game-playtest

Start by writing 3-5 specific, observable, actionable test questions, then build a per-player tracking form with timestamped observation fields. Set up screen capture and face camera recording, prepare a clean save state, and recruit at least 5 testers who are not on the development team.

A minimum of 5 testers per session is required for patterns to emerge over individual quirks. Findings classified as Major or Critical should be supported by behavioral evidence from at least 3 of 5 testers.

Ask what the game was about, what they were trying to do, which moment stood out as good, which felt confusing or frustrating, and what one thing they would change. Avoid asking if they liked it or would buy it, since social pressure distorts those answers.

No. AI agents like nunu.ai or modl.ai style bots are useful for regression testing, coverage checks, and softlock detection, but they cannot measure emotional response, aesthetic feel, or first-impression reactions. Use AI for coverage and humans for experience evaluation.

Have testers record with OBS or built-in OS tools, optionally with webcam, and distribute builds via a private itch.io link or Steam key. Block 90 minutes per session: 10 for setup, 60 for play, and 20 for a debrief captured in a shared form or document.

Skip playtesting when the build is so broken that testers will spend their time hitting bugs, when you only need to test an isolated mechanic, or when you want marketing testimonials. Fix critical bugs first, then observe behavior.