Over the last couple of years, interest in prediction markets as a forecasting method has continuously increased in the scientific world and in industry. Markets provide incentives for information revelation and can be used as a mechanism for aggregating information. So far, prediction markets have done well in every known comparison with other forecasting methods. Whereas information aggregation is only a byproduct of most traditional markets, prediction markets are set up with the explicit purpose of soliciting information. Engineered carefully, prediction markets can directly guide decision making. This paper describes the fundamentals of prediction markets as well as their key design elements. We thereby aim at giving insights into design decisions which have to be made by prediction market operators. Moreover, we contribute to the literature by giving an extensive overview on fields of application of prediction markets which have been discussed in academic literature.