Some very practical and down to earth suggestions on how to construct Bayesian priors that make a difference.