Google Blogoscoped

Friday, September 8, 2006

Digg vs Groupthink

Digg introduced some measurements to lower the importance of votes by closely connected groups of friends. I think the basic problem they’re facing is “groupthink” (as described by James Surowiecki and others): a group’s ability to measure something objectively sharply drops if the group is too like-minded, acting too much like a herd.

For example, if a group’s leader is so strong that he’s successful in suppressing alternative opinions, the group will only reinforce his measurements. This can lead to bad decisions which can lead to failure – like a dotcom bubble, an airplane crash, a space-shuttle crash, or the meltdown of a nuclear reactor. Whenever a group isn’t very diverse in opinion but puts strong trust in the opinion of others – trust without verification – things can go awry, as negative feedback is ignored (like criticism of a particular seemingly “proven” procedure; “we always did it like this and most people like it, so your criticism isn’t valuable”).

If you read up on what happened at the nuclear plant Chernobyl in 1986 (see this interesting book), the unfolding of events is simply astounding. There was a group of workers who worked with each other for a long time, so they trusted each other. They also started to trust the system of the nuclear reactor because it always worked. What they distrusted after a while were safety measurements; several got ignored that night before catastrophy. They continued with a dangerous safety test even after warning signs should’ve convinced them to stop. But they suffered from groupthink. They were a great team... maybe too great.

In Digg, groupthink leads to unimportant stories hitting the frontpage, a failure for the system. The Digg system of measuring a story’s importance is based on a semi-random set of people voting for or against a semi-random set of news stories. As soon as the site has one person favoring a submission based on what their friends favored, it’s deprived of the “wisdom of crowds” effect.

Digg’s recent measurements, while not known in detail, probably also lower the usefulness of “digg this” buttons to bloggers. If you have a more or less loyal readership, most of those clicking the “digg this” button will now be potentially identified as group of friends.

Google does some of the same counter-measurement, I think. They look at link networks – in Google links are votes, in Digg diggs are votes – and analyze how closely interconnected the network is. If it’s a very diverse network of seemingly unrelated websites linking to a particular site, then that’s good – the wisdom of crowds is at work. If however it’s a tight network it might well be a spam farm doing artificial linking (or perhaps a group of friends doing link exchanges).

If you think your blog, organization, project and what not suffers from groupthink, here’s some points you may want to ask yourselves:


Blog  |  Forum     more >> Archive | Feed | Google's blogs | About


This site unofficially covers Google™ and more with some rights reserved. Join our forum!