James Surowiecki is a staff writer at the New Yorker (the financial page) . His pieces have also been published in New York Times, Washington Post.

This is his book about how the many are smarter than the few. 

It all begins with an early twentieth century experiment by Sir Francis Galton, a noted British polymath. In this expirement, Sir Galton set out to prove that the average person did not know much & that knowledge was concentrated in a few. ( Some of us hold this view, whether knowingly or unknowingly :  especially when we criticize democracy )

 The experiment itself was conducted at a fair where people bought tickets to guess the weight of an ox. For six pence, one got a stamped ticket entered into the drawing, the best guess was to get a prize. The crowd that took a gander at this was diverse - there were farmers, butchers, blacksmiths and lay men who would not have a reason to know much about estimating the weight of an ox. However, much to his surprise, Sir Galton found that the average of the guesses was pretty accurate ( 1 lb off in approx 1068 points not bad  < 0.001 % margin) and more so that this was closer than individual guesses by cattle experts.

There in begins our book, the author postulates that collections of  independently thinking individuals ( not crowd moving blindly) are able to predict certain types of decisions and predictions better than individual experts.

The author defines three types of problems and then goes on to define the characteristics of a "wise" crowd.

 Coordination, Cognition and Cooperation are the types of problems.

Coordination : think people walking on a side walk, without bumping into each other.  Certain common understanding & culture has helped individuals predict the behaviour of the other person and optimize their own.

Cognition : Judgement problems - stock prices for instance. The markets have adjusted ( discounted) information quickly and better than experts. One possible reason is the difficulty of political interfernce for the market as whole.

Cooperation : where individuals form Networks that they can trust and things get done without a central authority dictating so. Think consultative decision making by individuals - considering the opinions of those you value before making a decision.

Not all crowds are wise ( ofcourse otherwise bubbles wont happen ;-)) 

A wise crowd is Diverse with each individual having sources of their own.

A wise crowd is Independent : Individuals within the crowd are unable to influence the opinions of others ie the individuals will resist such influence and express / act independently of each other.

It is decentralized - there is a local information element in all that their decision

Again - aggregation is the key - in the ox example as well - individual guesses may have been way off BUT the average of the individual guesses was accurate.

Ofourse we are all too familiar with failures of collective wisdom & this is where the author walks a stick wicket. For a book that is so well written, I think it suffers to an extent from confirmation bias - the author's knowledge of the subject and his conclusion influence his approach and examples.  The author says crowds fail when : 

 1) they are too homogenous - they have access to the same information, their thought process is essentially the same, and their approach is similar.  ( Diversity is key).

2) they are too centralized - the central authority governs - we need not look further for such crowds - the classic top down approach ( and I am sure quite a few us experience it) is a case in point ( and please let us not get started about the quality of decisions made by this crowd )

3)  they think with their emotions and lose objectiveness.

4) they are too divided ie there is no aggregation of their input. ( Remember the average guess was accurate) 

5) they are too imitative - and have lost their independence.

So what is the take away for people like us - to be valid input providers : 

Stay independent, expose yourself to diverse viewpoints, collect information from more and diverse sources with local elements.

For corporations and the society - the author laments the fact that policy analysis markets did not take off but he hopes someday they will. Existing markets for wisdom of the crowds are Delphi decision making ( and wideband delphi), Prediction markets ( v/s a think tank) , and the notorious opinion polls. ( It can be argued that the book, and its permise apply to value judgements as well as factual issues making the case for pluralism, democracy, diversity, and a free and independent media)

Overall, its a well written book and its definitely worth a read. For a book so heavy on experimental economics, it is surprisingly easy read.

Interested folks can follow this up at :

http://www.randomhouse.com/features/wisdomofcrowds/