Running Player Usability Tests

I recruited ten testers against the personas we’d already built for the game.

The sessions ran on PlaytestCloud, which records the screen and the player’s commentary, sends back a transcript, and shows where on the screen they tapped.

Each session was followed by a survey, capped at ten questions by the platform. I used them on the things watching wouldn’t tell me: what players thought the score meant, what they thought the bubbles were for, whether the UI made sense, and what they made of the art.

I built a profile for each tester first, so their answers came with some context: age, device, country, how much they played and what they usually played.

Then I read through the survey results and watched the recordings back, timestamping anything the survey hadn’t picked up. Each session ran fifteen minutes.

I coded all of it into one sheet by category: usability, mechanics comprehension, playability, bugs, and anything they volunteered as good or bad.

The main finding was that the numbers on screen weren’t reaching anyone. Ten testers gave four different answers for what the score represented, four for what the multiplier did, and three for what the coloured segments on the progress bar meant. That readout is the game’s main feedback on how a run is going.

The confusion went wider than the numbers. Testers split four ways on what they were trying to achieve in a run, and nobody could say why they were collecting bubbles or how bubbles affected their score.

For each issue I wrote up what players said, what I’d watched them do, and what I thought would fix it: animating the intro scene to show the objective, rewording the lose screen, moving the score next to the multiplier icon, and separating the distance indicator from the score comparison on the progress bar.

Not all of it was negative. The art style rated 8.6 out of 10 and testers praised it without being prompted. I also had them rate how fun the game was, whether they’d play it again and whether they’d recommend it.

Separately from the comprehension problems, I logged the playability issues by how often they came up. The two most frequent were deaths the player couldn’t avoid and target scores they couldn’t reach. I had screenshots for one of them: the player moves left, and there isn’t time to reach the right-hand platform before the train arrives.

All of it went into a report with the proposed fixes, which I presented to the whole company, C-level included. Several of the fixes were built. We ran the same process at every stage after that, through the rest of development and up to launch.