Someone says your game is “fun,” then cannot find the start button. The compliment is kind; the pause tells you what to investigate. Give a new player one goal, watch without guiding them, ask what they expected, and make one change you can test again. One person and one short round are enough to start. This guide begins after you have a playable version; if you are still choosing the mechanic, start with a game prototype.
What should you test in a game playtest?
Do not ask a single session to prove that the entire game is good. Choose one uncertainty you can observe: “Can a first-time player start a round without help?” “Do they notice when a point is awarded?” or “Can they tell when the game is over?” That question determines the task, what you record, and which change you make afterward.
For a small alignment puzzle, the test might be: “Can someone recognize that the crooked object should be dragged, not tapped?” Give the player the game and say, “Try to finish the first puzzle.” Do not tell them which object to move. For a party quiz, ask them to finish one round and explain how they think the score was calculated.
Write down the device, game version, and test question before anyone starts. A phone test can reveal small targets that a desktop test misses; a desktop test can reveal hover behavior a phone cannot use. Keep the first session short enough that you can watch every action.
Use a playable game to plan your first playtest
Perfect Grid is a public Eazo puzzle with five compact levels. Its first level asks the player to correct a slight visual misalignment by dragging. Try it once without reading an explanation. Then write down your first tap or drag, what you expected to happen, and what the screen actually showed. It is a quick way to practice observing an interaction before you watch someone play your own game.
For another small interaction to compare, Riso Nori Match starts with a card-flipping action and shows counters for flips and matches. In your own game, choose one signal to test: the start cue, the response after an action, or the ending. Watching these Apps can help you decide what to look for, but it cannot tell you how players will use yours.
How to run a ten-minute game playtest
Use this short script:
- Set the task: “Try one round and stop when you think you are done.”
- Watch: Note the first action, any pause, a repeated wrong action, and whether the player notices feedback. Stay quiet long enough to see what they try next.
- Ask after the moment: “What were you expecting to happen there?” or “What did you think that control would do?”
- Finish: Ask the player to describe the goal and ending in their own words.
Record actions separately from interpretations. “Tapped the score twice before tapping the cookie” is an observation. “The score is too prominent” is a hypothesis to check. Nielsen Norman Group's guidance on talking with test participants explains why giving hints or asking leading questions can change the behavior you are trying to learn from.
| Record this | Example note | What it tells you |
|---|---|---|
| Action | Player taps the crooked object twice instead of dragging it. | The intended drag may not be clear. |
| Words | “I thought tapping would straighten it.” | Their expected action differs from the one you designed. |
| Context | Phone, first attempt, no instructions. | The finding applies to this setup, not every player. |
| Possible change | Add a short drag cue beside the object. | A testable idea, not a proven fix. |
Action
- Example note
- Player taps the crooked object twice instead of dragging it.
- What it tells you
- The intended drag may not be clear.
Words
- Example note
- “I thought tapping would straighten it.”
- What it tells you
- Their expected action differs from the one you designed.
Context
- Example note
- Phone, first attempt, no instructions.
- What it tells you
- The finding applies to this setup, not every player.
Possible change
- Example note
- Add a short drag cue beside the object.
- What it tells you
- A testable idea, not a proven fix.
The rows above are illustrative. Replace them with what you actually see in your own session, including anything that contradicts your first guess.
Playtest feedback questions to ask players
“Was it fun?” is hard to act on by itself. Ask what the player tried to do, what they expected, and which moment made them hesitate. If they say an action was boring, ask when it became repetitive. If they say the game was confusing, ask them to show where they lost track. Do not ask, “Would a brighter button fix it?” before you know whether they saw the button at all.
You can ask a Skill to help organize your notes, but treat its output as a draft. In Eazo, browse Skills → Community for a relevant testing or review method, read its scope, then use / in Chat to insert it. Add a concrete goal: “Turn these anonymized observations into three possible causes and one question for the next test. Do not invent participants or results.” If no suitable Skill is available, the table above is enough.
How to use playtest feedback to improve your game
Open your own game in Eazo Creator and reproduce the issue in Preview. If the layout needs work, compare a few design directions or attach a style from Favorite designs; explain which element should be clearer. A saved style changes visual inspiration, not the rules or scoring by itself.
In Preview, you can turn on Comment, select the confusing element, and describe the problem at that location. This is useful when “the start button” could mean more than one control on the screen.
Then give the Agent a focused revision prompt:
In my alignment puzzle, a new player tapped the crooked object instead of dragging it. Keep the current puzzle rules. Add a short, dismissible “Drag the object to align it” cue beside the object on the first level. Do not change later levels. I will test whether the next new player tries dragging without help.
The observation table and prompt are an example for your own puzzle, not changes to Perfect Grid. After making a change, ask another person to try the same task. Save screenshots and notes so you can compare versions. If that person succeeds, keep testing as you make other changes.
Publish your game and test the updated version
Once someone can complete the core round and the main confusion is resolved, use Publish to create a stable production link. Open that link yourself on the intended device. After later edits, use Update before asking new players to test the revised public version. A preview or sandbox link is useful while building but is not the link to rely on for ongoing sharing.
Keep the observation sheet. Your next playtest can ask a different question—perhaps whether the ending is satisfying—without losing track of why you changed the first screen.
