Leave it to Google to bring a robust, problem-solving mindset to HR. In a recent NYTimes interview, Karen May, Google's VP for people development dismisses the reflexive approach to training that so many companies have:

Don’t use training to fix performance problems. If you’ve got a performance problem, there is a process to go through to figure out what’s causing it. Maybe the person doesn’t have the knowledge or skill or capability. Or is it motivation, or something about relationships within the work environment? Or lack of clarity about expectations? Training is the right solution only if the person doesn’t have the capability. But what I have seen in other places is sort of a knee-jerk reaction by managers to put someone in a training class if somebody isn’t performing well.

Having spent more than my fair share of time delivering training classes on time management, I can say with confidence that she's on the money. More often than I like to admit, my training classes were failures, if you measure success by sustained behavioral change. The failure wasn't due to quality of my teaching skills or the content. (At least, I don't think so!) Rather, it was due to root causes that were beyond my ability, or the ability of the participants, to fix.

As I've written before, time management problems are really just manifestations of dysfunction in one or more of the following areas: strategy; priorities; internal systems and processes; corporate cultural expectations; or individual skills. Training addresses the last area only -- but usually, the time management problem has its root cause in one of the other areas.

Remember: the performance problem you're seeing is more than likely just a symptom. And just as you'd look for root causes of defects in a manufacturing or administrative, process, you should look for root causes of "defects" in human performance. After all, your organization, and the people within it, are infinitely more complex than the products or services you provide. It only makes sense that any performance problems require at least a similarly probing analysis, rather than the simplistic fix of training.