Quick answer: Bug counts, ticket aging, severity distributions - all of it correlates with one number: crash-free rate. Track that one well, and the others sort themselves out.

Studios obsess over bug count. A better metric is 'what percentage of sessions completed without crashing'. It correlates with everything else you care about.

Define a session

Session = launch to clean shutdown. A crash, a hang >30s, or a process kill all break the session. Define this in code, instrument both ends, ship the count to your backend.

Slice by build, region, hardware

Crash-free rate dropping 3% overall might be a single platform's regression. Always slice. A flat aggregate hides real signal.

Aim for 99% as a floor

99.5% crash-free rate is excellent. 99% is industry baseline. Below 95% is a fire. Treat the metric like uptime - report it weekly, hold a postmortem when it drops more than 0.5% week-over-week.

Correlate, don't replace

Crash-free rate isn't the only number. But a project where it's stable above 99% almost always has its other quality dimensions in order too.

“Players don't see your bug count. They see whether the game crashed.”

Crash-free rate is the leading indicator your CEO can understand. Surface it on a wall monitor; the next two sprints will reorganize themselves around it.