MODELOS DE SÉRIES TEMPORAIS: A ANÁLISE DA ACURÁCIA DAS PREVISÕES DA DEMANDA DE UMA LINHA DE PRODUTOS EM EMPRESA DO SETOR DO VESTUÁRIO

Purpose – The purpose of this paper is to compare the performance of time series forecasting methods for a product line in a clothing company by analyzing the accuracy of demand forecasts Design/methodology/approach – This paper presents a case study in a large clothing company. Several methods were used to obtain both quantitative and qualitative data. Qualitative data were mainly used to describe the demand forecasting process and the quantitative data to make forecasts. Three time series models were applied to make forecasts and an accuracy analysis was done using different error measures. Findings – Regarding the three time series models applied in this case study, the static one is suitable for the product line considered, especially taking into account the impact of forecasting errors for carrying inventory and stockouts. We also identified advantages of quantitative methods and highlighted the importance of the forecast’s accuracy evaluation to choose an adequate model. Originality/value – There are few studies describing in detail the use of quantitative forecasting methods, specially addressing the forecasting process and error accuracy evaluation. This paper describes the use of three different time-series models to forecast the demand of the main product line in a large Brazilian clothing company. Furthermore, it suggests how to analyze the impact of forecasting errors on level inventory decisions and emphasizes forecast’s accuracy importance to support management decisions, a topic rarely addressed in the literature. Keywords - Demand forecasting. Time series forecast. Static methods. Winter Model.