A post on PM-Fu asks "how much detail is enough?"
One answer is
What is the amount of time, money or technical performance you are willing to put "at risk"? In other words what is your level of risk tolerance for finding out your late, over budget or the thing you're building doesn't work?
The amount of detail in the plan should define the boundaries of cost, schedule and technical performance measurement that answer that question.
- If you have a $10,000 budget and you're willing to risk $2,000 of it before you find out how it's going the the budget review is at 20%
- If you 8 weeks (40 days) to develop a feature and you have no slack you're late before you start. So pretend you have 1 week (5 days spread across the project) slack on the 8 week schedule. Ask how long can you wait to find out the slack has been consumed? 10% to 12% is probably about right. So progress reviews need to be every 4 days, may be 3 in the beginning or during the high risk parts.
