Metodología Basada en Modelos Econométicos para Predicción de Demanda en una Industria Cosmética

Rodriguez-Calderon, Carlos-Eduardo and Un Jan Liau Hing, Emilio Alberto (2012) Metodología Basada en Modelos Econométicos para Predicción de Demanda en una Industria Cosmética. PhD thesis, Universidad Nacional de Ingeniería (UNI).

Text (2012RodriguezCalderonMSc)
2012RodriguezCalderonMSc.pdf - Published Version
Available under License Creative Commons Attribution-NonCommercial-NoDerivs.
Download (3MB)

Abstract

Demand forecasting has been always a very difficult issue in every company. As a famous Intel’s mathematician says: Forecasting is “the hardest problem in math,” Any company, large or not, that manufactures and/or sell products, requires an adequate demand plan to compete in market. Accuracy errors in demand forecasting can cause problems both for inventory surplus or unsatisfied demand. Then it’s an important problem that should be treated with adequate scientific tests. Another important detail to be considered when forecasting (if it is determined by some factors) and to apply it with confidence, will be the availability of this data in future. Otherwise, these factors will not be able to be used for prediction. In this study, focused on a particular scenario, as complex as it would be a direct selling cosmetic company with its main showroom called catalogue book; after a primary data analysis and the correlation between variables, offers a possibility for demand forecasting with an appropriate accuracy level to guarantee a correct supply. This will generate in consequence a good customer satisfaction without oversupply costs. The proposed methodology lays the foundation that every company should set before applying a statistical technique: the basic principle that every model must be a faithful representation of reality and afterwards and with confidence we can choose the technique to be used. In our case, after certain adaptations, the chosen one was: econometrics. This work also considers other aspects beyond statistics such as elasticity of demand, also important for pricing.