I totally agree. I just today shipped code that lets clients off board, nearly two years after launch.
It is really hard as an engineer not to complete a feature, but for a small scrappy saas product that's exactly what you have to do.
I liken it to building a bridge while your clients are walking across it. You want to be fair enough ahead that they don't fall in, but not so far you take them the wrong direction or build paths they don't go down.
I've run engineering teams in SFBA for years and have, with a friend, built Cronitor. I've come to believe the most valuable "soft" skill in an engineer is just a relentless drive to ship, to know how to be both tactical and strategic in avoiding over-engineering.
With Cronitor, we call it the JIT Mindset: Don't actually write code until you have to, and focus on solving most impactful problems first. My co-founder wrote more about it here -- https://blog.cronitor.io/the-jit-startup-bb1a13381b0
In one sentence it's simply: Over-engineering something can be hard which can trick you into thinking it's real valuable work -- it's not. It's a trap.