Usually, doing prior predictive checks is really useful in these cases, to see what your priors mean on the outcome space. You can see this notebook for an example, and McElreath’s Statistical Rethinking for more 
1 Like