Can prediction markets make forecasting easier?



Pharmers Market, an online crowd-sourcing tool, will help determine whether prediction markets are useful for drug companies.


Its the pharma industrys biggest challenge. While a new drug that makes it to market can generate astronomical profits, a new drug that languishes in research and developmentto one day be tossed out as a bad ideacosts staggering sums of money, an average of $1.2 billion. To compound the problem, competing pharma firms rarely share information, and thus often end up spending $1.2 billion on the same bad ideas. Sharing information without jeopardizing proprietary ideas would save the industry billions of dollars a year.


 


Enter the Pharmers Market, a virtual market that uses collective intelligence to increase the flow of information among pharmas and forecast which drugs will succeed. Heres how the market works. A wide array of individuals, from researchers to clinicians to drug industry experts, bet virtual money on whether specific drugs in clinical trial will succeed or fail. The bets are anonymous, and thus any one participants identity is protected. However, taken together, these anonymous bets can indicate which drugs are likely to failbefore billions have been invested in them.


 


In an ideal scenario, the market can predict failures long before the event, says Ragu Bharadwaj, who worked with a group of students, professors, and doctors at the Sloan School of Management at M.I.T. and the Harvard Business School to devise the online market. Once the value is established, a pharma CEO could say, OK, Ive got to select ten molecules to go forward with from this set of fifty, and I wont get objective information from my own people, so this is another mechanism for me to find out which of these ten I should select.


 


Bharadwaj handed the Pharmers Market in as his thesis in February of this year. Now, its up and running with its first round of drugs for betting: a host of breast cancer compounds currently in clinical trial. Once the virtual bets are placed, and the reasoning behind those bets is aggregated, researchers can assess if the market is, in reality, a useful predictive tool for drug companies. If so, more drugs will be introduced and the market will become a virtual crystal ball for drugs in clinical trial, says Mat Fogarty, founder and CEO of Crowdcast, the prediction market company that powers Pharmers Market.


 


Fogarty founded Crowdcast in 2007 because, in his professional experience, there was no clear way of generating clear, unbiased forecasts within corporations, regardless of the industry. Basic questionslike, When will a product really ship? How much of it will sell? What will market share really be? And how much will it really cost?are often grossly miscalculated. If you take normal methods for gathering information, says Fogarty, its very much a one-way type of affair. You have a survey of how people are feeling, and what they think about this or that, and employees in different regions enter their forecasts into a spreadsheet.


 


The key difference with prediction markets, he says, is the feedback mechanism. Every employee in a company is encouraged to draw upon their personal expertise and place virtual, anonymous bets on a range of questions. If they bet correctly, theyre rewarded with real cash incentives, which further encourages them to bet in real-time, thereby expediting the transfer of information.


 


Crowdcast currently works with a range of clients, from Fortune 100 electronics companies to Fortune 100 pharma companies. The firms success speaks to the proven power of prediction markets within a corporation. Whats unknown is whether the Pharmers Market will succeed. Think of Pharmers Market like Wikipedia: difficult to monetize but incredibly valuable to the world, says Fogarty. If you could get more accurate information sooner about which drugs would pass trials, and the efficacy of different drugs, you could potentially save millions of dollars out of these drug evaluation costs.


Since you're here...
... and value our content, you should sign-up to our newsletter. Sign up here