For about five years I ran a small studio and we built a commercial narrative game, Book of Aaru, and shipped it on Steam. Unity, C#, a team that hovered between five and eight people, and a scope that was, in hindsight, heroic and slightly insane.
It still sells. I closed the studio anyway.
What a game teaches that a SaaS doesn’t
Games are unforgiving in a way business software isn’t. Nobody needs your game. There’s no procurement cycle, no switching cost, no “we already pay for it.” It has to earn every second of attention on feel alone.
That does something to how you build:
- Taste becomes a hard requirement, not a nice-to-have. A janky button in a B2B tool is Tuesday. A janky moment in a game breaks the spell, and the spell is the whole product.
- You ship the emotion, not the feature. The changelog nobody cares about; the way it feels is everything.
- Scope is the enemy. Five years on one thing teaches you, viscerally, what “the smallest thing that works” actually costs when you don’t respect it.
Why I walked
Because I’d learned what it had to teach me, and holding a studio open past that point is just inertia wearing the costume of commitment. The game earns its keep. The team went on to good things. And I got back the one thing a five-year project quietly eats: range.
I’d rather carry the taste forward into the next bet than defend the last one. That instinct, knowing when a thing is done even when it’s still working, is the most expensive lesson in the whole graveyard.
This was the goodbye. The build post-mortem (the restart, the allocation mistakes, the 1%-loop lesson) is in We rebuilt the whole game once. The mistakes survived the rebuild.