Vinsamlegast notið þetta auðkenni þegar þið vitnið til verksins eða tengið í það: http://hdl.handle.net/1946/22335
Accurate sale forecasts can help companies improve the operation of their entire supply chain. Although it is possible to create good forecasts with statistical forecasting methods, a judgmental forecasting is often needed. Forecaster has to have his/her own estimation of events that cannot be included in the statistical model. One of the things that need special attention is sales promotions. To develop software that can help the user define future promotions and make a forecast for the promotion effective is a significant step towards creating a more accurate forecast. To estimate future promotions, it is good to have the option of viewing and valuing past promotions. This thesis offers a method to classify promotions and a model to calculate the value of past promotions. It can be used for future promotions to estimate their value and their effect on the sale of the affected products. The model for calculating the value of promotion was tested on real data. Three promotions were analysed, statistical forecast was calculated to create a baseline sale, and an additional sale was then used to calculate the value of the promotions. This model can then be employed in the analysing tool AGR event analyser. This model will be easy to implement and will use easily accessible data. It is believed that it would improve overall forecasting, by viewing past promotions and use them to forecast for future promotions.