Trade your insight

Situation: In the last years, there has been a lot of buzz around Prediction Markets. These virtual markets work exactly like stock markets, only that instead of real stocks, the probability of future events is traded. Prediction Markets are already a very common tool to forecast the outcome of elections (here is one set up to predict the outcome of Switzerlands parliamentary elections held in October). You can buy and sell stock from every party participating in the elections. The final price of a stock will be equivalent to the percentage of the vote that the party wins. There are a lot of reasons why prediction markets usually work better than other forecasting tools, for example polls. First, because it is a market where you can win or lose, people have an incentive to really give the answer they think is right. Second, whereas in polls every opinion is equal, in prediction markets, people with more insight and knowledge will trade more (and more risky). This means that their opinion carries more weight in the forecast.
Prediction markets are also increasingly being used by businesses. Many companies have internal prediction markets where employees try to forecast, for example, sales. There are also quite a few companies who have specialized in setting up prediction markets.

Idea: To my knowledge, there neither is nor has ever been a prediction market for trend forecasting or consumer research. Firms still rely on marketing experts or focus groups to predict upcoming trend or evaluate new products. But I think that prediction markets would be a wonderful tool for market research, especially for predicting trends. The interesting part about a trend prediction market would be that traders would not trade with an existing and predefined set of stocks (as in the election market where the set of stocks is defined by the parties participating), but would be able to create their own stocks and sell them in the market. This guarantees that the market will not only tell the company which trends are likely to become big issues in the future, but also tells them about trends they maybe haven't even thought about before (an outcome that is, of course, also possible with a focus group, but much less likely because of distorting factors like group think).
The big challenge in creating trend markets or markets evaluating new products will be to find the right subset of traders. The set has to be focused and suited to the task at hand (there's no point in letting rural people from China trade if you want to predict trends for urban westerners), but it also must have a certain degree of diversity (if it doesn't, chances are that the group of traders suffers from a systematic bias and does not produce the right outcome).

Business model: Ideally, you already have an existing network of people (hint, hint) which could be used to trade in a trend market. I don't see trend markets as a stand-alone business, but rather as one of many different tools that can be used in trend research. However, a company or organization with access to a high-profile network of people (more hint) could sell access to this network and its trend markets to marketing and advertising companies (super-hint).

Problems: The idea seems somewhat obvious so I fear it has already been tried somewhere. And there is one problem within a trend market: Whereas in election markets the value of a stock can be defined very precisely (it's just the percentage of the vote the party got), it's more difficult with trends. Trend stocks would have do be defined very clearly. For example, it would be difficult to trade a stock stating that "social networks will be huge in the future", because it's not clear what "huge" means and when the "future" is. Because the final value of the stock after trading ended depends on whether the event it predicted really became true, stocks would have to be defined in a way that can be measured. To stay with the example, the stock would then have to state that "one year from now, 30 percent of the population in developed countries will be active users on a social networking platform".