#4a: Pre-change vs. post-change data.
Consider data available before a change is made (pre-change) and after (post-change).
Without knowledge of how a system came to be, pre-change data can provide only correlative connections. It will hint at the right decisions to make to change the system in the way you want. It can increase confidence that your decisions will be the right ones, but it won't reveal cause-and-effect. And it can't guarantee that decisions based on that data will have the intended impact.
Post-change data is different. This data lies in the future as a direct result of your changes. Once it becomes retrospective, this data can tell you the impact of your decisions in a way that pre-change data cannot. Measure it to learn what you do next — aim to maximise the amount and frequency of post-change data you have.