That's an invitation to chatter for days & I doubt if I know any more about this than you do. When you start to include all of the perturbations possible from three-axis flight control inputs, plus thrust line issues & physical structural variations, it can be quite daunting to sort out.

With problem airplanes, I always start with a physical structure check-up. Following a good eyeball check for alignment, plus a measurement of same, I twist & flex wings, tails & control surfaces to ensure that they aren't distorting excessively in flight -- & fix any suspected weak areas. I check for warps, slack in the controls, even throws, differential, loose skin surfaces, excessive throw (particularly elevator), incidence, C-of-G, lateral balance,etc. -- standard stuff, but it eliminates many problems straight away.

Thrust line issues are relatively easy to determine and to correct. That is usually where I start in-flight investigations -- if I suspect too much, or too little thrust setting, I test it by adding or subtracting thrust angle with shims & observe what happens. Following that, I start to systematically eliminate the remaining problems -- if I can.

Speed-dependent trim issues (not related to thrust line) are the next easiest problems to fix, so that's where I go, & for normal level flight (students airplanes) it isn't too hard to get things close.

It is for coupled aerobatic misbehaviour that things get interesting & frustrating. The problems are designed-in at that point & it is damned difficult to fix some of them (assuming that I can actually isolate a particular causitive agent), but I try it one step at a time.

I guess that's not very helpfull, because I'm sure that is what you do as well.