The fact that the average effect size in a population of an intervention is small doesn't mean that there aren't individual members in that population that benefit a great deal.

Over time I also think of mood as less of a one dimensional thing. People often change the way they judge their happiness, so you don't have a constant standard.

Some studies claim to examine "lifestyle changes", but they often do it in a really lackluster and low-compliance way

Getting compliance is really hard.