Suppose you have a task with an estimated duration of 10 hours using one resource. If we add an extra resource we tend to split the time in two and reschedule the task for 5 hours. We know that’s not always correct to do so because we have to consider the interaction time. So, what would it be the rescheduled time for the task? Are there theoretical/practical calculations on this issue? Saving Changes...
Sort By:
Mark Price PerryBusiness Driven PMO Evangelist| BOT InternationalOrlando, Fl, United States
Didier, for the reasons you mentioned for many projects the duration is not always cut in half just by doubling resources. Some folks like to add an additional amount of time to the task estimate, 10% to 25%, depending upon the complexity of the task as a practical measure. For example, a 40 hour task would have a duration of 40 hours for one resouce and perhaps a 22 to 25 hour duration for two resources. Good post, I hope to hear from others as well..! -- Mark Perry, VP of Customer Care, BOT International Saving Changes...
Thanks Mark. That’s a good technique. My dilemma is how to obtain accurate estimations. By my experience I believe that is deeply difficult, specially if we consider all the variables taking part in the estimation and execution of a task. That dilemma took me to the issue of allocating resources to tasks. Now, with this technique of adding a percentage of the estimated duration (E) of the task for interaction between the assigned resources, there is a moment in which the allocation of additional resources is counter-productive. Example: a task of 40 hours with one resource, an additional percentage of 5% by each added resource: with 2 resources (r) it would take 22 hours (h), with 4r - 16h, with 5r - back to 16h, 10r - 22h and with 20r it would return to 40h. But in real life it isn't that accurate, adding that we would have the difficulty to assign an interaction percentage for each task, depending on its complexity. Now, I say that "in real life it isn't that accurate", because by each added resource we would have to consider interaction time between that new resource and each one of the already existing ones. Then I wonder: Which is the ideal interaction percentage to be assigned to every task (I know it would depend on each task nature)? Which is the formula we have to use to calculate the overhead by interaction of and with each added resource? To answer the second question, we could use the following formula: (E/A)+((E*I)*(“The Adding” from 1 to A of A-1 {0 for 1 resource (A=1), 1 for 2, 3 for 3, 6 for 4, 10 for 5})), with E: Estimated duration with one resource, A: Assigned resources and I: Interaction percentage. On the previous example: with 1 resource (r) it would last 40 hours (h), with 2r – 22h, 3r – 19.3h (apparently optimum), 4r – back to 22h, 5r – 28h, 6r – 37.7h and 7r – 47.7h. Solved the previous issue, we still have the dilemma of estimating the optimal duration (E) for each task, for which, I believe, it is required experience, accurate and consistent statistics of past experiences and common sense; ah, and a magic wand would be also useful. What you think? Is there something written on these subjects? -- Didier Venegas, Project Manager, DC International, [email protected]. Saving Changes...
Mark Price PerryBusiness Driven PMO Evangelist| BOT InternationalOrlando, Fl, United States
Dear Didier, I think your a spot on target! As you mention, the hardest part of the equation is estimating the optimal duration for each task. Experience helps greatly, as does maintaining and updating the historical estimating database or log. Many IT departments and project organizations don't do this or have the time and/or need for this, but for those large project organizations and project-oriented delivery firms, this can be of use. Hope to hear from others. Regards. -- Mark Perry, VP of Customer Care, BOT International Saving Changes...
