I think the consequences of risking the project with poor experimental data are much greater than any time delays. If poor experimentation fails to uncover a particular design issue or if it is not accepted by appropriate regulatory agencies, then the delay to the project would be much greater than the alternative.