Deriving Project Estimates

Normally estimates are derived with a 50- 50 confidence.  But  is a fifty-fifty case okay while setting  your goals/targets/specification limits ? People will prefer to have a better confidence while setting their goals. So what should be done.? Some improvements needs to be brought in to the current system . Those improvements needs to be chosen in such a way that it will have a positive impact on the goals.

Now, consider a project scenario. Project A sets their effort estimates as 100 Person Days (PD).  This needs to be submitted to customer. In order to deduce this they took the below steps Step 1: Baseline the organizational productivity (LOC/PD), found 10 LOC/PD as the central value Step 2: Estimate the size of the project, found as 1000 LOC Step3: Convert the size to effort using productivity baseline, found to be 100 PD. And this is submitted to customer. This project team is aiming 20 % an improvement in productivity , i.e. 12 LOC/PD. From this statement, infect it is not clear whether they are aiming to achieve a productivity greater than 12 LOC/PD  with a 50% probability of success or with a greater probability. Now let them redefine the goal statement as ‘greater than 12 LOC/PD  with at least 80% probability of success’.  So this implies that the project team needs to have  improvement initiatives to make the increase in productivity with required probability of success . ……

With this, it is clear that the possible central value of the redefined specification limits must be  above to 12 LOC/PD. Using the normal distribution function, it  is seen that if central value of redefined goal is near to 13  ( assume standard deviation of 1), the probability of success would be greater than 80 %. In that scenario project team needs to deduce the internal effort estimates with current mean, i.e. 13 LOC/PD. Then the effort estimates becomes 77 days approximately. So this internal estimate shall be used for further resource planning and assignments within the project team.


The author is a Quality Assurance professional by experience. Part Quantitative data analyst, part consultant for quality and information security practices, part software tester, she is a writer by passion and blogs at and

You may also like...

Leave a Reply