Johanna Rothman is looking at the Costs of Multitasking and asks for suggestions both on other aspects of multitasking and on how to evaluate their impact (cost). I've trimmed out the details - be sure to check those, and the comments on this post have been particularly good.
There are three parts to multitasking:
- Stopping the work you're doing.
- Swapping out what you're working on.
- Swapping in the new task.
These all feel very centered on the person doing the multitasking. The Getting Things Done mode of thinking is very much centered on eliminating the results of this kind of multitasking. How many times do you look at a piece of paper or email laying in the inbox before actually doing something about it? How many times do you stop what you're doing to respond to email or the phone, only to realize at the end of the day that you dropped a dozen balls this way?
However, in my comment, I suggested that the costs associated with multitasking don't start to show up in measurable ways for the business until you begin to integrate multiple people across multiple projects. It is then that the delays and trouble with multitasking really start causing trouble for the organization. Frank Patrick's comment is similar to mine:
It's more than the additional cost of individudals' time. It's also the cost of late(r) delivery of the projects and their benefits.
One of the commenters pointed to a heavily-linked piece by Dmitri Zimine on How Two Hours Can Waste Two Weeks, which is a story about interruptions and how something that should have taken "two hours" ends up consuming much, much more time than necessary. It's related to multitasking in the sense that the story illustrates how important it is to understand priority of the work one is being asked to do. If the priority is not understood, it's easy to flip between tasks, or put off one task because the "swapping in" activity takes too long (and the other task might get priority again).