As a consultant and instructor in the realm of information management best practices, I spend a lot of time talking about the need to develop performance metrics so that future improvements can be quantified. And this is exactly right, for it is hard to know what kind of progress you're making if you don't know where you started from.
The flip side of this has to do with the establishment of the performance benchmarks you wish to reach (and hopefully exceed) – otherwise, you will have no idea whether you have accomplished what you set out to accomplish. Sounds obvious, doesn't it? Well, it is. But what's not obvious is the need to think about those targets and be sure they are defined in the most pragmatic way possible.
Case in point: a credit card company seeking to address its growing instances of fraud is disappointed when the percentage doesn't fall to zero. However, thanks to new technology and tightened procedures, it did cut the rate by 15%. So that's good news, right? Sadly, it wasn't received as such because the internal expectation was that all fraud would be eradicated. Was this a practical benchmark? Certainly not, and the result was a disenchanted organization rather than one that should have been enthralled by the progress it made.
Ironically, if this company had set its sights on a 10% reduction, it would have been thrilled with the way things turned out. But because it had miscalibrated its benchmark from the start, the story had a much less happy ending, and future initiatives will be that much more difficult to pursue as a result.
And this is my point precisely: it is much worse to set impractical goals and miss them then it is to set no goals at all. Neither scenario is a good one, but setting no goals at least leaves the door open to learning from the experience in developing these metrics the next time around. Setting bad targets inevitably taints the entire exercise, and results in a door that is barely left ajar if it is open at all.