Thursday, September 14, 2006

The difference between wise crowds and wise guys

My friend Kevin Maney - USA Today tech journalist extraordinare - wrote a excelent article this week on the wisdom of crowds or what I call "hive intelligence."


Internet types love sweeping ideas that come in the form of book titles that morph into catchphrases that can be dropped like little bombs into conversations to show how smart they are.

"Well, you know, his company was vulnerable to The Innovator's Dilemma, especially because he didn't have a World Is Flat strategy, although he could've seen it coming if he'd tapped into a Wisdom of Crowds approach."

This is at its most grating when said with a Thurston Howell III accent.

Anyway, the concept of "the wisdom of crowds" — the title of a popular 2004 book by James Surowiecki — is particularly hot among tech companies right now. Yet in the past week, the concept has run into a couple of potholes. One is a rhubarb at Internet darling; the other, the U.K. government's case of the hijacked wiki.

Both have people wondering if crowds are really all that wise, and what conditions have to exist to make them wise. Why is it that some crowds seem smart, while others turn into ugly mobs?

At its most basic level, the wisdom of crowds — let's call it WOC, so I don't have to type as much — means that the aggregated thought and knowledge of thousands or millions of people can be smarter than trained individual experts.

The WOC has been around forever — it's what democratic elections try to tap into. But the Net takes it to a whole new level. "The Internet provides a mechanism to get lots of diverse opinions and aggregate it in a quick and cost-effective way," Surowiecki tells me.

So if a company can use the Net to tap the collected intelligence of its employees, the employees will make better decisions than the CEO. IBM, Google and others have tried this. Wikipedia, written and edited by tens of thousands of unpaid contributors, should be better than an encyclopedia written and edited by specialists. News sites such as Digg, which lets users vote stories to the front page, should surface the best stuff more effectively than professional editors.

Except it doesn't always work that way. Pointing specifically at Wikipedia, Lauren Weinstein of the People for Internet Responsibility says that the Net has propagated a "basic fallacy that a wisdom-of-crowds approach could ever work, even theoretically."

Which might be extreme. But clearly some of the WOC mechanisms in use today have shortcomings.

Take Digg. Founder Kevin Rose set it up with a simple premise: If you like a news item or blog post, you can "digg" it by voting for it. Items with the most votes move to the front of the home page. But cabals of users started working together to boost certain items. Digg became less about the WOC and more about the ambitions of 1,000 or so savvy members.

So last week, Rose changed the formula. Digg will use computer programs to try to devalue bloc voting and give more value to what appears to be independent thinking. Digg's regulars revolted, causing a big Web stink. But Surowiecki thinks the site will now be better.

"The thing that makes the wisdom of crowds work is lots of diverse opinions and independent judgments," Surowiecki says. "Digg acknowledged it wanted more diversity of input."

So WOC needs the right structure to work, something the U.K.'s Department of Environment Food and Rural Affairs (DEFRA) found out the hard way.

DEFRA decided to let the public help it write environmental contracts, so it posted one on the Net as a wiki. The idea of a wiki — such as Wikipedia — is that anyone can log on and add or change something. In the case of DEFRA's wiki, the public mutilated it. In one of the more printable alterations, someone changed "Who are the parties to the environmental contract" to "Where is the party for the environmental contract? Can I come? Will there be cake? Hooray!"

"These sorts of issues have permeated the Net," Weinstein says. "Part of the problem is ... it doesn't take a lot of people to cause such disruptions."

Wikis, as it turns out, are not so much tapping the WOC as they are works in constant progress. Depending on who last wrote something, at any one time the wiki might be good, or bad.

Despite hiccups, though, there's still a lot of enthusiasm for WOC.

Google, in its own way, is a WOC device. It analyzes links that millions of people put on websites to come up with the most relevant search results. The reputation system on eBay uses WOC to self-police its users.

One of the more intriguing WOC sites is Hollywood Stock Exchange, aka It has about 700,000 members who use pretend money to buy and sell the "stocks" of movies and stars. (Most widely held star stock: Johnny Depp.) These are basically bets on how well a movie will do or how a star's career will go. HSX now sells its results to movie studios because its forecasts are more accurate than the studios' internal forecasts.

There are other prediction market sites aimed at politics, sports and other sectors. And in 2003, if you remember, the Pentagon experimented with a prediction market for terrorist attacks. Actually, it was a market for world events, including where and how future terrorism would take place. If hundreds of thousands of people were in the market, betting on the likelihood of such events, the WOC would probably wind up making pretty accurate predictions.

But when word got out, people just heard "betting on terrorism" and had an emotional reaction. Senators denounced it. "This is just wrong!" barked then-senator Tom Daschle, a South Dakota Democrat. And the program was canned.

The thing is, it was a sound, smart WOC idea. "It went down in flames for reasons I don't think were good," Surowiecki says.

In other words, it ironically was shot down by a crowd acting unwisely. In this WOC arena, there still seems to be much to learn.


Post a Comment

<< Home