I'm thinking maybe distance between speakers could be optimized first, followed by distance from back wall, followed by toe in?Sounds like a reasonable plan to me. Although I think by "back wall" you are referring to the wall behind the speakers, which is referred to by most audiophiles as the front wall.
Speaking of the back wall, though, meaning the wall behind the listening position, a point which might eventually prove to be significant is that reflections from that wall will tend to produce a dip in frequency response at frequencies (in Hertz) in the vicinity of about 281.5 divided by the number of feet between that wall and the listener's ears. So if that distance is around 3 feet, as appears to be the case, reflections from that wall will cause a suckout, to some degree, in the vicinity of 94 Hz or so.
The 281.5 figure, btw, corresponds to 1/4 of the speed of sound in typical indoor environments, in feet/second. Rear wall reflections at a frequency of about 281.5 divided by the distance to that wall in feet will arrive at the listener's ears exactly out of phase with the direct sound arrival. The resulting suckout can be quite pronounced IME.
Best regards,
-- Al