Forecasting on Prices for Crops Prediction markets

Objectives:

1. Development of prediction model of the organization and improvement of quality and accuracy of forecasts.
2. Increase staff motivation and efficiency of internal communications.
3. Disclosure of competences and potential of employees, search of personnel reserve in the field of analytics and forecasting.
4. Development of a social model of business organization by connecting employees of all levels to operational and long-term corporate decisions.

Time frame:

October 2013 - February 2014

Description:

Agribusiness is traditionally considered one of the most conservative areas of the domestic economy. Nevertheless, Agroterra States Concern tends to refute this statement, using the most modern approaches and technologies in its activities. In the absence of a developed open crop market, agroindustrial companies rely on closed analytical systems and their own experience in forecasting the cost of different varieties of crops.

Given the extremely high volatility and uncertainty of the market, this approach does not always yield results, as statistical systems do not react quickly enough to changes, and experience is limited to a narrow circle of experts. In these circumstances, we created the project "Forecasts Exchange", which allows to carry out price forecasting, and attracted all employees of the Group to participate in it. It was based on the FuturUS platform, built on the technology of prediction markets and collective forecasting. All participants, regardless of their position in the company, had the opportunity to influence price forecasting in the short and medium term.

Results:

In total, more than 2,500 participants' forecasts were processed. The overall results exceeded the client's expectations. The result of the system work was that the aggregated forecast either indicated the actual price, or a narrow range was revealed, where the actual price was. There were two leading signals about a significant change in the market situation, which contradicted the forecasts of the expert community, but as a result were justified. Meanwhile, the FuturUS system was also used as a monitoring tool: due to the constant aggregation of information from all levels of the company, TOP management was able to monitor changes in the "mood" of participants regarding a particular forecast and make the necessary decisions in advance.

The project on forecasting prices of different urgency, in turn, allowed to compare the markets of different urgency and build a common trend in the development of pricing policy. FuturUS uses a game mechanism ("stock trading") that allows the system to self-learn: the most successful participants have an increasing impact on the forecast, while employees who have not shown themselves enough, lose their "weight" in the system. As a result, a number of primary and middle-level employees were identified who showed analytical skills and understanding of the situation "from below" for accurate and early price forecasting. Some of them have been promotef to more responsible and relevant positions.

In addition, the project aroused interest among all regional companies, which allowed the exchange of information among all participants of the Holding.