/ Commentary

Social Media and the Wisdom of Crowds

James Surowiecki's The Wisdom of Crowds, published in 2004, has been influential in financial markets and corporate transformations. It also has plenty of ideas relevant to social networking's future.

The book's leading concept is that that collective intelligence -- aggregating diverse opinions -- can out-perform judgments made by alleged experts and executives. Estimates are better when they pool many invidividual investors' bets; decisions are better when produced by decentralized organizations than by top-down initiatives.

This work has plenty of cautionary tales, and the failures are what makes it worth reading. The author warns that the group isn't smarter when individuals start trusting public information more than their own private knowledge. He points to information "cascades" where people do something because everyone else is doing it. The clearest examples are the last two stock market bubbles, when investors put aside their doubts and started following the ticker tape. The product of this lemming-like behavior can be worse than decisions made by a central committee. I would add that the ability of the crowd to estimate is limited even more than Surowiecki acknowledges. Group estimates work much better for predictable bell-curve type distributions than for rare but catastrophic events like market crashes, as Nassim Nicholas Taleb demonstrates in The Black Swan.

At first glance, social networking would seem the perfect environment for crowd wisdom. Participants are often anonymous to one another. Contrary to organizations, the average social media participant has no paycheck, relationships or loyalties that could interfere with her honest assessment. Surowiecki's enthusiasm for applying stock market simulations to other forecasting like elections and foreign affairs would seem to work well on an internet site.

However, there are several barriers to making effective use of mass participation. The tendency of internet forums to attract people with anti-social behavior is a significant problem. While a troll would be quickly silenced in a face-to-face meeting (and unlikely to even be in the room), the internet poses no such barriers to entry. Financial motivation may be replaced with something equally bad: the need to aggrandize one's ego, and impatience and hostility towards other members who don't agree with one's reasoning or conclusions.

Also, internet forums have some of the same problems as business task forces and focus groups. They're rarely a random cross-section of people. Internet sites are often self-selected samples of people with keen interests in a topic and a willingness to write or vent about it. They're hardly disinterested observers. In addition, even when the posts are civil, the cyber-atmosphere isn't. People have the burden of introducing and proving themselves. Depending on one's personality and confidence in his writing, the internet may be either an incentive or detriment to venturing an opinion. My own belief is that forums are excellent for sharing and arguing about technical information, but less suited when it comes to fuzzier issues that require contributions from the quieter members of the group.

The Wisdom of Crowds distinguishes between problems involving collective agreement and ones requiring coordination or cooperation. But collective agreements are of two fundamentally different types. One type is a decision to act by picking among two or more choices (what should we do); the other is an estimation or forecast (what is happening, or what will happen). Social media in its current state is more suited to the latter. The limited ability to make decisions isn't only because the group is for discussion only -- it can advise but it lacks the authority to act on anyone's behalf. Good decisions often can't be made until you've gathered all the facts, and some of the best decisions are ones whose choices are framed by the discussion, not generated in advance. Therefore, to even advise on complex decisions, social networking needs tools where the options can evolve from the group debate. The poll or voting booth, with its canned multiple choices, is a poor substitute for this negotiation.

Nevertheless, I wouldn't bet against the ability of social media to acquire the necessary sophistication. The ubiquity of e-mail and instant messaging, plus the growth of mobile apps and virtual companies, suggests that decision-making is only going to get more distributed. The technology will become wiser, because it has to.

The Wisdom of Crowds, by James Surowiecki, Anchor, August 16, 2005, ISBN-10: 0385721706 ISBN-13: 978-0385721707

This post was originally published on my other blog, https://tech.surveypoint.com