In moderated usability testing, a researcher is present during the session, in person or on a video call, and can ask questions as the participant plays. In unmoderated testing, the participant plays alone, usually at home, and the session is recorded for the team to study later.

Which one to use depends on the question. Moderated sessions suit questions about what a player is thinking, and unmoderated sessions suit questions about how a player behaves when nobody is watching. Below, the two are compared on probing, natural play behavior, cost, and turnaround.

Probing

Probing means asking follow-up questions in the moment: "What were you expecting there?" or "What made you go back to the map?"

Moderated. When a participant hesitates at the crafting screen, the moderator can ask what they are looking for while they are still hesitating. The answer is usually more specific than anything the participant could recall afterward. A moderator can also follow something unexpected, such as a player inventing a strategy the team never designed for.

Unmoderated. There is nobody to ask. Questions come before the session or after it, in a written form or a recorded prompt. By the time the participant answers, they are remembering the moment, and people tend to remember the most intense part of an experience and the way it ended better than the rest. Some unmoderated setups add data recorded during play, such as facial expressions or concentration, so the team can see where reactions happened even without asking.

Choose moderated when the question is about reasoning: why players choose one path, how they understand a system, what they think a screen is for.

Natural play behavior

Moderated. Participants who know someone is watching tend to play differently. They are often more patient with confusing screens, less likely to quit, more likely to read instructions, and more polite about what they dislike. Think-aloud, which moderators often ask for, also slows players down, because talking and playing a fast game compete for attention. That matters little on a turn-based game or a shop screen, and a lot more in an action or rhythm game, where it can change the result.

Unmoderated. The participant plays on their own phone, in their own room, at a time they choose. This is closer to how they would play the game after downloading it from the store, and if they get bored and want to put it down, they do. So unmoderated sessions show more reliably where players would quit.

Choose unmoderated when the question is about behavior: where new players stop, how long they keep playing, whether they find a feature on their own.

Cost

Moderated. Each session needs a researcher for its full length, plus scheduling, and often a note-taker. Sessions happen one at a time, so a study of 10 participants means 10 separate blocks of researcher time. In-person sessions add lab space, devices, and travel.

Unmoderated. Participants play in parallel, without a researcher present. The researcher's time goes into setup and analysis. Unmoderated studies can still take substantial analysis time, because someone has to watch the recordings, but the sessions themselves are cheaper to run.

Choose unmoderated when budget is tight, or when you need more participants than a moderator could reasonably sit with.

Turnaround

Moderated. Scheduling sets the pace. Finding times that suit both participants and the researcher, across time zones, can stretch the sessions out over a week or more before analysis starts. On the other hand, the researcher has sat in on every session and starts analysis already knowing the main themes.

Unmoderated. All participants can play within a day or two of the build going out. The time goes into reviewing recordings afterward. Tools that tag key moments or flag where engagement dropped shorten that step.

Choose unmoderated when you need results before a sprint ends or a release date arrives.

Side by side

Which one, for which question

Some teams use both in sequence: an unmoderated study to find where players struggle, then a smaller moderated study that asks a few players why. Our Beginner's Guide to FTUE Playtesting has a section on choosing between the two for a first study.

Sample size does not change much

Moderation changes the kind of evidence you get, more than the number of participants you need. Five participants per round, per player group, is the usual starting point for a qualitative usability study either way. Unmoderated studies often recruit one or two extra people to cover drop-outs. How many playtesters do you need? covers the numbers.

How we run studies at Emhance

Our studies are unmoderated. Participants play at home on their own device while their webcam records their face, and we analyze facial expressions and eye movement second by second. That gives a record of emotional intensity and concentration across the whole session. We count a moment as engaged when both are high together, so the team can see where engagement rose and fell without anyone in the room. We pair it with short interviews before and after play, and use the debrief to ask participants about the moments the data flagged. See how Emhance playtesting works, or try the blink signal behind concentration in our 30-second demo game.

For the wider context, see game user research: what it is and what teams use it for and remote playtesting without a lab.

Book a demo to see an unmoderated Emhance study on your own game.