Is it the voice coils you want to line up or a "center of push" point on each cone ? I'm guessing you would want a point perhaps 0.7 of the way out from the center of the cone, so that the cone area behind the line matched the cone area in front of the line.
Another approach would be to use small, not-very-deep drivers so that the time difference between woof, mid and tweet is pretty much trivial... which is the trend these days anyways.
Looking at the front of my M60s... we have perhaps 1/2 inch fore-and-aft distance between the cones of woofer and midrange, and perhaps 1/4 inch between midrange and tweeter.
Speed of sound is perhaps 13000 inches per second, so 1/2 inch makes a difference of 1/26,000 of a second. The crossover from woof to mid is about 200 hz, let's double that and say 400 hz. The phase difference from speaker alignment would be 400/26000 of a cycle, or about 6 degrees (actual calculation is 1/26000 second delay divided by 1/400 second for one cycle of the signal times 360 degrees). That would result in slight visible distortion of the wavefront but nothing more.
Mid to tweet alignment makes more of a difference but still fairly small :
1/4" misalignment is roughly 1/52,000 of a second difference in arrival time. Crossover is ~2.2 KHz, let's use 4.4 KHz for maybe 12 to 15 dB drop. Phase distortion is 4400/52000 of a cycle, or about 30 degrees, which would result in some noticeable distortion of the waveform.
BUT (and it's a big but) the mere fact that we are using cones as drivers means that we are getting a distribution in arrival times corresponding to the depth of the cone for the driver which carries that frequency range. The woofer cone is at least twice the depth of the woofer to midrange misalignment, and the midrange cone is about twice the depth of the midrange to tweeter misalignment. If you use a flat driver (a la the old KEF B139) my understanding is that you get all kinds of other intereresting effects from the edges of the flat baffle so that's not an easy answer either. My gut feeling is that the crossover elements have more impact on waveform distortion than physical alignment of the drivers anyways.
