So, let's take this scenario- I want a full TKL of one color, and the left shift doesn't make it and all other keys do. The entire set is useless to me without it, therefore I won't pay for my invoice. I'm sure this is going to happen, and thus the way this group buy is set up is doomed to fail- classic tragedy of the commons.
What if it was like this: we take this initial phase more as a vote of how many people are down for what colors/keys, and create sets made of the most popular combinations. To address keys that people would still get without needing entire sets, we can isolate keys that made it into the MOQ that didn't make it into sets, and ask people who signed up for them if they still want them without the other keys that made it, and then tally up the new total to see if they still pass the threshold. Also, certain less common keys that are within sets can be added at any quantity, to potentially give those sets an effective lower MOQ (like R4 2.25, R4 2.75, R3 2.25, spacebars, etc).
The other positive effects: lower barrier to entry, which will attract many more people. You'll be able to make renders, people can imagine the products on their keyboards much more easily.
There will be some people that are left unhappy, but most of these people would likely not have had their keys made anyway. It seems to me this strategy maximizes the total happiness across everyone.
What do ya'll think?