The valuable chance to try new things

I value trying to improve the experience constantly, but I don't want to inconvenience my players, but I need their input. So what I need to do is set expectations, have a cut off time for the experiment (Ideally in 20minutes to try it out, if more time is needed 30 minutes abort Unix alt f4 and deal with it out of game).
What I have to gain out of all this is really happier players, more consistent quality delivery (by capturing some of the variables), and possibly more efficient adventure prep.
I don't mind a hard abort or failure because in science and failure I learn and have solid proof what does not work (and video proof via onair). Since the scientific process is mostly about finding what does not work to find out what works, learning from failure is a step in the right directions.
If I don't experiment or try the idea is is just a waste of space and hot air. Its better to fail and move on and learn.