I have a slightly different approach.
Firstly there are scheduling errors. No schedule prediction is accurate so
you need some way of improving accuracy from first guesses!
Secondly there is contingency required from Risk Analysis.
To handle schedule inaccuracies I find that I can most accurately predict
how many hours it will take if everything goes well. To get a figure that
more accurately predicts reality (after the tenth interruption and lost
productivity because of a shortage of good coffee!) I add the % chance of
the work taking twice as long. For example if I predict a task will take 10
hours if everything goes well but I feel there is a 20% chance of the task
taking 20h, I will add 20% of 10 which gives me a prediction of 12h. How do
you know if your schedule prediction is accurate? When actual hours worked
bracket your predictions. If your hours are consistently under or over
estimating then you need to adjust your prediction method until you
consistently under and over estimate tasks equally.
With Risk management you need to predict the hours of extra work that will
be incurred if a Risk does happen. Multiply that by the % chance of it
happening and you get the hours you need to add to your contingency buffer.
Add the contingency hours for all risks and you have a total block of work
you need to schedule to cover all risks. Do this by adding one or more extra
tasks. If this amount is too high, then you need to mitigate or reduce your
risks in some way. If that isn't acceptable either, then your schedule and
risk predictions are saying there is a very high chance this project is not
going to finish on time. If you still need to go ahead then you need the
Google methodology for project management:....... click the "I'm feeling
lucky" button!!
--
Rod Gill
Project MVP
NEW!! Project VBA Book, for details visit:
www.projectvbabook.com