How we decide what to build next
Every prototype gets two weeks and one question. If we can't answer the question, the prototype dies — and that's a success.
We build a lot of things that never ship. That is intentional, and we think the process is worth describing because it is the main thing that keeps a small studio from sinking eighteen months into a bad idea.
Two weeks, one question
Every prototype starts with a single question written at the top of the design doc. Not a pitch, not a genre — a question we genuinely do not know the answer to.
For Fortress it was: what happens if you can't capture by moving? For Signal Drift: is a communication gap fun if we never let players close it?
Two weeks later we either answered the question or we did not. If we did not, the prototype stops. No extensions.
Killing things is the point
Most questions turn out to have boring answers. That is useful information and it cost two weeks instead of a year.
A prototype that dies on schedule is doing its job. A prototype that gets extended three times is a decision nobody wants to make.
The ones that survive earn a second round with a harder question. Fortress went through five.
Questions about this post? Write to [email protected].
← All newsKeep reading
