Two measurements around chest below breast, abd around breast across nipple. Measurement 1 is the number part of a bra size. The difference between measurement 1 and 2 gets converted to a cup, bigger difference equals larger cup size. So a 32D and a 46D both have boobs that protrude bast their chest the same amount, but visually it'd be much more noticeable on one woman over the other, but the 46D still has significantly more breast tissue.
Large breasted heavier women are almost never actually D cups. And iirc in the US 36D is the average, and the scientists who did that study find that increasing weights had no impact on rising bust sizes, but rather it seems more women are actually measuring properly now and realizing they were undersized all along.
Obviously with the general squishyness of the area of measurement, it's actually pretty easy to fuck it up, especially if no one's helping. And there's SO much misinformation about cupsizes that loads of women never wear a properly fitted bra their entire lives.
Add on to that, that different countries and companies do different cup size formulas (although there is nominal standardization at least within a given country, think shoe sizes) and it's a wonder all women don't just go braless or do wraps to spare themselves the headache.