The Role of Predictive Markets in Storytelling

Posted on | July 2, 2009 | Comments

In 2004, I took a position at MIT’s Technology Review, the nation’s oldest technology and science magazine published out of one of our countries most venerable institutions. My job: oversee the redevelopment of the magazine’s website from a simple placeholder for some of its articles into a fully-functional, modern news operation.

The site was a wreck, for sure. The content had migrated through several databases, with each migration stripping the data of more and more information rendering much of it useless to us. There were very limited tools that allowed us to interact with the data. Much of what we had was corrupted or “parent-less”, meaning we had data but no way to identify what the data once did.

But the one thing we did have was a predictive market, a thriving community of users — small, but thriving — who signed up for our service and made “bets” against each other on questions that we posted. The predictive market, which was written about in James Surowiecki’s excellent book The Wisdom of Crowds: Why the Many Are Smarter Than the Few and How Collective Wisdom Shapes Business, Economies, Societies and Nations, as MIT used it was meant to be a market that was integrated into the editorial process, set alongside stories that journalists generated as a way to create a genuine flow of information between the readers/users and the journalists.

Editors would post questions (e.g. Apple will sell more than 1 million iPhones by the end of the year) based upon stories they were writing about the technology and business of Apple, then use tools to pull feeds and information from other sources into the question page, which people so inclined to bet on the prediction would visit to place bets and discuss the news, which would lead to more information being created around that idea, which would turn into a feedback loop for the journalists who could track the actual numbers alongside the community.

That was the plan anyway.

Despite the predictive market’s prominent play near the front of Surowiecki’s book, the editors at Technology Review magazine never bought into the idea of the predictive market despite my best efforts to convince them. Every metric measurable (other than user numbers) was off the chart. Time on site, return visits, communication, pages viewed were TK times higher than on any other part of our site. In many instances, the advertising spaces within that section allowed us to deliver the numbers we’d sold on the site.

In other words, that dedicated audience was exactly the type of person that you want to foster in a digital world. Which isn’t to say that those people were the same demographic as the other readers on the site. We don’t have figures on this, but it’s a fair assumption (I think!) to guess that the kinds of people who send hours betting on predictive markets are different than those who are interested in quantum computing articles. That’s the assumption, anyway, although I’m not sure that is a good enough reason to dump the market. Because there are other predictive markets that are quite…literate, in the Classics sense of the word.

There are, of course, a variety of ways to approach the predictive market.
Jason, my editor, wanted us to incorporate Steward Brand’s Long Bets into the Technology Review site, which I botched because we had neither money nor resources to make it work. Unlike predictive markets, Long Bets allow people to place bets against individuals to see who is correct on a particular topic or posit predictions (which people can challenge — and turn into a bet) on a specific topic.

For instance, Martin Tobias posted this prediction:

In 2010 more diesel passenger vehicles will be sold in the US than hybrids.

You can click on the link and see Tobias’ argument along with who has challenged him (nobody at this point); however, if you disagree, you can place a bet against Tobias. When the time frame elapses, one of the two of you will win the bet. This type of “vanity” predictive market, though, is more about enticing a conversation about a subject than aggregating the masses, which is what news organizations would want to do (although frankly I wouldn’t view this as an either/or; this is a both/and). Either way, reporters could cull stories from the citizen engagement that comes from such involvement.

The most famous of these markets, the Iowa Electronic Market, is run by the University of Iowa’s College of Business and has more accurately predicted the past several presidential elections than any other polling agency. The project is both a research and a teaching tool. But places like Predictify, which have raised $4.8 million in funding, allow a range of people — including advertisers — to ask questions that people can bet on.

***

But predictive markets don’t seem, on the surface, to be about storytelling. Gathering collective wisdom, yes. Storytelling, no.

The rest of this Chapter Section is at my Book Wiki. Click over, request access and join.

Comments

Where am I?