Right, so you don't design it to sound 'right', you design it to work right and then fiddle. As you don't know what, in a technical sense, makes it sound right, that fiddling is just random. If that fiddling isn't just random, then you do know what makes it right, in a technical sense, and thus ought to be able to measure that. What I'm chipping away at here is the claim to 'designing' the sound. If it is design, then it is quantifiable and, by corollary, if you can't quantify it then it isn't design it's something else which I don't have a name for.
What differentiates this something else from audiophoolery? I'm not trying to be insulting by asking that, I'm inviting you to offer a reasoned defence of your methodology that differentiates it from all the other unquantifiable arsing around much beloved of the audiophool community.