Forecast accuracy is a major challenge for manufacturing organizations. Forecast error can be a direct cause of stockouts, inventory growth and/or costly changes in the mast schedule. Despite the considerable time and effort expended on forecasting, forecasts are still very inaccurate. In summation, forecast inaccuracy is expensive and the confusion and frustration resulting from it is extensive. This confusion has led some advisers to recommend the elimination of forecasting altogether (Goddard 1989). Others, however, emphasize the need for forecast accuracy in order to permit an appropriate and effective response by the manufacturing organization. The importance of a relatively accurate forecast to manufacturing performance calls for understanding of the factors which affect forecast accuracy. Many researchers have analysed how the use of various techniques affects forecast accuracy. These studies focus on understanding of the technical accuracy of various methods: they do not suggest how manufacturers might work to improve forecast accuracy. Most of this research as focused on the forecast developers rather than the users of the forecasts. This study examines the effects of institutional factors on forecast accuracy from the perspective of the manufacturing executive. This study investigates the effects of institutional factors on forecast error. The questions it addressed include: Who should be involved in forecast preparation? How important is forecast error measurement for future accuracy? What should the primary purpose of the forecast be? The data was obtained from a sample of manufacturing professionals in the UK, reporting information about practice within their manufacturing organizations. Statistical results from this study suggest that firms with a culture supportive of new technology have lower forecast errors. Multiple functional involvement in forecast development does not improve forecast accuracy. Top management (president/managing director) involvement in forecast development reduces forecast accuracy. Forecasts are best developed for sales planning and then distributed to other functions for their use. The statistical results suggest quantitative techniques have no effect on forecast accuracy. Single values of the forecast and forecast error measurements lead to improved forecasts. And finally, organizations respond to forecast inaccuracy by frequent modification.