Ask any PM about the parametric methods he is following for cost estimation and he will point to ‘COCOMO‘, ‘FP‘ or lately accepted ‘Feature Point Analysis‘. But next ask him what was the success rate, he won’t give you exact figure but sum up his answer by adding clarifications- None of the methods is exact so they do also use non-parametric methods too. So what is in parametric that most of PM cannot escape and still they don’t religiously follow that.
The exact answer is very fundamental. Parametric approach is very accurate, no doubt about it, however does our software development process is also accurate? Anybody who has spent at-least an year in the development can answer with big ‘No’. So anyway to go for approach which is far from being accurate, because of Murphy’s law (whatever the time you allocate for specific task, the task will always consume it). So what is the best way. Now to answer that we have to consider that whenever a process is not matured there will always be chances of betterment, henceforth always new approaches will keep on appearing and will try to minimize the risk associated. Now what is major flaw with parametric approach? Parametric approach fails to consider that a software development is more a art than a perfect science. The best ways are still to come or to bring the truth out there will never be accurate ways to conclude estimation and if Customer is too rigid Development team will always be at lose.
