Starting with the question of what we want to learn from the experiment, we can define how we will observe and measure it, and finally create the cheapest, quickest, and simplest MVP to test our assumptions, measure the effect, and use that learning to formulate next steps.
