Most failed products aren't badly designed or badly built. They're well-executed answers to a question nobody checked. Product design is the discipline that asks the question properly: who has this problem, how are they solving it today, what would make them switch, and what is the smallest thing we can put in front of them to find out.
We start with discovery — customer interviews, market and competitor review, analytics on the existing product if there is one — and turn what we learn into a prioritised set of opportunities rather than a feature list. Each one gets sized against user value and engineering cost, so the roadmap you end up with is a sequence you can defend to a board, not a list of everything anyone has ever asked for.
Then we validate before committing. Design sprints and clickable prototypes let you test the risky assumption in a week rather than discovering it in month five. What we hand over is a definition of the first release that everyone has agreed to, a roadmap for what follows, and the evidence behind both — plus a measurement plan so the next round of decisions is made on data rather than on the same instincts that produced the wish list.