Planning for Scarce Resources

One of the concerns that comes frequently from clients is that their people work on lots of different kinds of stuff so planning their time for projects is difficult. And it's true. A common circumstance in IT shops, in R&D labs and in many other situations, is for someone who developed something originally to get pulled into responding to questions and supporting it for eternity. This is the support dilemma. It worries project resource managers because the work load for the support people can be unpredictable. The support call cannot be planned or scheduled.
So the temptation is to throw up one's hands and say because this group of activities cannot be planned, it's pointless trying to plan anything for that resource. This extreme is unnecessary. You will have to live with a bit of uncertainty but there are good statistical approaches to dealing with uncertainty.
The first, essential step is just to list the different kinds of stuff that a resource might work on. Then estimate the rough order of magnitude of effort for each kind. Support tasks will be an average over time - not necessarily constant (more urgent support calls during year-end, product launch, etc.). That will give a base of committed time for that resource - leaving a known amount for other (project) work. Support work is typically treated as high priority.
Historical data may give some clue about the level of variability in the level of support work - an average of 2 hours per day may in fact be fairly predictably 1.5-2.5 hours every day - or 1 day in 4 completely taken up with support. This pattern is important because it affects the confidence in the ability of the resource of being able to work to a project timetable.
One simple approach is to schedule only a part of a scarce resource's time for high priority projects. The balance, the float, is provided by assignments to low priority projects. If it's important to have someone dedicated to a single high priority (please, not to 5 high priority projects!), then you cannot afford to have them dragged off to provide support.
This paper includes some interesting references to resource planning and a description of the rolling wave approach - which has been mentioned in other posts on PMThink!
Managing Project Performance: A Proposed Model (Part 1 of 3)