There's an underlying question to the whole endeavour: are conventional bra sizes doing a good job of a hard task, or a bad job of an easy task? Is it relatively simple to classify the shapes of women's bodies, and standard back/cup sizes are the wrong tool, or is aforementioned classification really hard, and standard sizing is doing a sterling job of it?

I suspect it's probably neither, and standard sizing is doing a terrible job of a hard task. I tried making a few simple topological models, and it turns out breasts are actually quite complicated, but there's no reason to assume the standard sizing scheme is optimal for the task, so gains can definitely be made somewhere. I don't think just trying to figure them out is an option, though. It's a question of how women's bodies categorise, and the only way to answer that is with the data.