Things like targets or published measurements that have no particular purpose in your process improvement cycle are risky because they will change behaviour in undefined ways. You should always avoid them if you can.
Measurements should be taken in order to answer a question. You should avoid taking measurements if you have no questions. Similarly, you should try to measure the minimum that you can that will answer your question. This will minimise the change in behaviour.
Before you can have metrics you need a defined process (because by definition a metric is a measurement that is used to change your defined process). Avoid taking metrics when you are unsure of your existing process. When you change your process as a result of a metric, use measurements to answer the question, "Did we affect things we didn't expect to affect". If the answer is "Yes", then take more measurements to answer the question, "Is this change something we need to do something about".
In this way you will be able to control your process. One last thing. Do not attempt to protect against gaming other than to not measure things. People naturally optimise and will do it whether they are aware of it or not. Simply observe the effect of your measurements and decide if they are what you want. If not, then remove the measurement.