It is often assumed that the accuracy of formal statistical methods of forecasting is naturally superior to methods which are not based on statistical principles. Nevertheless, non statistical methods of forecasting are very widely employed in practice, especially within small businesses, because they are generally less sophisticated and easier to understand. This paper evaluates the accuracy of a simple rule of thumb that was employed by a British company in its forecasting procedure. The results are compared with the performance of statistical methods under different conditions in order to ascertain whether any improvements in forecast accuracy could be gained.