The traditional profit/utility maximising models are only useful up to a point, beyond that they are garbage, because many people are generally "satisfied" at a given level of a specific variable. A recent article in Spiked magazine dissed a couple of serious attempts to answer the question, both the "enoughism" of John Naish and the academic analysis of Oliver James were rejected as anti-consumer defeatism, though I think each has its merits.
The British Psychology Society offers plenty of theories, but I think the true answer is probably to be found in the relationship of mood-influencing chemicals to neural networks. Yes, you don't need to be a brain scientist to be happy, but you need to be a brain scientist to understand happiness.
But what a marketer might say is that a good brand can make you happy.