Predicting the demand for promotional or newly launched items is an ongoing challenge for consumer manufacturers. While the variability in demand for these events is tremendous, promotions constitute a large percentage of total revenue. Although many statistical methods have been used, General Mills wanted to know if other methods could provide better forecasting results.

How did CTL address the problem?

Two masters' students explored how predictive markets could be used to forecast the demand for promotions. Essentially, predictive markets are a formal mechanism for collecting and condensing disparate information from a wide variety of sources. Each participant in the market provides the best estimate of the demand based on his or her own particular and specific set of inside information. The market “price” for each potential demand projection changes based on the predicted prices. Similar to the financial markets, these predictive markets arrive at a suggested value or, as in this case, an estimated demand.

What were the results and impact on the company?

General Mills conducted a number of markets on various items, including a wide number of participants from across the firm. They found that the predicted results were in line with the estimates generated from the complex statistical methods. Additionally, they found that including a broad range of participants increases employees’ sense of ownership.