I asked a contact a Realization* to remind me about a conversation we had several years ago, where he suggested that we probably didn't need to do time tracking. I am convinced now, but I wasn't back then. And now that I'm convinced, I didn't particularly recall the reasoning as to why it is not helpful.
Here are the three general reasons that he described and some thoughts about the value or damage they bring about.
Ensure people are busy. If you can "see" idle time, you can eliminate it. The one thing we know about projects is that they never go off as planned. If everyone is busy all the time, how will you be able to react and keep the project moving if critical people are being kept busy on other activities. I'll leave it to you to think about the busy-ness of non-critical people.- Make sure people meet their estimates. If each task finishes on time, then the project will finish on time. This is false. Activities naturally take shorter or longer than thought at the planning stage. Keeping people to their estimates guarantees they will never finish early - and frequently finish late. And this guarantees late projects.
Find out where the time is being spent on projects in order to guide improvement efforts. Assuming you aren't participating in the behaviors above, this could work. However, note that success in multi-project environments is measured by shorter cycle times, not high individual utilization. Improvement efforts need to be guided by reducing cycle time, not by increasing utilization.
Why else might people want to track time in a project environment? Yes, I am aware that you need to charge your customers and track time against internal project cost accounts. But I suspect much of that could be done with standard algorithms, rather than forcing your people to punch the clock every week.
* FYI: Realization sells Concerto, software that helps execute Critical Chain Project Management. So the perspective here is on project management environments.