Metrics are defined processes for measuring things. They are not amounts. When creating a metric, be aware that people tend to modify their behavior so that measurements improve, even when there is no direct reward. (See BehavioralEffectsOfMetrics.) This can result in counterproductive behavior. For example, the KLOC metric (which measures lines of code produced) can result in verbose code with minimal code reuse. Coming up with a good metric is hard. Here's a few things to consider when creating a metric: 1. What problem am I trying to solve? How will this metric help? 1. What impact will this metric have on people's behavior? How will they subvert the metric to make themselves look good? See also PerformanceIndicators