One additional thought: this is also why I find placement and proximity experiments only become interesting fairly late in the process. As long as other noise dominates, they tend to be inconclusive. Once things are quieter overall, even small positional changes can sometimes become more informative, or not, which is equally useful to know.