eThesis - School of Business electronic theses

Sales forecasting is considered to key management activity in a company. It is crucial to determine an optimal forecasting model for sales, and strong determination of the model affects to entire business process of the company. Forecasting by quantitative approaches applies historical data to predict, so that the company can use forecasts for the future to decide business strategies and achieve goals. Previous research mainly considers the sales forecasting of manufacturing industry rather than optical service company, meanwhile, there exist plenty of studies which handle medical or natural environment field of Botswana. However, the empirical evidence of sales forecasting about the optical company in Botswana is lacking. The focus of an empirical research in this thesis is to bridge such research gap, and provide evidence how to determine an optimal forecasting model by quantitative approaches for the optical service company in Botswana. Before to start the empirical research, literature of forecasting methods was reviewed in respect to qualitative and quantitative approaches both. Forecasting methods by quantitative approach went into details with examples because those methods were only used for the empirical research. According to research question types, case study method was conducted for the empirical research, and data of the case company was collected from interview of general manager by phone and e-mail. Background and original sales data of the case company were presented in the beginning of the empirical research, and the data was analyzed with chart and graph by using Microsoft Excel 2013. Outputs of forecasting model application were illustrated with actual sales data in graph. Furthermore, forecasting accuracy measurements were described in column chart. Results of the analysis revealed that a forecasting model by simple regression analysis is the optimal one for the case company. It shows the highest accuracy in all forecasting measurements, and the differences to other models are significant.