Playtest survey questions that get real answers
A copy-paste bank of playtest survey questions for indie games, plus the rules for writing questions that do not produce polite, useless answers.
Good playtest survey questions ask about what the player did, not whether they liked it. Ask about specific moments ("what were you trying to do when you died the first time?"), keep the post-session survey under eight questions, and put one question before play to capture expectations. Every question you write should be one you would change the game over, and if you would not, cut it.
Below is the bank I actually use, organised by when it gets asked. Take what fits and delete the rest.
Why most playtest surveys produce useless answers
Four failure modes cover almost all of it.
The question asks for a verdict. "Did you enjoy the combat?" invites a yes from anyone who does not want to upset you. "What was the last fight you lost, and why do you think you lost it?" invites a story you can act on.
The question asks the player to be a designer. "What would you add?" gets you a wishlist shaped by whatever they played last week. Players are reliable witnesses and unreliable architects. Ask about their experience, then design the fix yourself.
The question arrives too late. Confusion decays fast. A player who spent ninety seconds hunting for the interact key in minute three will genuinely rate the controls as fine in minute twenty, because they figured it out and their brain has already filed the struggle away. This is the strongest argument for asking something during play rather than only after.
There are too many questions. Answer quality falls off a cliff somewhere past eight questions. The eleventh answer is a shrug, and it is also the one you needed most.
What should you ask before they play?
Two or three questions, no more. The point is context you can slice results by later, plus one question you can only ask before the game spoils the answer.
- How much have you played games in this genre? (multiple choice: never / a little / a lot)
- What are you playing on? (multiple choice: desktop / laptop / phone / tablet)
- Based on the title and the screenshot, what do you expect this game to be about? (open text)
That last one is the most undervalued question in playtesting. It tells you whether your capsule art, your title, and your store framing set up the game you actually made. When expectation and experience disagree, players report the game as bad when it is really mismatched.
What should you ask during play?
Mid-session questions are the ones that catch the thing a post-survey never sees. Keep them to one sentence with one tap of an answer, and trigger them on an event rather than a timer where you can.
| Trigger | Question | Format |
|---|---|---|
| First death | Did that death feel fair? | Yes / No |
| Third death on the same encounter | What are you trying to do right now? | Open text |
| Tutorial complete | How clear was that? | Emoji scale |
| First time opening a menu | What did you come here to do? | Open text |
| Boss defeated | How did that fight feel? | Emoji scale |
| Idle for 45 seconds | What are you looking for? | Open text |
| Level complete | How hard was that section? | Rating 1 to 5 |
| Optional item found | Did you find that on purpose or by accident? | Multiple choice |
Two rules keep these from becoming an annoyance. Never fire more than three or four in a session, and never fire one during anything that needs the player's hands. A question that interrupts a fight is a bug report about you.
This kind of trigger is exactly what the PlayProbe Unity SDK does: you emit an in-game event, and the survey appears in the moment, then lands in the same results dashboard as your pre and post answers. The setup is covered in how to playtest a Unity game. On other engines you can get most of the way there with a screen recording and a post-survey question that references a specific moment.
What should you ask after they play?
Order matters here. Open with the broadest, cheapest question so people warm up, put your hardest question third or fourth while attention is high, and end with the one open text prompt that people write real paragraphs into.
The core eight
Use these as a default and swap two or three for whatever your build is actually testing.
- Overall, how was that? (rating 1 to 5)
- In one sentence, what is this game about? (open text)
- Was there any point where you did not know what to do next? (yes / no)
- If yes, what was happening at that moment? (open text)
- How difficult did you find it? (multiple choice: far too easy / a bit easy / about right / a bit hard / far too hard)
- What was the best moment? (open text)
- Would you keep playing if there were more? (yes / no)
- What is the first thing you would change? (open text)
Question 2 is the sleeper. If five testers describe five different games back to you, your pitch is not landing, and no amount of balance work fixes that.
Question 7 is worth more than a satisfaction score. Someone who rates you a 3 but would keep playing is telling you the foundation works. Someone who rates you a 4 out of politeness and would not keep playing is telling you the opposite of what the number says.
Swap-ins by what you are testing
Onboarding and clarity
- What did you try first, and did it work? (open text)
- Did you know what the goal was in the first minute? (yes / no)
- Which controls did you discover on your own? (multiple choice, list them)
Difficulty and balance
- When you died, did it feel like your mistake or the game's? (multiple choice)
- Was there anything you could not beat and gave up on? (open text)
- Did any strategy feel obviously better than the others? (open text)
Systems and progression
- What did you spend your currency on, and why that? (open text)
- Which upgrade did you want that was not there? (open text)
- Did you understand what the upgrade actually did before buying it? (yes / no)
Art, audio and feel
- How did the game feel to control? (emoji scale)
- Was there any sound you wanted to turn off? (open text)
- Did anything look like it was interactive but was not? (open text)
That last one finds more UX problems per word than any other question I know.
Store and marketing readiness
- Would you have clicked on this game in a store listing? (yes / no)
- How would you describe it to a friend who likes this genre? (open text)
Question format cheat sheet
| Format | Best for | Watch out for |
|---|---|---|
| Rating 1 to 5 | Trend lines across builds | Meaningless in isolation, only compare to your own previous test |
| Yes / No | Clean splits you can filter results by | Loses nuance, so always pair with a follow-up open text |
| Emoji scale | Quick reactions, especially mid-game | Not precise enough for balance decisions |
| Multiple choice | Anything where you already know the plausible answers | Writing options that miss the real answer, so add "other" or an open text nearby |
| Open text | Everything you did not anticipate | Expensive to read at volume, so keep it to two or three per survey |
PlayProbe supports all five of these formats in pre, post, and mid-game surveys, and aggregates the closed formats into distributions automatically so you are not counting rows by hand. Open text answers can also carry tags, which makes the pile of paragraphs sortable when you have thirty of them instead of five.
Rules I hold myself to when writing questions
- Never ask two things in one question. "Was the combat fun and responsive?" produces an answer to whichever half the player noticed.
- Never suggest the answer. "How much did you enjoy the new dash?" has already told them the dash is good.
- Ask about the last time, not the general case. "The last time you got stuck, what happened?" beats "do you often get stuck?"
- Make every question survive the "so what would I change?" test. If no answer to it would alter your next build, delete it.
- Keep the wording identical between builds. If you rewrite a question, you have started a new dataset and lost your trend.
- Say how long the survey takes and be honest about it.
Turning the answers into changes
A survey is only worth writing if it produces a decision. Read the closed questions first to see the shape, then read the open text looking for the same problem described in different words by different people. Three people saying "I did not know where to go" and one saying "the level is a maze" are one finding, not two.
For the full session structure around these questions, start with how to playtest your indie game. If you do not yet have people to send the survey to, how to find playtesters covers where I go looking.
If you want these questions to live somewhere other than a Google Form, PlayProbe is built for exactly this: pre, post, and mid-game surveys in one flow, one link for testers, no account or download on their side, and the results aggregated for you. It is free to start, and the question bank above will fit inside the free plan comfortably.