If you develop a simple theory about how much water people drink in the summer relative to how much they drink in the winter, will your theory be correct for all people? If not,is your theory worthless in describing human behavior?
No theory about human behavior is ever likely to accurately describe all people. That, however, does not mean that the theory will be useless. To think about this, let us look at the theory that you ask for in this question.
Of course, the theory in this question would be that (at least in places where there are significant differences in temperature between winter and summer) people will tend to drink more in the summer. We will make this hypothesis because we can imagine that hotter weather leads people to sweat more and, therefore, to drink more. However, this will probably not be true for all people. There might be some people who do not spend much time outside and who exercise vigorously (indoors) in both summer and winter. These people would not reduce the amount that they drink in the winter.
Even so, it is useful to create such theories. If the theory is correct, it will describe the behavior of the typical person. If it does this, it can, for example, help a business know how much bottled water to stock in the winter compared to the summer. It will not capture the behavior of each individual, but it will predict aggregate behavior fairly well.