University of Twente Student Theses


Forecasting promotional demand in the FMCG industry

Weustink, T. (2020) Forecasting promotional demand in the FMCG industry.

[img] PDF
Abstract:Promotions have become more and more important for the Grolsch during the last years. In total, 70-80 percent of the total sales volume is sold during a promotion. In the current situation, they forecast the promotional volume based on historical sales, experience, and human knowledge. Besides, it is known which factors influence the volume but they do not know the impact of each variable. The demand planning department believes that the current forecasting process could be more efficient, more effective and that they can achieve a higher forecast accuracy by standardizing the process. Therefore, we develop in this research a prediction model for forecasting promotional demand. First, we determine a method for cleaning the historical sales to obtain the sales during a regular, non-promotional period (baseline sales). Based on these baseline sales, we calculate the promotional volume, which is the difference between the total sales and baseline sales. We apply Random Forest to predict the promotional volume of a promotion based on multiple input variables. The model is able to achieve a higher forecast accuracy than the current forecasting method.
Item Type:Essay (Master)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:31 mathematics, 50 technical science in general, 54 computer science
Programme:Industrial Engineering and Management MSc (60029)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page