My short answer: I did not change direction, I kept asking the same question in a different laboratory. Psychology taught me to ask why people behave the way they do. Advertising is one of the few fields where that question gets measurable answers every single day.
Three habits that carried over
Working from a hypothesis. The first thing you learn in experimental design is to write down what you expect before you start. Before a test I write what I expect and which result would prove me wrong. Without that, a test turns into storytelling after the numbers arrive.
Distrusting the instrument. In psychology, whether a measure actually captures what you intend is its own question. The digital equivalent: is that "conversion" in GA4 a sale, or a button click? Most of the "bad performance" cases I have seen were measurement problems, not performance problems.
Not trusting the average. The most durable sentence from statistics class. You may see 4 ROAS at account level, containing campaigns that range from 12 down to 0.8. Decisions come from the distribution, not the mean.
What it changes in practice
The clearest difference is in creative selection. Instead of "I think this one looks better", I ask which message matches which intent: a category with existing demand needs persuasion, a category being introduced needs explanation. The same image performs differently in each.
The second difference shows up in client conversations. When a number looks bad I ask first whether it is actually bad, or whether the expectation was set wrong. Those are different problems with different fixes.
Without romanticising it
A psychology degree did not teach me to write ads; I learned that in the field. What it gave me is a method: do not believe without measuring, do not generalise from one case, and treat your own intuition as just another hypothesis. Those habits are worth far more than the "behavioural marketing" label.