Smarter Than You

You say you can't trust anyone to help you run the company? Wrong. Trust everyone.

Last march a guy I know, named Jim, decided just like an estimated 28 million other Americans to enter his local office pool picking winners in the annual NCAA men's basketball tournament, known to anyone with a television as March Madness. Entertainment purposes only, of course. At stake was just pride. In Jim's case, 25 portions of it (his entry ante). Not to mention the four figures' worth of pride going to the pool's winner, and more for the runners-up. "A lot of pride," Jim notes. "Possibly very entertaining."

To compete, Jim would have to fill out the single-elimination brackets that would whittle 65 teams down to one champion. For every correct pick Jim would get points--one for each first-round game winner, two for second-round winners, and so on until the correctly named winner of the final game would yield 32 points. Total the most points in your pool and, well, you get to be the proudest.

Jim knew he faced long odds. For one thing, given his place of employ at a highfalutin consulting firm, his particular pool would be contested among people who are unusually competitive and hard-working and better educated than is strictly necessary. For another, he didn't really know anything about college basketball.

But he had a solution. Inspired by having recently read James Surowiecki's The Wisdom of Crowds, he decided to make no attempt at forecasting whether 10th-seeded St. Mary's of California could upset seventh-seeded Southern Illinois--or at trying to pick the outcome of any other game, for that matter. Instead he sought the most aggregated public bracket he could find (he used's cumulative bracket, which resulted from the tallied picks of more than a million average joe forecasters who sent in their individual guesses) and followed it slavishly in filling out a bracket of his own. Would his crowd-built bracket outperform those of his more knowledgeable pool competitors? Jim would see. At least, he would see right up until the final game, which is the only one for which he personally picked against the crowd because he just "knew instantly in my gut" that the crowd's pick was wrong. Which means that he banked everything on the wisdom of the crowd until the last pick, when, you could say, he blinked.

But that's another book--Blink by Malcolm Gladwell of Tipping Point fame--for another time. Jim's experiment (we'll see how it fared in a minute) provides the perfect setup for calling attention to a book vastly less celebrated than Gladwell's (which are not undeserving) and vastly more instructive for entrepreneurs. None of the better-read business books of the past year has nearly as much to say to company builders as Crowds does, and now that it's just been released in paperback, it's time to make sure it gets read.

Surowiecki's insight is this: The collective wisdom of a crowd (even a crowd of average, inexpert laypeople) will under the right conditions be smarter and more accurate than the individual decision of even the most exalted expert. Collective decision-making trumps individual expertise, giving the lie to what Surowiecki calls "our excessive faith in the single individual decision maker." The deduction of markets trumps the deduction of experts.

Surowiecki supports his claim with frequently astonishing (and hugely persuasive) evidence offered by dozens of experiments, as well as examples of collective wisdom in action. He shows collective thinking at work on problems both simple and complex. Simple: At an English country fair in 1906, almost 800 people guessed the weight ("slaughtered and dressed") of a 1,198-pound ox. The average of the crowd's guesses was off by only a pound. Complex: After a U.S. Navy sub disappeared in the North Atlantic in 1968, leaving only the merest scraps of information that might help locate it, a number of people with varied backgrounds (mathematicians, submarine specialists, salvage men) were asked to independently posit what had happened to the sub and where. None of the individual guesses accurately located the sunken sub. But when the individual guesses were combined, making a kind of collective guess, that location was a spot just 220 yards from the sub, enabling the Navy to find it.

Elsewhere in Crowds, Surowiecki shows how collective wisdom underpins the effectiveness of Google, and how, on the TV game show Who Wants to Be a Millionaire?, "asking the audience" always yielded better advice than "phoning a friend" (even though the friend was chosen explicitly for being smart and the audience was just a bunch of folk who stumbled into a television studio that afternoon). He also describes many examples of businesses using internal "decision markets"--a stock-market-like process for capturing the wisdom of crowds by enabling a collection of people to, essentially, bet on the answer to a question. (In a business, the question might be, Which potential new product would sell the most? Or, Which of our competitors will pose the biggest threat a year from now?) Crowds includes the story of Hewlett-Packard setting up an internal decision market to forecast printer sales. Twenty or 30 employees, drawn from different parts of the company to ensure diversity of knowledge and opinion, bought and sold shares according to what they thought future sales would be. The market for predicting each period was open only a week, and employees participated in their spare time--and yet the market's forecasts were more accurate than those of the company's expert forecasters 75% of the time.

The mechanics of why all this works are complicated, but what's essential to take away is that crowds possess almost unfathomable stores of inchoate information--they know more than they think they do. And experts, even the genius ones, know less than they think they do. What's "mystifying," as Surowiecki says, is how little interest businesses have shown in decision markets that tap the collective wisdom of their employees. "Strategy is all about collecting information from many different sources, evaluating the probabilities of potential outcomes, and making decisions in the face of an uncertain future. These are tasks for which decision markets are tailor-made."

Crowds lays siege to the cult of the expert. Also the cult of the CEO, which may or may not be the same thing.

If Surowiecki's right, entrepreneurs especially should care. Crowds lays siege to the cult of the expert (also the cult of the CEO, which may or may not be the same thing); it points out the shortcoming of any team or organization that overrelies on hierarchical decision-making. It also underscores the peculiar weaknesses of very small groups: Small groups are typically too insular to contain broad information, are overinfluenced by strong individuals, and are underexposed to diverse outside points of view. The upshot is that small companies are handicapped by collecting and considering too little raw, surprising data, as well as by naturally operating in founder-as-hero, individual-decision-maker mode. Surowiecki quotes the author of a study of corporate failure who identified a recurring theme: "The remarkable tendency for CEOs and executives of new ventures to believe they are absolutely right."

Crowds may do a better job than any other book I've read at helping explain why so many entrepreneurial companies become stagnant (or worse) after capitalizing on an initial business idea--why they have such trouble with second products and getting to that proverbial next level. More exciting, though, Crowds also offers solutions that smart CEOs can apply to that challenge.

Which are?

  • Recognize what makes a crowd insightful and try to create as many of those conditions as you can among your information and advice sources. For a group to be smart it must be diverse (in terms of knowledge and point of view, not necessarily sociology), its members must offer views independently (uninfluenced by other group members), and it must be decentralized (its members draw on unique local experience and information). That means that if you run a 12-person company, be sure to cultivate as many unusual outside advisers and raw data providers as possible.
  • Invent internal decision markets, even if crudely informal. It's especially useful to invent processes that provide anonymity--since honest information flow inside any business is quickly and inevitably obstructed by politics, sycophancy, and confusion of status with knowledge.
  • Respect your own limitations as an individual decision maker, even if you're an expert when it comes to the decisions at hand.

All of that, in essence, is what our NCAA tourney forecaster Jim attempted to do. He found a diverse, independent, and decentralized crowd to address his problem; he seized a decision market (the aggregated bracket) that would reduce the crowd's incalculable private information to a set of concrete choices; and he acknowledged his own limitations by almost entirely ceding his decisions to the group. (Until he correctly picked North Carolina over Illinois, the crowd's choice, in the final game.)

How'd Jim do in his pool? Second place, out of a couple hundred entrants. Highly entertaining, he reports. Just how entertaining, he forgets.

And, he says, he will never again make a strategic decision without input from some kind of crowd.

Michael S. Hopkins, an Inc. editor-at-large, has written for the magazine since 1987.

Last updated: Sep 1, 2005

Register on today to get full access to:
All articles  |  Magazine archives | Livestream events | Comments

Or sign up using: