As I understand it, Occam's Razor stipulates that the potential solution with the least variables and assumptions is the preferable first mode of investigation (not necessarily the best), as this makes it easier to determine the effect these variables have on the end result and makes (scientific) repeatability easier, i.e. going with the most easily testable method/solution first.