Sales surprise, the ratio of actual sales to forecasted sales for a year, is a key determinant of firms’ inventory turnover. While the previous literature on empirical inventory management frequently uses Holt's double exponential smoothing method for obtaining future sales forecasts, this method suffers from measurement error. In contrast, we directly use the data on management forecasts publicised by listed companies in manufacturing and retail industries for the fiscal years 1997 through 2014, which are available in Japan, and evaluate the adequacy of the exponential smoothing sales forecasts. Our analysis reveals that sales surprise measured via both means positively relates to inventory turnover in the majority of industries, consistent with previous literature; however, the two variables imply discrepancies, particularly when economic conditions change.