Onboarding First-Time Players

I pulled apart the onboarding in six games: Talking Tom Gold Run, Talking Tom Hero, Blades of Brim, Bowling Crew, Clash Royale and Tennis Clash, recording what each one taught, when, and how. Every one of them had a way to stop a new player failing a lesson, by pausing until the right input arrived, slowing time until it did, or letting the crash happen and rewinding it.

Spacing was the one I took most from. Clash Royale teaches across five practice matches, adding one thing at a time and never moving on until the last one’s been used. Tennis Clash does the same with upgrades and serving. That’s the rule I designed to.

I started with a grid of every feature the game would eventually ask a player to understand: when to teach it, how, and a priority from 0 to 3, where 0 was critical and 3 meant it was fine if the player never worked it out. I wrote the why in the player’s own voice, as “I need to collect bubbles to fill my multiplier”, so every lesson had to justify itself as something the player wanted. A narrative wrapper column turned the list back into a story about Sackboy trying to become the best runner of the lot.

Then I sequenced it in FigJam across eight runs, through rounds of feedback with the Lead Designer, before another UX designer and I built it into the full user journey flow. Runs 1 and 2 are the forced tutorial, and after that the forcing tapers: by Run 4 the player can equip an item, check missions, buy the season pass, open a lootbag, or just run again.

The coach’s tapping finger only appears two seconds after the screen it sits on, and the navigation arrows only appear if the player hasn’t already read the obstacle themselves. Crashing during the tutorial doesn’t end the run: Sackboy comes back through a portal at the same spot, and free continues carry on until the Marathon unlocks.

I wireframed every sequence in Adobe XD and wrote each one up as a step-by-step spec covering what appears, after how long, what’s forced, and where the player lands next. That went to Art, Design, Code and QA, and came back with changes every round.

We built it into a clickable prototype so people could walk the sequence instead of reading it, since a tutorial is mostly timing. Art produced the UI assets and their own supporting document, down to the states of the username field, where DONE stays greyed out until the name is between three and twelve characters. I stayed with Art and Code through implementation until what was on screen matched what I’d specified.

I defined the analytics events alongside the analysts and tagged them by priority, then we watched the funnel through soft launch and past hard launch. Between February and July 2023, 617,021 players went through it.

99% started, 91% pressed the run button, 82% finished their first run and won it. 74% started Run 2, 65% reached the missions and marathon sequence, 60% started entering a name, 54% finished Run 3. By Run 5 it was 41%, by Run 6 36%. The decline was steady the whole way down, with no single step that players fell out of.

We also split the same events by first hour, first day and third day. The day-three line sits about three points above the first-hour line early on and eight or nine by the end: 15% of players had won a Daily Duel within an hour, 22% within three days. Most of those missing from the first-hour funnel came back and finished later.

The platform split surprised me. Android brought 455,300 players to iOS’s 161,721, nearly three to one, but iOS completed more of the funnel at almost every step and finished on 31% against Android’s 22%.

One step did drop sharply. 34% of players saw the coach unlock the Daily Duel and 25% started one, a nine-point fall where the steps either side cost one or two. It’s the last thing the onboarding forces, and the first that asks a player to compete against a stranger. If I did it again I’d let the sequence fire once the player had gone into the Daily Duel on their own.