Ask the boring questions first
Budget, platform, deadline. Nobody enjoys that conversation at the start and everybody pays for skipping it.
Studio

I started in mechanical design, where being wrong is expensive and obvious. That is still how I scope software: by asking what it has to survive rather than what it should look like.
Along the way I learned to render, then to shoot, then to build the thing being rendered. Now I make products and the whole apparatus around them, which means nothing waits on anybody else.
Route here
How the work runs
Budget, platform, deadline. Nobody enjoys that conversation at the start and everybody pays for skipping it.
A mockup that demos well proves nothing. I want the thing early enough that being wrong is still cheap.
A product that never leaves the folder is a hobby. Some of mine have been.