I think that having matched cells is less important than people think. In fact, if you do the math, if you have unmatched cells, it is better to parallel them at a cell level than to have multiple 4s packs with unmatched cells.
Adding resistance in parallel;
R= 1/( 1/R1 + 1/R2 + 1/Rn)
If you have a number of cells (say 12 cells) with a wide range of internal resistance, and you group them into 4 groups of 3, then add the resistances of the cells in each group, the spread of resistances will get much closer together. It is impossible for that to not happen. If you actually know the IR of each cell, you could group them so that all 4 groups had the same IR, even though the individual cells could vary significantly. So the 4 "cells" in your 3p4s pack would be balanced.
Having the cells in parallel assures that the voltage of the cells is kept the same. So, while the SOC of each cell might vary a bit (but not much)in the flat region, as soon as the battery gets into the knee (either charging or discharging) the cells will balance to the same SOC. All the paralleled cells will reach 2.5V or 3.65V at exactly the same time. This would work even with very different cells, even different capacities. You can parallel a 200Ah cell with a 280Ah, and you will get 480Ah.
If you built multiple 4s packs with unmatched cells, each pack could have radically different IR values, and might not share the load equally.
Now, real world, not theory: It doesn't matter. You probably don't have the capacity to measure the differences of what option may or may not work better. If you are pushing the cells *very* hard, charging/discharging at their max rate, then it might make a difference on whether or not you exceed a single cells charge rate based on the IR of the cell it is paralleled with. But at any "normal" rate, it won't matter, and everything will be good again when the cells get into the knee.
One possible answer to "bad things often happen" isn't really a matched cell issue, but a state of charge issue. Before you connect 2 cells in parallel, make sure they are at (nearly) the same voltage.