Designing a simple Bayesian model benefits from a design loop with three steps. (1) Data story: Motivate the model by narrating how the data might arise. (2) Update: Educate your model by feeding it the data. (3) Evaluate: All statistical models require supervision, leading possibly to model revision.
