The value of milk hinges on its physicochemical functionality under processing conditions. We examined composition-functionality relationships with individual milks from 24 New Zealand dairy cows, sampled at 3 times over the season. Milks were classified into type A or B, according to the shape of 3-point heat coagulation time versus pH profiles. Milk type changed over the season for half of the cows in the study. Best subsets regression suggested that different factors controlled heat stability in the 2 milk types. Urea concentration was key for both types, but for type A milks, osmotic pressure and milk solids were the most important predictors of heat stability, whereas casein micelle size and ionic calcium predicted heat stability for type B milks. This study revealed that milk type is prone to change over the season, and the findings suggest that optimizing heat stability could be achieved by different means for type A versus type B milks.