Predicting new product success with prediction markets in online communities

Authors


Abstract

The prediction of new product success is still a challenging task. Traditional market research tools are expensive, time consuming, and error prone. Prediction markets have been introduced as a viable alternative. Utilizing inputs from various participants in game-like environments, they have been shown to produce accurate results by combining dispersed knowledge via market-based aggregation mechanisms. While most previous studies use employees or experts as a sample, we test whether online consumer communities can be used to predict the sale of new skis via prediction markets. Sixty-two users took part in the study. The prediction market was open for 12 days before the main skiing season 2010/2011 began. The outcomes of the prediction markets were compared with the actual sales numbers provided by the ski producers. The mean average errors were between 2.74% and 9.09% in the four markets. Overall, it can be concluded that the prediction markets based on consumer communities produce accurate results.

Ancillary