STEM Unicorn student engagement and revenue recovery designing a gamified e-learning platform across 4 user portals
Role
Duration
Team
Platform
Key Metric 1
Site Admin, School Admin, Teacher Admin, Student one connected system
Key Metric 2
Reversed after replacing physical-only classes with an online gamified platform
Key Metric 3
Design Thinking process Empathize, Define, Ideate, Prototype, Test
Summary
Overview
STEM Unicorn had a problem that snuck up on them they were running a STEM and entrepreneurship education programme entirely through physical classes, and students were checking out. Engagement dropped. Revenue followed. I joined a team to design their full online platform from scratch: four separate portals for Site Admins, School Admins, Teachers, and Students, built around gamification, decision-based learning, and the kind of progress visibility that actually keeps students coming back. My individual scope covered user research, user flows, wireframes, design system, high-fidelity prototype, and the full student experience.
Students weren't disengaged from learning they were disengaged from how STEM Unicorn was teaching it.
Problem Framing
The Core Problem
Three things were failing at once. Physical classes were boring students into disengagement. There was no online platform, so students who couldn't attend in person had no alternative. And even the students who showed up lacked motivation to push through the harder material there was nothing pulling them forward between sessions.
Why it mattered
STEM Unicorn's revenue dropped measurably over the year before this project started. Their own research team traced it back directly to the educational approach. That's a hard finding to sit with the product itself was the problem. Patching the marketing wouldn't fix it. The learning experience needed to change fundamentally, and the platform needed to exist before any of that could happen.
Supporting data / evidence
Research identified three consistent user complaints lack of motivation to study, difficulty and boredom in physical classes, and no interest in existing course-related games and quizzes. The target demographic ranged from students aged 18-25 through to teachers and school administrators up to 55, spanning middle to upper-class economies with English as the primary platform language. Every persona had a different relationship with the platform, but all of them were pointing at the same root failure: learning felt disconnected from anything real or engaging.
Role & Team
My Role
The Define and Ideate stages were collaborative I worked with Harsana Kuwar, James Kuwar, and Nikolaos to map pain points, build the How-Might-We board, and run the Crazy 8 ideation sessions. After that, I worked independently. My solo deliverables were user research, user flows for all four portals, wireframes, design system, and the high-fidelity prototype.
Team Composition
Four-person team for research and ideation. Solo execution from prototyping through final UI. The team collaboration happened exactly where it needed to at the stage where four perspectives on a complex multi-portal system genuinely improved the outcome.
Stakeholders & decision-makers
STEM Unicorn's product and education leadership. Platform specification requirements — including the gamification mechanics, jackpot question algorithms, booster shops, and multi-language support were defined by the client and refined through the design process.
Approach
Strategy & framework
We used Design Thinking across five stages Empathize, Define, Ideate, Prototype, Test. The collaborative stages front-loaded the complex decisions: four portals with overlapping permissions and distinct user needs required careful definition before anyone opened a design file. The affinity diagram and Crazy 8 sessions in Ideate were where the gamification mechanics started taking shape as actual interface patterns, not just feature list items.
Research methods used
User research and empathy mapping across four distinct user types Site Admins managing licences and content, School Admins managing classrooms and student rosters, Teachers managing individual classes and progress, and Students navigating courses, games, and quizzes. Pain point mapping and How-Might-We brainstorming in the Define stage. Card sorting and affinity diagramming in Ideate. User flow creation across all four portals before wireframing. Usability testing post-prototype with iterative improvements.
Key insight(s)
The permission architecture was more important than the UI. Four portals sounds manageable until you start mapping what each role can and can't do Site Admins upload content school by school, School Admins can pause classrooms but can't remove them, Teachers can pause individual students, and Students only see the classroom they were invited to. Getting that permission logic right in the user flows before designing a single screen prevented a category of errors that would have been nearly impossible to fix in high-fidelity.
Challenges
Primary challenge
Designing a gamified learning platform where the game mechanics had to be genuinely educational not just motivational decoration. The jackpot question algorithm, investment simulation, booster shop, and leaderboard system all needed to reinforce STEM and entrepreneurship concepts, not distract from them. Gamification that feels like a reward layer on top of boring content doesn't fix boring content. It just delays the drop-off.
How I navigated it
I designed the game mechanics as learning flows, not reward flows. The investment simulation where each investment represents 10–20% of current valuation with auto-generated percentage returns is a real financial decision-making exercise wrapped in a game interface. The booster shop exists within the game's economic logic, not outside it. The leaderboard shows top 10 publicly but keeps individual rankings private for lower-ranked students, which protects motivation for the students who need it most.
What I Did differently
The multi-language support content and interface localisation by country and school was scoped into the platform but deserved its own dedicated design sprint. Designing a system that adapts content school by school, with language varying by location, creates a content management complexity that the Site Admin portal carries most of the weight for. I'd push for a separate IA session specifically for the content localisation logic before the next phase of development.
Solution
What I designed
The multi-language support content and interface localisation by country and school was scoped into the platform but deserved its own dedicated design sprint. Designing a system that adapts content school by school, with language varying by location, creates a content management complexity that the Site Admin portal carries most of the weight for. I'd push for a separate IA session specifically for the content localisation logic before the next phase of development.
Key design decisions
1. Separated licence states for School Admins. Active and expired school licences are visually separated in the Site Admin view not mixed in a flat list. This sounds minor but it's the kind of thing that turns a 30-second task into a 5-minute hunt when you're managing dozens of schools. The expired licence data remains accessible even after expiry, which matters for continuity when schools renew. 2. Leaderboard visibility logic protecting motivation. Only the top 10 students appear on the public leaderboard. Students ranked below 10 see their own position privately. This was a deliberate decision based on the research finding that lack of motivation was a primary drop-off cause a fully public leaderboard where a struggling student can see exactly how far behind they are doesn't motivate, it discourages. The private ranking keeps every student in the game. 3. CSV roster upload instead of email-by-email invitations. School Admins and Teachers can upload a CSV of student names and emails to register an entire class at once, rather than sending individual invitations and waiting for acceptance. In a school environment managing hundreds of students, this isn't a convenience feature it's the difference between the platform being usable in practice or not.
Results & Impact
Quantitative results
Four fully designed and user-flow-validated portals delivered across a five-stage Design Thinking process. Complete gamification system designed decision games, jackpot questions, investment simulations, booster shops, and leaderboards within a unified educational framework. Wireframes and high-fidelity prototype completed and tested with iterative improvements applied.
Qualitative outcomes
The platform addresses all three user complaints identified in research lack of motivation (tackled through gamification and streaks), physical class limitations (removed entirely by the online platform), and disengagement from existing quizzes (replaced with investment simulations and jackpot mechanics that have genuine unpredictability). The permission architecture, validated through the user flow stage, handled the four-portal complexity without exposing users to actions outside their role.
Business impact
STEM Unicorn now has an online platform where none existed before which was the single biggest structural cause of their revenue decline. Students who couldn't attend physical classes have a full alternative. Students who were disengaged have a reason to come back. The gamification layer creates the daily return behaviour that subscription-based EdTech platforms depend on, and the School Admin and Teacher portals give institutions the management visibility they need to justify platform adoption at a school level.
Next Steps
What's next (and why)
The Metaverse integration flagged as "coming soon" in the original platform brief is the natural next phase. The visual novel and in-class experience layer that STEM Unicorn is building toward requires its own UX work, particularly around how the existing Student Portal transitions into an immersive environment without losing the progress and game data students have already built up.
Open questions / hypotheses
Does the leaderboard visibility logic actually improve retention for lower-ranked students or do they find other ways to disengage? That's testable within the first semester of platform usage and would directly inform whether the private ranking approach is worth keeping or needs a different design. And how does the investment simulation perform as an actual learning tool are students making better financial decisions by the end of a course than at the start?
Learnings
The thing I keep thinking about from this project is how much the permission architecture shaped everything downstream. Four portals, four sets of permissions, four different relationships with the same underlying data getting that logic solid in the user flows before touching visual design was the right call, and I'd make it again on any multi-role platform. The UI is the easy part. The hard part is making sure the right person can do the right thing and nothing else. Nail that first, and the rest follows.