One of the most important and famous quotes from Peter Ferdinand Drucker, the founder of modern management and perhaps the world’s most famous management consultant of all time, is: "You can’t manage what you can’t measure."
Although I agree with the overall spirit of this, there is one aspect we must keep in mind. As we are dealing with people, measurements must be used with supreme care. If not, you are taking the risk that they will do more harm than good. Keep in mind that people who are measured will most likely do what they can to affect the measurement and its course (In other words, the actual measurement process will determine the rules of the game and its results).
Parkinson's law: “work expands so as to fill the time available for its completion” can be expanded in this case to “A person who is measured will do what it takes to meet the measurement goal.” Add to this the day-to-day pressure from management and customers and the anxiousness of the employee to show positive results.
What do you get? A frustrated employee who focuses on the measurement instead of doing his job. For example, if the employee is measured on the number of bugs opened per day, he will direct all his efforts to open as many bugs as possible. He will not take the time to investigate whether the bug is real or not.
Conversely, if the metric is related to the quality of the Bug, this can result in the same team member spending his time documenting every little detail. This may look good on paper, but it doesn’t add value to the development.
Often, organizations focus on the measurements themselves without considering how they may affect their employees. To succeed in creating metrics that will not fall into the trap of "Parkinson’s law for measurements,” use only metrics that make a difference.
The metric should serve a tangible goal – One of the most frustrating things is using metrics just because they look good to senior management. Before I approve a metric, I ensure it answers a specific question or serves a particular goal.
It should be easy to measure – If a metric is too hard to measure, not fully understood or overly complex, it shouldn't be used. Using it will probably fail to provide the expected benefit, even if the measurement goal offers value.
It should be one piece of the puzzle – Even the best metric, if used in isolation, will not fulfill its full potential. Only if it is used as part of a group of metrics can it provide real value and a balanced picture of the measured activity.
It should provide a return on investment (ROI) – Consider a metric’s forecast ROI before deciding whether to use it in your project.
It should display a trend – Do not use metrics to compare different teams; metrics should be used to demonstrate a trend of a single team through different sprints and projects.
The team should embrace it – Agile teams are already working under pressure, forcing metrics on them will not work. A metric is only applicable when the team understands its purpose and how it will help them improve.