I build products. I define the problem, decide what gets built first, and check whether what shipped actually gets used.
I care about one thing: models where you can say in advance what comes out for a given input. That goes for a product, a team, or a decision-making process. What can be predicted can be explained, and what can be explained can be fixed.
This is where I keep the calls I made along the way, and whether they turned out to be right. The wrong ones stay up too.