How to playtest your first video game

Nov 18, 2026|5 Min

Unity Team - Unity Technologies

Game design
An abstract game window with three floating platforms. Blue dotted arcs show the intended jumps, while an orange dotted path falls short into the gap and ends in a glowing ring. A dark notes card beside the window lists three lines, the first marked in orange. Image generated by Claude Opus 5.5, November 9, 2026.

• A playtest is a session where someone who didn’t make your game plays it while you watch.

• You don’t need to be a developer to run one: you need something playable, one or two questions, and a few people who aren’t you.

• Start testing as soon as the main thing players do works, not when the game feels finished.

• Stay quiet while people play. Where they get stuck can be the most useful thing you learn.

• When you build with AI, turn what you saw into specific changes, one at a time, then test again.

A beginner’s guide to playtesting your video game: what playtesting is, why it matters, how to run a session, and how to act on feedback when you build with AI.

To playtest your game, put it in front of people who didn’t make it, watch what they do, and change the game based on what you see. You don’t need a team or a budget to do it. You just need something playable, one or two clear questions, and a few people who aren’t you.

This guide walks through playtesting for anyone making a first game, including people building with an AI game maker like Unity Spark.

Once you’ve played your own game a few times, it gets hard to come to it fresh. You know where to go, what each button does, and which jump needs a running start. A new player knows none of that. A playtest shows you how the game feels to someone new, and tells you what to fix before you spend more time building.

What is playtesting?

Playtesting is watching someone who didn’t make your game play it, so you can learn how it feels to a new player. During a session, a tester plays while you note where they get confused, get stuck or lose interest. Then you use what you saw to improve the game.

It helps to think of testing as two jobs. The first is checking that the game works: nothing freezes, no button stops responding, no sudden stutters, no way to fall out of the world. The second is checking that it works for people: whether they understand what to do, and whether it feels the way you hoped.

You can do a lot of the first job yourself. The second job needs other people. That’s the playtest.

Why does playtesting matter?

You’re the one person who can never play your game for the first time. Because you know how it’s meant to work, you fill in the gaps without noticing them. Outside players catch what you’ve stopped noticing, and they help cut through assumptions you may not know you’re making.

Timing matters too. Feedback on how the game plays is easy to act on early. The same feedback, arriving after you’ve built levels, menus, and a story around a confusing core, can mean redoing a lot of work.

When should you start playtesting?

Start playtesting as soon as the main thing players do works, even if the art is placeholder and there’s no sound yet. If players will spend most of their time jumping, test the jumping. If they’ll spend it matching tiles, test the matching. Waiting until the game feels ready can mean building for a long time on something that doesn’t work for players.

If you haven’t pinned down that main action yet, start with how to find your core game loop and a one-page design document for your first game.

New to making games?

How do you run a playtest?

Here’s a simple process you can repeat every time your game changes.

Playtest loop: Diagram titled “Spotlight on the playtest loop.” Five numbered steps in a row, linked by arrows: decide what to learn, make a short test version, watch people play, look for patterns, and change one thing. An arrow labeled “Back to the start” returns from the last step to the first. A note reads: keep your questions the same between rounds so you can see whether a change worked. Image generated by Claude Opus 5.5, November 17, 2026.

A playtest loop: Each round of playtesting feeds the next, so small changes add up. Image generated by Claude Opus 5.5, November 17, 2026.

1. Decide what you want to learn

Before each playtesting session, write down one or two questions you want to answer. Sessions without a goal tend to produce vague feedback. Good questions are specific, and you can answer them just by watching:

• Can players work out how to jump without being told?

• Do they head toward the exit, the way you designed the level to guide them?

• Does the first enemy feel fair on a first try?

• Does the opening create the feeling you set out to create?

• Can players tell when they’ve won or lost?

2. Make a short test version

Build a version of your game around those questions. For many games, 15 to 30 minutes of play is enough. Include only what you want to test, and make sure the main action doesn’t crash. A simple objective screen or a few lines of instructions keeps testers from getting stuck before they reach the part you care about. At this stage, working matters more than looking finished.

3. Try to break it yourself first

Before anyone else plays, try to break your game on purpose. Most people test their own game the same way every time, so play as a few extreme types of player instead:

• The cautious player: Avoids every risk. If they win without trouble, your challenge may not be doing its job.

• The all-in player: Uses one strategy and ignores everything else. If that strategy always wins, the other options may need rebalancing.

• The random player: Picks actions without thinking. If random play beats a game that’s meant to reward skill, something’s off.

• The wanderer: Ignores the goal and does nothing for a minute. Does the game still make sense?

• The button masher: Presses everything at once. Does anything get stuck?

Write down what you tried and what broke. A simple log with the date, what you tested and what you noticed makes patterns much clearer later.

Diagram titled “Spotlight on five ways to break your own game.” Five cards, each with an icon: cautious (a shield, avoids every risk), all-in (a target, uses one strategy and ignores the rest), random (a die, picks actions without thinking), wanderer (a winding path that misses a flag, ignores the goal), and button masher (four buttons, presses everything at once). Image generated by Claude Opus 5.5, November 17, 2026.

Player types: Playing like someone else is the fastest way to find what breaks. Image generated by Claude Opus 5.5, November 17, 2026.

4. Find a few people who aren’t you

Friends and family are a good place to start when looking for playtesters, though they’re often too kind to tell you what’s wrong. Mix people who play a lot of games with people who don’t, since they’ll give you different results. If you can, include players with different needs too (accessibility basics covers what to look out for).

It also helps to mix people who’ve heard about your game with people who haven’t. Those who know about it come in with expectations, and those who don’t show you what a stranger would make of it.

How many people should playtest your game? For early tests, three to five usually surface the most common problems (treat this as a rough guide for games, not a rule). Several small rounds work better than one big one: fix what the first few people find, then test again. If your game is for distinct groups of players, test a few people from each.

When you want more testers, look for online communities where players expect to try early games, such as forums and chat servers. Share a short version, say what feedback you want, and reply to comments. A small thank-you, like a credit in the game, can help. Testing is real work, so respect people’s time.

5. Watch without helping

You can run a session live, with the tester sharing their screen or sitting next to you. Or you can ask them to record themselves playing and send you the video with a few answers. Live sessions show you every moment of confusion as it happens. Recorded sessions let you test with more people than your calendar allows.

Either way, the rule is the same: don’t help. When someone gets stuck, resist explaining the controls or pointing to the door. Their confusion is what you came to learn, and stepping in erases it. Ask them to think out loud, so you hear what they’re trying to do, not just what happens.

If you ask for recordings, give clear instructions, even for steps that seem obvious. And ask permission if you might share clips later.

6. Ask a few focused questions afterward

Skip “What did you think?” Instead, ask three to five questions tied to what you wanted to learn, such as “What was the first moment you felt lost?” or “What moment stood out, and why?” With a small group, questions people answer in their own words tell you more than a one-to-five rating.

7. Look for patterns, then change the game

One person struggling with something is an opinion. Three people struggling with it in the same way is a problem worth fixing. Give repeated observations the most weight. A one-off comment about taste, like a color palette someone doesn’t love, can wait unless it matches your own instincts.

Testers are good at showing you where a problem is. They aren’t always right about how to fix it, so treat their suggestions as clues rather than instructions. After any big change, test again: a fix in one place can move a problem somewhere else.

It’s normal to feel strongly when you read feedback, good or bad. It’s about the game as it is today, and the game will keep changing.

How do you playtest a game you build with AI?

Playtesting a video game stays basically the same when you build with an AI game maker like Unity Spark. What changes is how you turn what you saw into the next version of your game. When you shape a game by describing what you want, a few habits help each playtest turn into clear changes:

  • Describe what players did when prompting edits: “Three testers walked past the door without noticing it” gives you far more to work with than “make the level better.” It also points to a specific change, like making the door stand out.
  • Change one thing at a time: If you change several things between rounds, you won’t know which one helped. Make one change, test it, then make the next.
  • Keep your questions the same between rounds: Asking the same one or two questions each time shows you whether a change worked.
  • Keep a log of each round: Note the date, what you changed, and what testers did. It helps you spot patterns and undo changes that made things worse.
Playtest log example

A simple log shows whether each change helped.

Round and date
Round 1, June 2
What changed
Nothing yet – first test version
What testers did

Three of five walked past the door

Round 2, June 9
What changed
Made the doorway brighter
What testers did
One of five walked past the door
Round 3, June 16
What changed
Added a light above the door
What testers did
All five found the door

A video game playtest checklist

Before you invite anyone to playtest your game, run through this list:

  • I’ve written down one or two questions I want answered
  • My test version is short, works, and covers only what I want to learn
  • I’ve tried to break the game by playing in ways I normally wouldn’t
  • I have at least three testers who aren’t me, with a mix of experience
  • I’ve decided whether the session is live or recorded, and written clear instructions
  • I know I won’t help when someone gets stuck
  • I have three to five focused questions ready for afterward
  • I’ll look for patterns across testers before I change anything

Make playtesting part of how you build

Playtesting works best as a habit you repeat while your game grows: make something playable, watch a few people play it, change one thing, and test again. Each round shows you more of what a new player experiences, and you don’t need to be a developer to run one.

Ready to make a game worth testing? Unity Spark is a browser-based game maker from Unity. Start with prompts, build with artist-created assets from the Unity Asset Store, and refine your game without writing code. It’s heading into closed beta, and the waitlist is open now.

Learn more about Unity Spark

Header image note: An abstract game window with three floating platforms. Blue dotted arcs show the intended jumps, while an orange dotted path falls short into the gap and ends in a glowing ring. A dark notes card beside the window lists three lines, the first marked in orange. Image generated by Claude Opus 5.5, November 9, 2026.