Personalising Vocabulary in 4 Minutes a Day
Flipi turns rigid drills into playful micro-lessons, lifting Day-7 retention to 21 %
Problem
Existing vocab apps feel rigid and irrelevant. Learners cannot choose the words they need, lose motivation quickly, and rarely form a habit.
Outcome
Two weeks after launch our twelve-person beta mastered 517 words. Analytics showed 21 % Day-7 retention and average sessions of 3 min 40 s, proving that short micro sessions beat longer drills.
Role
Product Designer & iOS Developer (solo)
Team
Independent project with continuous feedback from 12 classmates, 2 German tutors, and 1 iOS mentor
Timeline
Oct 2024 → Ongoing
Project Intro
The idea was born in my A2 evening class. Every week I copied new words into spreadsheets because the mainstream apps forced me through generic word lists. I wondered: could an app hand the steering wheel to the learner and still feel effortless?
Spark & Framing
I jammed about everything that hurt about current tools and every hunch on how to fix it. A whiteboard session mapped pains, opportunities, and one North-Star goal: “Master twenty self-selected words per week in under four minutes a day.”
Immersion Research
I wanted to understand what already slows learners down. Ten coffee chats with classmates about Duolingo, Quizlet, spreadsheets and other learning methods produced an affinity wall of five themes: Control, Cadence, Progress, Friction, Context. These guided every later decision.
Prototype ➜ Pivot
I began with a deliberately skeletal card to verify the core recall loop. After one week, interviews showed it felt “too mechanical—no clue how to pronounce or use the word.”
A quick redesign added only the context learners said they missed.
Bare-bones | Context-plus | |
|---|---|---|
UI elements | German word · “Show answer” | + progress bar · emoji feedback · audio icon · word-type tag · example sentence |
Learner reaction | “Empty, not helpful.” | “Fun, clear, I can say and place the word.” |
Impact | 11 % Day-7 retention | 21 % Day-7 retention · +11 % quiz accuracy |
The richer card deepened learning without lengthening sessions—proof that a pinch of context can outperform strict minimalism.
Expanding the Practice Arsenal
When the richer flashcard felt solid, testers asked for fresh ways to practise. I added two new modes, and the spaced-repetition engine now serves every word in all three modes until the app marks it learned.
Mode | Core mechanic | What it adds |
|---|---|---|
Multiple Choice | Pick the correct translation from four options | Quick confidence check, instant feedback |
Combine Letters | Drag scrambled letters into the right order | Playful challenge, stronger spelling recall |
With I Knew / I Didn’t Know, Multiple Choice, and Combine Letters rotating in sequence, Flipi began to feel like a mini-game. Daily sessions climbed from 1.2 to 1.7 per learner, words practised per session rose from 14 to 18, and total daily time in app increased from 4 min 15 s to 6 min 02 s, all without stretching the single-session format users already loved.
Data-Driven Refinement
Google Analytics funnels showed the steepest drop-off right after sign-up—users weren’t creating their first deck. I folded topic selection directly into onboarding: new learners now pick 1–3 interests (e.g., Travel, Shopping, Morning Routine) and land with a pre-seeded deck ready to practise.
Metric | Before | After topic-onboarding |
|---|---|---|
Install → First practice | 33.3 % | 51.4 % |
Day-1 activation | 68 % | 83 % |
The change turned an empty home screen into an instant “Aha!” moment and became the single biggest lift in my retention dashboard.
Human Proof
I did not schedule weekly meetings or set up fancy surveys. Instead, I told testers to ping me the moment something felt great or went wrong. A voice note, a screenshot, or a quick text on WhatsApp was enough. If the issue needed more than a couple of messages we opened a fifteen-minute Zoom and worked it out together. Those bite-sized conversations kept the feedback loop fast and honest.
My Contribution
I ran every phase myself—from digging up user pains and sketching the North-Star, to prototyping in Figma, coding the SwiftUI app, wiring Google Analytics, and shipping weekly TestFlight builds. Owning the full loop taught me two things: good research saves weeks of re-work, and tight data hooks turn late-night hunches into clear next steps. By speaking directly with testers, pushing code the same day, and watching dashboards light up, I learned how to blend product thinking, design craft, and engineering speed into one continuous rhythm.
Lessons Learned
Building Flipi solo showed that designing for a pain I share—while chatting directly with the people feeling it, keeps empathy sharp and priorities obvious. A quick WhatsApp ping or micro-Zoom turned unknowns into the next clear build, proving that real voices streamline the roadmap better than long research sprints. Tiny, playful tweaks like the scramble game or a streak flame lifted retention more than heavyweight features, and the biggest wins came from listening to users and doubling down on what mattered most to them.