A Japanese learning app that puts writing first, for learners who can say the words but can’t write them.
System Usability Score, above the industry benchmark
came back within 48 hours of their first session
would recommend it to other learners
Most Japanese apps teach you to recognise and say words. Writing comes later, if at all. Kakimasu flips that: characters are the foundation, practised stroke by stroke, so learners build literacy alongside vocabulary instead of after it.
I designed it end to end over three months, from the first interviews to a tested prototype, aiming to turn character practice from a chore people avoid into something that builds confidence.
Duolingo, Busuu and the rest have made vocabulary feel like a game. Writing, though, is treated as a side skill. Early conversations with learners kept surfacing the same frustration: months into an app, they could order food but couldn’t write the simplest characters.
Before designing anything, I wanted to know whether this was a real gap or just my own experience.
Beginner to intermediate Japanese learners.
How the big players handle writing, if at all.
What learners complain about, unprompted.
Writing confidence and practice habits.
struggled to write basic characters after 6+ months on popular apps
gave up within 3 months because they felt “illiterate”
avoided writing practice because it felt overwhelming
interviewees called writing their “biggest weakness”
How might I make writing Japanese characters approachable for beginners, so literacy grows alongside vocabulary instead of waiting until later?
Paper prototypes with stroke-by-stroke guidance, tested with 8 learners. The big insight: people needed to understand why stroke order matters, not just follow it.
I mapped the journey from first meeting a character to mastering it, and landed on four core moments:
Confirmed the writing-first approach and showed pronunciation needed to be built in.
Tuned how strictly strokes are recognised and added cultural context.
Polished the timing of feedback and the light game elements.
Three screens from the tested prototype: pick a course, pick a lesson, then write.
Characters never show up alone. They arrive inside real words you can use straight away.
Each character is broken into numbered strokes with arrows. Drift off course and a gentle haptic nudges you back onto the path.
Every stroke sequence plays its native pronunciation, so sight, movement and sound reinforce each other.
No wall of completion percentages. Progress shows up as small, specific wins: “3 new strokes mastered today” or “perfect stroke order on your first try”.
Usability testing with 23 people, plus how they used it afterwards.
called the stroke-order system “game-changing” for how they learn
used the core features without any onboarding
System Usability Score, above the industry benchmark
average session, versus 5–6 minutes in competing apps
returned within 48 hours of their first session
sessions per week on average
“For the first time, I understand why stroke order matters. It helps me remember the characters.”Interview participant
“I actually look forward to writing practice now. The app makes me feel like I’m improving, not failing.”Beta tester
Treating writing as the foundation, not an advanced skill, set Kakimasu apart from vocabulary-first apps and met a need they ignore.
Even a 2–3 second delay noticeably hurt learning. Real-time stroke correction became a must-have.
Characters learned inside real words or cultural context were retained about 40% better than isolated drills.
The best predictor of people coming back wasn’t game elements but how confident they felt, so I started measuring confidence, not completion.
Going after a specific gap made for a more compelling product than polishing what already exists. Learners knew exactly what they were missing; the apps just hadn’t asked.
Designing around the limits of handwriting recognition taught me to bring developers in early. Some of the best solutions came from working with those constraints instead of against them.
Most of all, it showed me that confidence is a UX outcome you can measure. Tracking how people felt about their progress, not just what they completed, is what made them want to keep going.