
plan and run a playtest at home with family members or other adults I know that respects their consent and privacy, observe without leading, separate what players did from what they said, and run a structured self-playtest when no tester is available.
Instruction
A playtest is a small piece of research with people. Professional studios run playtests with rules drawn from research ethics, and so will you. Good ethics and good data turn out to be the same thing: testers who feel safe and unpressured behave naturally, and natural behavior is what you need to see.
Who your testers are. You test in person, at home, with people you already know: parents, grandparents, brothers and sisters, cousins, or other adults your family knows and trusts. Never look for testers online: do not post your game on forums, chat servers or social media to find strangers, and do not create accounts to do so. Let a parent or guardian know when you are running a test. Three or four people you can sit beside will teach you more than a hundred anonymous comments.
Informed consent. Before anyone plays, tell them in plain words: what you are testing (“the game, not you”), roughly how long it takes, what you will record (notes, and a screen recording only if they agree) and how you will use it. Tell them they can stop at any time without giving a reason. Then ask whether they agree. Write this as a short script and read the same script to every tester, so every tester starts the same way, even when the tester is your own grandmother.
A parent’s OK for any tester under 18. If a tester is a minor, such as a younger brother or a cousin, their parent or guardian says yes first, before you observe them at all, and before you record anything. Tell the parent what the child will do, that you will take notes about the game and not about the child, and that the parent is welcome to stay in the room. Never record a minor’s face or voice without that parent’s OK. The safest default for every tester is a gameplay-only screen recording with no face, no voice and no camera. Write “parent OK: yes” and the date in your log, without names.
Privacy. Collect only what you need. Refer to testers as Tester A, B and C, not by name (not even “Grandma”), in your notes and reports. Store notes in your project folder on your own computer, not on social media or a shared public drive.
Confidentiality works in both directions. In industry, testers of unreleased games are often asked to sign non-disclosure agreements (NDAs), promising not to share what they saw. Designers, in turn, keep testers’ comments confidential and never mock them. At home that means: do not post what your testers said or did, and do not tell other relatives who said what.
Testing people who love you. Family members want you to succeed, so they tend to be kind rather than honest. Say clearly that honest reactions help you most, and then trust what they do more than what they say. A relative who says “it’s great!” after falling into the same pit five times has given you two pieces of data, and the five falls are the stronger one.
Content and comfort. Warn testers about anything potentially upsetting (horror, flashing lights, loud sounds) before they start, and offer a way to skip it. Make the test accessible: let testers use the controls they prefer, allow breaks, and do not treat a tester who struggles as a failure. A struggling tester is showing you a problem in the design.
Facilitating without leading. The hardest part is staying quiet. Give only the information a real player would have. When a tester asks “what do I do now?”, reply “what do you think you should do?” and write down that they were stuck. Do not say “oh, you just have to…”: every explanation you give hides a problem your players will hit when you are not there. Ask testers to think aloud (“say what you are looking at and what you are trying to do”) and prompt gently with “what are you thinking?” if they go quiet.
The self-playtest, when no one is available. Sometimes nobody is free. Then test yourself, but in a structured way: leave the build alone for at least a day; start a screen recording; play as if you had never seen the game, using only what the screen tells you; count attempts and note every pause; afterward, watch the recording and fill in the same observation sheet. Label these sessions S, not as a tester letter, and weigh them lower: you already know where everything is, so a self-test can find broken things and confusing text but can never tell you what a fresh player would understand.
Observation versus opinion. Record behavior and words separately. Observed: “Tester B died at the second pit four times; paused 6 s before the blue platform.” Said: “‘I didn’t know I could wall-jump.'” Behavior is strong evidence; opinions are useful but weaker, because people often cannot explain why they feel what they feel, and they may say what they think you want to hear. When a tester suggests a solution (“add a double jump”), write down the problem underneath it (“the gap feels too wide”) and solve that problem your own way.
Honest reporting. Report what happened, including the results that disappoint you. Choosing only the flattering comments is a form of dishonesty, and it leaves the real problems in the game.
Accepting criticism. As a designer, thank every tester, argue with none and decide later. When someone asks you to try their work, the same rules apply in reverse: be specific and kind, because you are describing your experience, not grading a person.
📖 Read for this lesson
The reading opens over the lesson; close it to come back to exactly where you were. The checks below draw on it.
Formative check
Work through these before moving on. They are not graded — they tell you whether the standard below has landed yet.
A tester says, “What am I supposed to do here?” What is the best facilitator response?
Your 13-year-old cousin says it is fine to record their face and voice while they play your game, so you may record them even though their parent has not said yes.
Which note is an observation rather than an opinion?
Write your consent script (no more than 90 words). Then run a five-minute practice playtest in person at home with one family member or other adult you know; if the tester is under 18, get their parent’s or guardian’s OK first and record nothing but the game screen. Record two observations and one quotation, labeled separately. If no tester is available, run a structured self-playtest after a day away and log it the same way, labeled S. Finish with one sentence on whether your tester (or you) felt free to stop or speak honestly, and what in your script helped.
How confident are you that you can run an ethical playtest and record observations separately from opinions?
Practice
Work these on paper or in your notebook, then open Check your answer. Aim for all of Fluency and Application; try at least one Challenge.
Printable version: this unit’s practice workbook (PDF)
Fluency
Build speed and accuracy with the core skill.
- List four things a consent script must tell a tester before they play.
Check your answer
Answer: What you are testing (the game, not them); roughly how long it takes; what you will record and how you will use it; that they can stop at any time without giving a reason. Then ask whether they agree.Read the same script to every tester so everyone starts the same way. - Who may be your testers, and what is never allowed when finding them?
Check your answer
Answer: People you already know, in person at home: parents, grandparents, siblings, cousins or other adults your family knows and trusts. Never look for testers online or create accounts to do so.Let a parent or guardian know when you are running a test. - Label each note as an observation or an opinion: (a) “Tester A paused 8 s at the fork.” (b) “Tester B: ‘The music is annoying.'” (c) “Tester C died at the spikes three times.” (d) “Tester A thinks it needs a map.”
Check your answer
Answer: (a) observation; (b) opinion (a quotation); (c) observation; (d) opinion.Behavior is strong evidence; opinions are useful but weaker.
Application
Use the skill in context. Show your reasoning.
- Your 10-year-old cousin wants to test your game while you record a camera video of them. What must happen first, and what is the safest way to record?
Check your answer
Answer: Their parent or guardian must say yes first, before you observe or record anything, after hearing what the child will do. The safest default is a gameplay-only screen recording with no face, no voice and no camera.Log “parent OK: yes” and the date, without names. - A tester says, “Just tell me how to get past this bit.” Write your facilitator response and what you note down.
Check your answer
Answer: Response: “What do you think you should do?” Note: the tester was stuck at that point (where and for how long).Every explanation you give hides a problem your players will hit when you are not there. - A tester says “Add a double jump.” Following the lesson, what do you write down, and what do you do with it?
Check your answer
Answer: Write down the problem underneath the suggestion, for example “the gap feels too wide,” and solve that problem your own way.Testers are good at noticing problems, less reliable at designing solutions. - Your grandmother says “It’s great!” after falling into the same pit five times. What two pieces of data do you have, and which is stronger?
Check your answer
Answer: Her opinion (“it’s great”) and her behavior (five falls at the same pit). The five falls are the stronger evidence.Family members tend to be kind rather than honest, so trust what they do more than what they say. - Rewrite this note so it protects privacy: “Grandma Rosa got lost at the fork and said my game was confusing.”
Check your answer
Answer: “Tester B hesitated at the fork for about 10 s. Said: ‘Which way am I supposed to go?'” (or a similar wording with no name, behavior and quotation separated).Use letters, not names (not even “Grandma”), and record behavior and words separately. Store notes on your own computer.
Challenge
Stretch problems. Expect to think before you write.
- Nobody is available, so you run a self-playtest. List the structured steps from the lesson and explain why the session is labeled S and weighed lower.
Check your answer
Answer: Leave the build alone for at least a day; start a screen recording; play as if you had never seen the game, using only what the screen tells you; count attempts and note every pause; afterward watch the recording and fill in the same observation sheet. It is labeled S and weighed lower because you already know where everything is, so it can find broken things and confusing text but not what a fresh player would understand.A strong answer gives the steps in order and the reason for the lower weight.
Review
Keep earlier skills sharp.
- Review (Lesson 1.2): your death data comes from the hazard’s print line. Why is that observation stronger than a tester saying “the spikes are unfair”?
Check your answer
Answer: It records where deaths actually happened, which is behavior; the comment is an opinion.Both are useful, but behavior is the stronger evidence.
Home Playtests: A Consent Script, a Parent’s OK, a Self-Playtest Protocol and an Invented Session Log
SLS original technique sheet for at-home playtesting; the session log and the names in it are invented and illustrative · SLS original
Why this reading. The lesson asks for a consent script of no more than 90 words, an observation sheet and a practice playtest at home, or a structured self-playtest when no tester is available. This reading gives a model consent script (with its word count), a model request for a parent’s OK when a tester is under 18, a self-playtest protocol, a blank observation sheet, an invented session log, and a table that sorts behavior from opinion so you can check your own notes.
Who can test: the home rules
- In person, at home, with people you already know: parents, grandparents, siblings, cousins, or other adults your family knows and trusts.
- Never strangers online. Do not post the game on forums, chat servers or social media to find testers, and do not create accounts to do so.
- A tester under 18 needs a parent’s or guardian’s OK before you observe or record anything.
- Default recording for everyone: the game screen only. No camera, no face, no voice.
Model consent script (81 words, SLS original)
Thanks for helping. I am testing the game, not you. You will play for about ten minutes and say aloud what you are thinking. I will take notes and, only if you agree, record the game screen; I will never record your face or voice. Notes are used only to improve the game, and you will be Tester B, not named. Honest reactions help me most. You may stop at any time without giving a reason. Do you agree to play?
It says what is tested, how long, what is recorded, how the data is used and that the tester may stop, and it asks for honesty, which matters when the tester is someone who loves you. Read it word for word to every tester, even family. Add a content note if your game has flashing or loud moments.
Model request to a parent, for a tester under 18 (68 words, SLS original; Sam is an invented name)
I would like Sam to playtest my game here at home for about ten minutes. Sam will play and think aloud while I take notes about the game, not about Sam. I will not record Sam’s face or voice; with your OK I would record only the game screen. You are welcome to stay and watch, and Sam can stop at any time. Is that OK with you?
Ask before the session, not in the middle of it. If the parent says no to screen recording, take notes only. In your log write parent OK: yes and the date, never the child’s name.
Structured self-playtest (when nobody is available)
| Step | What to do |
|---|---|
| 1 | Leave the build alone for at least a day. |
| 2 | Start a screen recording, then play as if new: use only what the screen shows you. |
| 3 | Count attempts at each obstacle and say aloud each time you pause. |
| 4 | Watch the recording and fill in the observation sheet below. |
| 5 | Label the session S, not a tester letter. Report it separately and weigh it lower: it finds broken things and unclear text, but cannot show what a fresh player understands. |
Observation sheet
| Time | Where (level or spot) | Observed (what the tester did) | Said (in quotation marks) | Problem underneath |
|---|---|---|---|---|
Invented session log (SLS original, illustrative; Tester B is an adult relative)
| Time | Observed | Said | Problem underneath |
|---|---|---|---|
| 0:40 | Tester B walked past the key twice | Where is the door? | Key is not noticeable |
| 2:10 | Fell into the second pit four times; paused 6 s before the blue platform | I did not know I could wall-jump | Wall-jump never taught |
| 4:30 | Asked the facilitator what to do; facilitator replied What do you think you should do? | Add a double jump | The gap feels too wide |
| 6:00 | Finished the level with 3 pickups missed | It was great, really | Unknown: kind words, but 3 missed pickups; probe with a follow-up question |
Notice the row at 4:30. The tester offered a solution (double jump). The note records the problem underneath (the gap feels too wide) and leaves the fix to the designer. Also notice that the last comment, it was great, really, is weaker evidence than the three missed pickups, which are behavior. Family testers are often generous with praise, so behavior counts for even more.
Sort these notes: observation or opinion?
| Note | Type |
|---|---|
| Tester C fell into the same pit three times, then jumped earlier | Observation |
| The level is too hard | Opinion (rewrite as a count of deaths) |
| Tester A said the music was bad | Said, recorded as a quotation and weighed as opinion |
| Tester D never used the side path | Observation |
| In session S, I paused 4 s at the fork | Observation, from a self-test: weighed lower |
Facilitator lines
| Tester says | Weak reply (leads) | Better reply |
|---|---|---|
| What do I do now? | Just jump on the red box. | What do you think you should do? (Note that they were stuck.) |
| Is this right? | Yes, you are doing great. | What are you thinking right now? |
| I don’t want to hurt your feelings… | Oh, it’s only a prototype. | Honest is the most helpful thing you can be. What happened there? |