In early 2000, the newspapers and TV in India reported with jubilation the successful test flight of a newly built light combat aircraft. But the detailed reports also indicated a somber side: The prototype delivery was more than five years late, and the project had cost more than 10 times its initial estimated cost. Many projects across the world suffer a similar fate. It seems as if the estimated cost and time are never enough to execute a project. Improper estimation is the bane of project management in many engineering disciplines, and software engineering is no different, as the poor record of software projects unambiguously illustrates.
In a services business, improper estimation hurts even more. Asked to identify the one thing he wants from projects, Nandan Nilekeni, the managing director of Infosys, answers, "No surprises." No surprises on the customer satisfaction front, and no surprises on the revenue and profit front. More often than not, the cause of surprises that come late in a project on these two fronts is improper estimation of effort or schedule.
There are no quick and ready-made solutions for the estimation problem. Project managers, however, can improve their estimation by using tested guidelines that are based on past experience and data. This chapter discusses the approaches used by Infosys project managers for estimating effort and schedule. It includes examples from real projects and the estimates for the ACIC case study.