Article ID: | iaor20032408 |
Country: | Portugal |
Volume: | 22 |
Issue: | 2 |
Start Page Number: | 147 |
End Page Number: | 166 |
Publication Date: | Dec 2002 |
Journal: | Investigao Operacional |
Authors: | Themido Isabel Hall, Grilo Lus Miguel |
Keywords: | statistics: regression, retailing |
This paper presents the development of regression models to explain the sales of two consumer groups of products that present remarkable variability and seasonality (personal hygiene paper pulp based products). Models for the market share of two groups of products and for the sales of the total market were also developed in order to evaluate alternative approaches to sales forecasting. Using data from Renova, the Portuguese leader in this market, several linear models were developed. In these models price is an important explanatory variable. Another variable that is also significant in some models is the market share of other brands, where own brands have the major contribution. For the sales models, the lagged sales variable has a significant explanatory power, which reflects the importance of the brand under analysis and consumer habits.