Forecasting has proven to be a useful tool that can constitute a competitive edge in today’s competitive market. Indeed, forecasting deals with uncertainty by predicting the future either based on human judgement (qualitative forecasting) or based on historical data (quantitative forecasting). This paper explores and compares different quantitative forecasting methods for the sales of four sizes of a product belonging to a well-known company. These stock keeping units (SKUs) pertain to a carbonated beverage with high daily sales such that everyday hundreds of boxes are sold. To that end, this analysis focuses on one specific geographical area which is Fez, Morocco. Both the stock keeping units and the location were chosen with no prior motivation. In a first instance, we clean our raw data set. Then, we delve into exploring different classical and neural network forecasting methods, and comparing their accuracies to determine the most precise one, using the statistical programming language R. The results revealed that the Neural Network model outperforms the other methods.