Redesigning SudoStudy; an AI-powered Ed-tech platform, to help students study smarter
- Role:
- Senior Product Designer
- Timeframe:
- 6 months (then ongoing product consulting)
- Business goals:
- Lift WAU, increase sessions per user and decrease first-session drop off
Business impact:
+66.7%
WAU
(from 1,280 to 2,187)
+27%
Sessions per user (practice mode usage)
-31%
In first-session drop-off
None answer the only question he cares about:
what should I study next?
He clicks a few menus.
Closes the tab.
Plans to “study later.”
I struggle to stay consistent with my schedule…sometimes I don’t even have one...it’s hard to figure out which topics I should focus on
-Pugh, 16 years old, O Levels
I get overwhelmed by all the topics and menus…I often skip sessions...
I don’t know where to start next.
-Cole, 17 years old, A Levels
I am a bit lenient when self-assessing, I prefer external feedback to gauge progress…delay in feedback makes I lose interest
-Grace, 16 years old, O Levels
make the next step obvious and rewarding - fast.
What I walked into
- No onboarding. New users hit a combined auth modal: sign up, sign in, magic link, password, everything at once. It sometimes failed. Rage-clicks were common. People bounced.
- Practice mode was cluttered. Competing CTAs (“Give feedback,” “Buy subscription”). Redundant choices. No clear exit. Multi-select topics that didn’t match how students actually study.
- Home didn’t guide the next session. Students couldn’t see strengths, gaps, or coverage. They felt lost between subjects and topics.
- Desktop first, but mobile behavior mattered. The UI didn’t respect quick, focused sessions.
The bet:
If we reduce choice, surface the next best action, and fix first-run friction, students will activate faster and come back more often.
My role and how we worked
I led end-to-end design across research, IA, UI, prototyping, and developer handoff. I joined sprint planning, wrote user stories with acceptance criteria, and partnered tightly with the founder, PMs, and a small dev team. We shipped in two-week increments, with clear “measure next” hooks in each release. Marketing later used the new visuals in campaigns and on the website.
Evidence that shaped the plan
Session replays & funnels (PostHog):
- Confusing auth modal → immediate drop-offs after sign-up/sign-in
- Unclear navigation in practice and quizzes
- “Too much, too soon” options causing paralysis
- UI elements (buttons, instructional cues) needed to be more explicit
Interviews (20 students):
- Students prefer one topic per session.
- They want external feedback and quick wins.
- They need a sense of progress (how am I doing?) and coverage (what have I touched?).
These insights drove three product bets:
Fix first-run friction - authentication and onboarding
Problems
New users landed in a catch-all modal that mixed sign up, sign in, magic link, and password flows. It was unreliable and unclear. Many bounced before seeing value.
Decisions
- Separated authentication paths and hardened error states.
- Introduced a short onboarding that sets intent and momentum: pick subject(s), set a starting point, get a clear “Start practicing” CTA.
- Wrote crisp, action-first copy.
Trade-offs
We kept onboarding short to ship fast and measure impact. Deeper personalization moved later.
Results
Immediate drop-offs fell 31%. More students reached their first real practice session.
Make “what’s next” obvious - the Home reframe
Problems
Home didn’t guide the next action. Students couldn’t see progress or gaps.
Decisions
- Added Health Score (how often you answer correctly) and Coverage Score (how much of a topic you’ve touched).
- Surfaced Recommended Topics using those two signals.
- Added a daily streak to encourage frequent, shorter sessions.
Trade-offs
We deferred advanced recommendation tuning to focus on activation and return usage first.
Results
Students had a clear next step, and weekly users rose from 1,280 to 2,187.
Reduce cognitive load - Practice mode, before & after
Problems
- Competing CTAs distracted from the task.
- No obvious exit from a session.
- Multi-select topics didn’t match real study behavior.
- Redundant options added friction before the first question.
Decisions
- Added a clear ‘Exit practice’ control to restore a sense of safety.
- Put difficulty and topic choice first, in a simple left-to-right flow.
- Switched topic selection to single-select (radio) based on interviews: one topic per session.
- Used progressive disclosure. Secondary actions live behind an options menu.
- Kept the question area clean: answer → feedback → next.
Trade-offs
Single-select limits “batching,” but it aligns with how students actually work and reduces choice paralysis. We bookmarked multi-topic practice as a future advanced mode.
Results
Students spent less time deciding and more time practicing, and practice per user rose 27%.
What didn’t work (and what I learned)
The initial subject empty state CTA under-performed until I swapped a descriptive card for a clear primary button. Simple beats clever when users are anxious to start.
Outcomes and measurement
+66.7%
WAU
(from 1,280 to 2,187)
+27%
Sessions per user (practice mode usage)
-31%
In first-session
drop-off
How I measured
- Post-release window: ~2–4 weeks, all active users.
- Baseline: the prior 2–4 weeks.
- No major promos or unrelated product changes during this period.
How I work
- Translate user signal into product bets with clear guardrails.
- Ship in small, measurable slices; instrument the next answer into each release.
- Keep copy, layout, and controls brutally simple, especially for first-run flows.
- Treat design as a team sport: I co-planned sprints and wrote user stories with acceptance criteria; handoffs were detailed and predictable.
🤔 Reflections
This project reminded me that great UX isn’t just about cleaner screens, it’s about building systems that support users end-to-end.
Grounding every decision in research and real behavior helped turn scattered features into a more focused, usable learning experience.
Even small friction points can pile up, but thoughtful design rooted in user needs makes a measurable difference.
