I am fine with this and think it is a good idea in principle. Execution of the proposel is the hard part. I would say it should be done with a time limit that allows enough time to gather sufficient data to see if it is working. Once the trial period has run for 50% of the set time the data can be analyzed to see if it is going in the right direction. If not then it can be reversed at the end of the trial period or adjusted. whichever coarse of action seems best based on the data.