A Bayesian methodology is described for designing experiments or surveys that are improved by utilizing available prior information to guide the design toward maximally reducing posterior uncertainties in the interpretation of the future experiment. Synthetic geophysical tomography examples are used to illustrate benefits of this approach.