the interface and the wording used on the site.

Digg’s response

Later this past week, Digg responded to the controversy by changing its algorithm:

“This algorithm update will look at the unique digging diversity of the individuals digging the story. Users that follow a gaming pattern will have less promotion weight. This doesn’t mean that the story won’t be promoted, it just means that a more diverse pool of individuals will be need to deem the story homepage-worthy.”

I think this is the wrong approach. By keeping the above features the same…Digg is asking for gaming. As gaming occurs, they’ll have to change their promotion algorithm. Then more gaming will occur to override the new algorithm, which they’ll then have to change. In any social system gaming will occur, but I think Digg’s problems are much more basic: their features simply beg to be gamed. Better to focus on the independence of voting, not on the algorithm. By making much of the ranking and voting hidden, the diversity of the site would increase. It’s exposing information that leads to sameness.

What would change mean?

Even if Digg were to change, however, to alter some of the features above to make voting more independent, we still couldn’t be sure that they would work. People test the boundaries of all social tools, finding every which way to bend them to do something useful. Sometimes it’s fine, sometimes it really does hurt the quality of the site.

Digg couldn’t just say “let’s move the digg voting widget somewhere else” and be done with it. That would introduce a new set of problems, based on the new context. However, they did add a new feature lately whereby the Digg widget shows up right on the posts themselves. That could potentially solve a lot of these problems, getting the voting mechanism much closer to the content people should be reading before voting on. Though it isn’t clear whether or not this is part of the solution, it seems like a step in the right direction. (I’m trying it out below – you may not see it if you’re in an aggregator that strips scripts)

Published: September 12th, 2006

Currently working on:

The What to Wear Daily Report: The most informative 30 seconds of your day. An email that delivers clothing recommendations and other helpful info based on the weather. Remarkably useful. It's free to sign up.