Minutes 0–5: reach the actual experience

Game lane: get from launch to the first genuine NPC exchange. Note load time, whether the key setup (if any) worked first try, and whether dialogue is generated or canned — ask something a script cannot anticipate, like advice about a hobby you name. Chat lane: reach one real reply in a private browser window without surrendering an email. Pick a premade character; the builder is a feature tour, not the test.

In both lanes the first measurement is identical: did you reach the advertised experience without a wall the marketing never mentioned? Write down the wall if you hit one. The wall is data.

Minutes 5–10: instruction pressure

Give the character one benign fact, one preference, and one boundary in a single message. Something like: call me Ray, I collect film cameras, keep this conversation non-explicit. Spend the next few exchanges off-topic, then circle back — “what would suit my collection?” — without repeating yourself. A coherent product uses the fact; a shallow one has already dropped it.

Then apply gentle pressure to the boundary: request a warmer tone while restating the limit. Respecting a stated boundary under a tone shift is the single most revealing behavior in this niche, in either lane. A product that ignores it in minute eight will ignore it in week three.

Minutes 10–13: break something on purpose

Game lane: alt-tab out, come back, and if saving exists, quit and reload. Does state survive? Does the NPC remember the run? Chat lane: refresh the page, then close and reopen the tab. Note precisely what persists — character, history, your fact — and whether an account prompt now appears, and what it claims to unlock.

Do not repair the session by re-pasting context. You are grading the product’s continuity, not your own patience.

Minutes 13–15: read the money screen cold

ItemWrite downDo not assume
Upgrade or purchase promptNamed feature and shown priceThat paying fixes what you disliked
Credit or API meterBalance and reset/refill ruleThat credits map 1:1 to messages or rounds
Exit routeExact cancel/delete/revoke stepsThat uninstalling stops billing

Player questions

Why only fifteen minutes?

Long first sessions let novelty and attachment overwrite judgment. A timer keeps the evaluation about the product’s behavior instead of your mood at minute forty.

What if the product fails one check but I enjoyed it?

Then keep it with open eyes — as entertainment, not as the winner of a test it failed. The playtest exists to make that distinction visible before money moves.