South by Southwest (SxSW), the annual music, film, and interactive conferences festivals in Austin, has been underway since Friday. Recently it’s attracted many big figures of the Web 2.0 “scene” to give talks and lead discussions at panels including (and definitely not limited to) Matt Mullenweg, the creator of Wordpress, Jimmy Wales, the founder and president of the Wikipedia Foundation, and Joshua Schachter, the creator of del.icio.us. And even if you’re not there, the blogosphere is publishing extensive coverage (as you can search on technorati), and the SxSW website is publishing “podcasts,” which are surprisingly really cool. The other day I downloaded and listened to a presentation on the wisdom of crowds by James Surowiecki, a New York Times columnist and writer of a book actually called The Wisdom of Crowds.

His argument is convincing.

The “true collective judgment of a group as a whole” is often actually better than the “smartest” person in the group. Surowiecki cites the popular game show “Who Wants to be a Millionaire”; polling an audience yields a correct answer ninety-one percent of the time, but calling an expert friend yields a correct answer only sixty percent of the time. The common jellybean contest yields an average number that is actually only within three percent off the perfect count. Google’s PageRank taps into the Web’s “collective intelligence” by giving the highest priority in search results to websites that are the most linked to. Google became so famous because often the results returned are just what the person searching was looking for. Wikipedia is also great even though anyone can edit any page… but I think everyone knows that already. There are even more examples, but in any of these cases there is a diverse group of “average” people with an “average” intelligence as a whole, although together, the group becomes intelligent.

To most, this is completely unintuitive, so what’s the reasoning behind this possibly being true? Collective intelligence does require conditions, but if these conditions are met, there is really some ample logic for the group as a whole to be smarter than any one. Each person contributes their own pieces of knowledge and thought, that, by themselves, may be rather useless, but together combine well, and casting off their individual erroneous assumptions. The more various pieces of knowledge a group has for making a decision the more likely a group is too make a good decision.

In contrast, the problem with (most) experts is that although they do know a lot, they’re not able to notice their own blind spots or biases. Surowiecki says, “Almost across the board, expert judgments are very poorly calibrated. They’re not actually worse calibrated than amateurs, but they’re no better than amateurs. So, again, it’s not that experts don’t know more; it’s just that they don’t know what they know and what they don’t, so to speak. Now there are two exceptions to this rule– that is, bridge players and weathermen…” One must “cast his net more widely” in order for the blind spots to fade away. This is one of the essential conditions of good crowd wisdom– diversity.

Now here are all of the conditions together, as Luke Wroblewski discusses on his SxSW coverage of the panel at his blog:

• Bottom Up: Everyone must have an equal say. The group’s decision must represent the true collective judgment of the group.

• Diversity: Not from a human factors perspective but cognitive diversity (different approach tools, perspectives, etc.). Diversity expands the range of possible solutions and helps a group think more clearly.

• Independence: Individuals within a group need to make judgments based on their own knowledge not on the ideas of others. Most group decision-making fails because there is too much of a focus on consensus. Instead of tapping into the intelligence of each individual, you are trying to find the lowest common denominator that everyone agrees to.

Expanding on independence, there are two possible problems with achieving the correct circumstance– imitation and peer pressure. One can easily get by many things by simply learning through imitation. However, if there are many imitators in a group, there won’t be a diverse pool of knowledge. Similarly, people care about their reputations and will often jump on others’ “cool” thoughts so that they’ll be in better social standings. In a decision-making group, it is important that people do not simply jump on others’ ideas but offer their own input, either by agreeing or disagreeing with good reason.

Surowiecki offers an interesting perspective on the wisdom of crowds. Crowds may be “dumb” and go wrong, but they can also be collectively very intelligent. I wholeheartedly recommend listening to the downloadable mp3 of the panel. It’s an hour long, but really, his lecture is only for the first three quarters of an hour; the rest is Q&A. I look forward to spending some more of my spring break time listening to more SxSW panels… and possibly writing about them.