Do a week with potential confirmation bias, then a week random as you suggested and then a week with the mark covered so you don’t know which is which (but know which you used)… That should avoid confirmation bias, and let you know which you actually prefer