It'll cost you, what, $5k to try it once? You have your other variables tracked well enough that you should be able to see the difference pretty well and see i

It'll cost you, what, $5k to try it once? You have your other variables tracked well enough that you should be able to see the difference pretty well and see if it's worthwhile. If you go for the low hanging fruit first (a holiday) where the effect should be biggest you might get a nice surprise - or possibly not, but at least you'll know for sure.