An empty room is rarely one failure. It is a trail of assumptions that went untested.

The work is learning which signal you missed, then building a system that catches it earlier next time.