Because studies have to be funded, it is important that the funds are used efficiently. Researchers learning from their work realize that having a perfect research design does not guarantee that all will go well, only that it can be replicated by someone else.

Just like architects can design aesthetically pleasing structures, you still need engineers to test them for feasibility. So must researchers test their research design to see if it will work before actually conducting the study. For example, if they are planning a survey, they may have a small group that have similar characteristics as the study population take the survey first and provide feedback on survey design, etc. Modifications can then be made to the survey before it is administered to the actual study population.