ABC7, the San Francisco affiliate of the ABC network, has implemented Inkling Markets on the station's Web site in the ABC7 Futures Market. ABC7 asks its viewers questions about local and international news items to generate both short- and long-range predictions. Some questions range from "Who will be Barbara Walters' most fascinating person of 2007?" to "Will the price of oil top US$100 a barrel in 2007?"
Ellen Conlan, vice president of station marketing and research at ABC7, says, "We refer to the prediction market results in our newscasts. It performs the function of a poll, but it gives us a better indication not of what the predictors want to happen -- as is the case with a poll -- but rather what they think is most likely to happen. So it tends to be very accurate."
Predictify offers the same type of service, except that, in addition to working with companies, it opens its prediction markets to the public. Anyone can pose a question to the market at any time.
"We also collect demographic information about those users, and we allow the question-askers to filter the data based on those demographic attributes. So you can understand not just what the crowd is saying but who they are, which allows you to see that women are more optimistic than men about a certain topic, and evaluate what that means for you," says Parker Barrile, co-founder and CEO of Predictify.
Aggregation and scarcity in predictive markets
Predictive markets are known to be extremely effective in forecasting political elections. They also help candidates see which issues they lead on, as well as the areas where their messaging needs work.
Smaller civic organizations can also benefit. The University of Iowa's on-campus club the University of Iowa Democrats -- the largest Democratic club in the state of Iowa -- uses Predictify's tool to pull greater insights out of political public opinion polls.
Atul Nakhasi, a junior at the University of Iowa and president of the University of Iowa Democrats, is amazed at the insights.
"We're going to have 120,000 caucus-goers in Iowa on the evening of Jan. 3, and if you just have two or three people making decisions, it's not going to be reflective of the whole group. But now, if we can expand this throughout our entire membership and organization, we're going to start reflecting the thoughts of the state of Iowa and coming up with a more accurate picture than what the polls are saying," Nakhasi says.
The key to success in using predictive markets is to create the right amount of scarcity. If the online "market" that a company creates has too much of what it values as currency, participants' "purchases" don't predict anything because everyone is buying everything simply because they have the means to do so.
For most b-to-c companies, it's much easier to predict outcomes by studying the average budget-conscious consumer than it is to study multibillionaires.
The cautious approach to collective wisdom
Several factors are contributing to the rise of crowdsourcing in today's competitive business landscape. Foremost is the fact that organizations can gain hard-to-obtain information without a lot of investment, since the economic barrier to entry for crowdsourcing initiatives is very low.
Also, there are fewer moving parts in crowdsourcing than in traditional external collaboration projects, such as focus groups and customer surveys. Once the project is launched, participation is not likely to require screening committees, nor active monitoring. The data-rich nature of the Web -- when coupled with a well-designed database and UI -- takes care of that automatically.