While this is an outstanding book, I highly recommend his first work, Getting Things Done. Since this doesn't have a consistent narrative but is instead broken up into numerous tiny essays, it will be harder to get the maximum benefit from his approach to personal productivity from this alone.
Readers who "got" Getting Things Done don't need my advice on this one...they've already bought it I'm sure.
David Allen is probably the smartest personal productivity coach in print. I would buy Getting Things Done for every employee in my organization, and I would have copies of this one lying around to remind people and elaborate on some of the finer points.
Oh and I would like to add one point. I believe there is one thing missing from Mr. Allen's algorithm. That is finishing. I think his plan is outstanding for getting unstuck: figure out the next action, and do it without hesitation. But I don't find any attention paid to how to decide how many actions are "enough" for a desired outcome of a project.
You can always find some next action, and founder in what software engineers like myself call "permanent beta" or "feature creep." Yet external constraints are best not relied on exclusively for these decisions. It's best to volunteer a ruthless focus on the essence of your project's deliverable, isn't it?
So I would like Mr. Allen to write his next book about finishing projects, if he is able to develop insights into that stage as strong as his insights into the process of the middle stages.