I think timbre matching is important but like you I've noticed many HT speaker sets that obviously can't be timber-matched perfectly due to differences in design between the speakers.
I think voice matching is more important than timbre matching, and I think voice matching is not the same thing at all as timbre matching. I'm not exactly sure what the actual terminology is.
At home right now I have 2 different speakers, Bohlender-Graebener and Infinity Kappa, both with similar ribbon/cone hybrid designs. They have VERY similar timbre qualities. Voices and constant notes sound the same. However when I try to play them simultaneously they sound awful. I think the drivers are out of phase or something. And the crossovers are different. I think this is the main thing with voice-matching. The raw drivers used should all be the same for equal transient response, and the crossovers better make sure all the speakers are phase-aligned or else when you play them together it will sound weird and awful.
But timbre matching is desirable also but I would say it is much more critical to have all your speakers operating in phase, etc.