Monday, May 14, 2007

The Wisdom of the Crowds

Are crowds really wise? That's what James Surowiecki says in the first book on my summer reading list. The subtitle of the book is "Why the many are smarter than the few and how collective wisdom shapes business, economics, societies and nations." I started reading this book last summer and had to put it down for classes in the fall/spring. It was one of the optional books for a finance class, I chose to read two of the other books (The Smartest Guys in the Room and Barbarians at the Gate).

When I picked the book back up, I only had a few chapters to read. Problem is, I've forgotten a lot of what was in it. So it goes back on my reading list, but down a few entries. The basic premise of the book is that if you take a group of people, any people, knowledgable or not, you will get a better decision out of them that you do the experts. Our government is one example (politics is discussed in the last chapter).

He cites other examples, one where a group of non-scientists were polled about a missing submarine. The people were given facts and asked to guess where the sub might be based on a map. The answers were all averaged together (as I recall, by lattitude/longitude) and the search started with the non-scientific answer. The sub was soon found very close to the speculated position.

The postulates in the book are significant for business that typically attempts to find the perfect expert to solve a problem (can you say "CEO pay"?). The conclusion is that crowds can do a better job than individuals.

There were some contradictory examples cited, the Columbia disaster is one I recall. These are examples of "group-think", where the group heads on direction and won't return to "normalcy". However, he cites these as examples where a few people override the group. Groups are good.

I'm not 100% sure I agree with the book. In one example where I tried a test he suggested, the crowd I chose came up with a vastly different result than he predicted. I had students count candy in a fish bowl and, using his averages, estimated the results. I had some extremely wild guesses that blew the average. If I ignored these wild guesses, the average was close, but that seems to violate the principal of the book. My experiment may have been flawed, I want to try again.

Overall, a decent read for busines. Next time I read it, I'll focus on some of the other aspects.

No comments: