The way I understand it, Davinci resolve has a native colour space thats generic. By inputting it in through the Fx colour space you are using DJIs (In this instance) colour setup. The output Fx will convert the colour grading etc that you’ve done into a film standard that is used by YouTube for example ie Rec709a.
You dont necessarily need to do it but you should get a more accurate rendering if you do. Also, when you get a bit more savvy, you may wish to use planar trackers, masking etc, using a node in between input and out output nodes by adding one in series (Alt+s) and using it for that affect only, in doing so it doesn’t bugger up all your other settings that you’ve made.