Risks of sales forecasting based on historical data and their impact on calculated business value using the income capitalization approach
Keywords:sales forecast, error function, forecast error, business valuation, time series, regression analysis
AbstractPurpose of the article: This paper deals with statistical methods of sales forecasting and their impact on calculated business value using the income capitalization approach. Our aim was to present several statistical methods used in practice for sales forecasting and demonstrate the limitation of their use. Scientific aim: The main scientific aim of this paper is to answer following questions: Will the use of error function, presented in literature, lead to choosing the most accurate method for sales forecasting? Even under the situation of unstable development of the company’s environment. How accurate is possible to calculate the business value? Methodology/methods: For creating a sale prognosis we used basic statistical methods e.g. time series and regression analysis using one dimensional model. For forecast error evaluation we used following error function: mean error, mean square error, Spearman’s coefficient, index of determination, Thiel’s index and so on. For calculating business value we used the DCF entity and EVA/MVA entity model. Findings: The method selected on the basis of error functions, presented in literature, lead to largest forecast error among presented methods. Under the situation of sustainable development of environment, this error function has a limited used. Conclusions: We agree with Little, Damodaran, Makridakis, Taleb ho claim that the forecast of sales based on historical data could not lead to most accurate results. We suggest to combine statistical methods with judgmental forecasts as it is presented in work of O’Connor, Remus and Griggs.
ORIGINAL SCIENTIFIC ARTICLE