
Notable quotes from the opening keynote by James Surowiecki, author of the "Wisdom of Crowds".
(This post is not intended to be very structured - it is cleaned up notes taken while listening.)
Started by giving examples of how wise the crowd can be...
- Google’s page rank algorithm based on their realisation of the "remarkable hidden order and intelligence under the surface of the apparent chaos of the web".
- Uses example of a crowd of bettors at a race track and shows that the odds are usually right.
- Refers to another example of prediction markets - the Iowa voting one for presidential elections since 1988, more accurate than the Gallup polls 75 % of the time.
- Companies that use markets internally: HP, Google, Siemens, (see the Wikipedia article on the subject for other names of corporations)
Prediction markets get around the problems created by hierarchies and bureaucracies: hard to move information, incentives to hoard information, tell bosses what they want to hear.
The only premium placed by the wisdom of crowds is for being right, no matter who you are, where you are in the organisation.
What's the catch? Only works under certain circumstances:
How do we make the crowd smarter?
1. Aggregation: A way of aggregating a lot of individual judgements to produce a collective judgement (this is not a suggestion box where ideas go in and are then filtered out)
2. Diversity: real case for diversity is that it improves the bottom line and decision making. Referring to cognitive diversity - different ways of thinking about and framing problems.
Diversity expands the range of information available to you
Example: Has a group of engineers solve a problem. Then has them solve one together. One group is the best of the best, the other a mixture. The B of th b does well, but is almost always out-performed by the other random group.
The ones at the lower end of the spectrum, less successful than the B of th B, - is bad for them as individuals, but put ito a group it makes the group smaller
Less likely to get hung up for a wrong answer
Diversity of perspective: geographical, training, age,
People make mistakes, so want to ensure they aren't all in the same direction. Errors cancel each other out.
In a homogeneous group, the more people talk to each other, the dumber they become. They become more convinced they are of the rightness of their positions.
You need to find ways to break that apart - disrupt that.
You need a devil's advocate. Must not be the same person everytime. That person serves an automatic role of being the devil's advocate. The others listen to him out of routine. We listen, then do what we want to do anyway.
3. Independence: have people relying as much as possible on their own knowledge and opinions. Give their thoughts, not others.
People want to appear reasonable in the eyes of others. Therefore, tend to do what others do. If you're wrong, you're still close to the others. If you go out on a limb, have a different judgement, more risky. Those people tend to get punished in some way in most organisations.
Humans herd together in conditions of uncertainty.
Paradox: groups are smartest when the people in them are acting as much as individuals as possible.
Make people free to offer genuine opinions. Must be able to accept disagreement.
Consensus: Finding consensus too quicklty can often be a mistake. People must be able to disagree strongly before making a decision. Imitation makes sense for individuals, but not for the group. Leaders have to learn to step back.
Recent Comments