A pre‐tender building cost estimate is an important piece of information when making decisions at the project planning and design stage. The important project characteristics influencing the accuracy of pre‐tender building cost estimates are examined and practical improvement for increasing the accuracy of estimates are considered. A quantitative approach is used to address the research problem. Analysis of data from 56 projects and from a postal questionnaire survey of 102 quantity surveying firms suggests that the accuracy of pre‐tender building cost estimates varies according to project size and principal structural material. When eight identified project characteristics are controlled in a multiple regression analysis, the accuracy of estimates is influenced by project size. The estimates of smaller projects are more biased than the estimates of larger projects. It was discovered that pre‐tender building costs are more often overestimated than are underestimated. Overestimated forecasts are incorrect by a larger amount than underestimated forecasts. Data analysis also revealed that the accuracy of pre‐tender building cost estimates has not improved over time. The majority of the respondents are somewhat dissatisfied with the accuracy of estimates in the industry. Probability estimation and simulation of past estimates, reducing quantity surveying and cost engineering skill turnover, incorporating market sentiments into estimates, early involvement of the quantity surveyor at the brief stage, and proper documentation of experience gained in the estimation of projects should help firms increase the accuracy of estimates for new projects.