Project cost control: a new method to plan and control costs in large projects

摘要

Purpose - The purpose of this paper is to reengineer the process of cost management in large projects. Design/methodology/approach - Considering the fact most large projects overrun their budgets because of the long time period needed to completion, it was decided to reengineer the cost management. Accordingly costs allocated to packages were reviewed and changed on a dynamic basis. Different types of contingencies were provided. Concepts of package contingency and project contingency were introduced. These were based on the project buffer and feeding buffer concepts popularised by Eliyahu Goldratt. Findings - The re-engineered method of cost control worked well and yielded better than expected results, leading to the setting up of a new world record in the completion time for setting up a million tonnes per annum continuous, tandem cold rolling mill to roll steel sheets. Research limitations/implications - In view of the total success of the reengineered approach, which was tested out on a large project over three years, it is felt that other projects could also try out this technique, especially since it is along the lines proposed by Eliyahu Goldratt who is an authority on project management. However the success can be better understood if the results of the testing become available. To that extent the contents of the present paper have limitations. Practical implications - Large projects can deploy the methodology and complete their projects on time and under budget. Social implications - The reengineering of the cost management was done primarily with a view to complete projects under budget. Since many governments spend many billions of dollars on publicly funded projects for the welfare of citizens, the use of this technique could have a salutary effect on the cost. Originality/value - The method was innovated in the company by the author's team and deployed in a live project over four years for the first time to achieve world-class results.