How can companies harness hidden knowledge located throughout the enterprise? Supporters of prediction markets claim they offer a way. Prediction markets resemble financial trading sites, but instead of buying and selling stocks, traders buy and sell predictions. A company that wants to operate a prediction market can provide their employees virtual cash to trade and give those who do well over time small prizes.
So how do prediction markets work? Let’s say a shoe company with an active prediction market is rolling out a new style of boots. Employees at the shoe company could buy a prediction that the boots will hit stores on time. If the project succeeds and meets its schedule, the prediction pays $1. If there are delays, the prediction will pay nothing. Conversely, skeptical employees could bet against the boots arriving on time. They would receive $1 only if the boots fail to arrive on time. The price of the prediction signals how likely the entire market thinks the event will happen. If everyone believes the boots are on schedule, the price of predicting the project will be on time might be high — 80 cents or so. However, if fears grow that the boots will be late — perhaps a key supplier is having production problems — the price of the on-time prediction will start falling. The falling price is a sign that all is not well on the project.
A prominent example of a political prediction market is the Iowa Electronic Market run by the University of Iowa, which allows participant to predict the outcomes of various elections. You can check it out at www.biz.uiowa.edu/iem/.
The advantage of prediction markets is that they add up the opinions and estimates of a wide range of employees. Thus, they are more likely to pick up on potential problems that management may not be aware of. Furthermore, the betting component forces people to “put their (virtual) money where their mouth is.”
If this sounds farfetched, it is not. Corporations like GE, Best Buy, and Hewlett-Packard are already using prediction markets to increase innovation and improve forecasting and decision making. Many more are testing the feasibility of this idea in pilot projects.
The Postal Service has a large, widely dispersed workforce that knows a lot about how things are going on the ground. A prediction market might help uncover this hidden knowledge. For example, if the Postal Service wanted to find out more about whether bulk mail volume would pick up, it could run a prediction market for BMEU employees who might know about customers’ plans. Or if the Postal Service were rolling out a new piece of equipment, it could run a prediction market on how successful the equipment would be. There is one thing, however, that could limit the use of prediction markets at the Postal Service: many employees do not have access to the Internet at work.
What do you think? Could prediction markets be a potentially useful tool for the Postal Service? Would lack of Internet access limit its use? If the Postal Service did implement prediction markets, what sorts of questions should it ask?
This topic is hosted by the OIG’s Risk Analysis Research Center (RARC).