Launch was v1, not the end. With no PM pulling me toward new features, I spent significant time after launch doing something most teams deprioritise: reading the friction data and fixing small things systematically. Over four years, they compounded into the product quality that mattered at acquisition.