Case Studies/LMS
Product Design Case Study · EdTech · Multi-role LMS
LMS — architecting concept-based learning across three roles
A concept-based learning platform for private schools, architected grade by grade and subject by subject for students, teachers and admins, and delivered on mobile and web in seven months.
- My role
- Product Designer, end to end
- Domain
- EdTech · K–12 private schools
- Duration
- 7 months, zero to production
- Platform
- Mobile app & web
01
Overview
A learning platform built around concepts, not memorisation
This is a concept-based Learning Management System for private-sector schools. Instead of testing students on recall, it breaks every subject into its underlying concepts and asks concept-oriented questions designed to build real understanding, mapped to each school’s curriculum, grade by grade.
The platform serves three roles (students working through curriculum and concept quizzes, teachers managing classes and tracking performance, and admins overseeing the school) and ships as both a mobile app and a full desktop web application.
My role: I was the product designer and owned the full arc, not just screens: product design, application architecture and testing, end to end. That included architecting how a curriculum is represented in software: every grade, subject and concept, and how a student’s progress is measured, marked and reported.
02
Challenge
Students needed to understand a lesson’s roots, not just its answer
The brief was direct: provide concept-based learning that helps a student understand a lesson and the roots behind it, not just get the surface-level answer right. Schools had curriculum content and question banks, but no structured way to teach or test the concept a lesson was built on.
Standard multiple-choice testing rewarded memorisation over understanding, gave students no visible sense of progress, and gave teachers no early signal of who was falling behind on a specific concept, only on a whole subject.
“Students don’t disengage because the content is hard. They disengage because they can’t see any progress. The motivation layer isn’t a nice-to-have; it’s load-bearing.”Research insight
03
Research
One platform, three very different jobs
Before any wireframe, structured research ran across all three roles: 15 students across grade levels, 10 teachers across subjects, and school admin stakeholders, to understand how each group experiences learning, assessment and school management.
Students · 15 personas
Want to know exactly where they stand. Motivated by streaks, scores and visible progress. Need quizzes that feel like a challenge, not a test.
Teachers · 10 personas
Need to release quizzes and chapters on schedule, compare sections at a glance, and see who is struggling before it’s too late.
School admins
Need institution-wide visibility, teacher-training tracking, syllabus and calendar control, and audit-ready reports.
04
Approach
Herbert Simon’s seven stages, because the real work was defining a concept, not a screen
The core challenge wasn’t visual design. It was defining what a “concept” is inside a curriculum, and whether a good way to teach and test it already existed. Herbert Simon’s seven-stage process fit directly: define the thing first, and only converge on components once they’re meaningful and buildable.
- Define concept learningWhat a “concept” means for a given grade and subject, and what mastering it should look like.
- Study the marketWhether concept-oriented teaching and testing already existed for that subject, and how.
- Explore presentationHow each concept should reach a student: a quiz type, a concept tree or a guided exercise.
- Prototype in FigmaThe concept and its interaction pattern, for both mobile and web.
- Choose componentsBased on the concept itself and what was realistic to build.
- Build into the systemChosen components added to the shared design system, ready for engineering.
- Close the loopWatch whether students actually learned, and feed that into the next concept.
05
Architecture
Three flow architectures, one application
I defined the application architecture end to end, as three flow architectures that work together.
The core data model behind every role in the platform.
Student functional flow
How a student moves from grade and subject selection, into a concept, through a quiz, to a visible result on their mark sheet.
Concepts architecture
How each subject’s concept tree is structured: the concepts a student must master, and how they build on one another.
Admin & teacher review flow
How teachers release chapters and quizzes and review performance, and how admins approve and monitor it all.
Marks & course architecture
How a quiz score rolls up into a chapter grade and a subject report, and how schools configure course programs without breaking the model.
06
Design decisions
The interactions that made the difference
Concept-based quiz types
Drag-and-drop matching, fill-in-the-blank and image-based questions, built to test understanding rather than recognition.
Motivation layer
Streaks, animated score reveals, grade badges and leaderboards built into the student experience as structure, not add-ons.
Question navigation dots
A persistent, colour-coded question map so students always know where they are and can jump freely.
Teacher analytics at a glance
Chapters vs completed, quizzes released vs taken, and section comparisons in one dashboard view.
Same architecture, every screen
Once the web app’s concept trees, examinations and mark sheets were defined, I rebuilt the same structure for the mobile app, the format most students and teachers reach for first.
07
Testing
Checking for missed paths, not just bugs
Testing focused on whether the happy path completed end to end for every role: a student finishing a quiz without a dead end, a teacher releasing a chapter without missing a step.
Flow-completion testing
Every role’s happy path checked end to end for dead ends and missing steps.
Concept testing
Validated that each question genuinely tested understanding of its concept.
Feasibility testing
Checked each concept and flow was realistic to build and maintain.
A/B testing
Compared competing presentations of the same concept to see which one students learned from.
08
Outcomes
From zero to a production LMS on two platforms
Mobile app and web delivered
A complete mobile app shipped first, with the same architecture scaled into a full desktop web app.
Concept-based learning, not just MCQs
Drag-and-drop, matching and image-based questions replaced flat multiple choice across every subject.
Built-in motivation system
Streaks, score animations, grade badges and leaderboards woven into the core student experience.
One consistent data model
Grade, subject, concept tree, marking system and course programs unified into one model.
09 · Conclusion
What architecting a curriculum taught me
The hardest part was never the screens. It was defining a “concept” precisely enough that a marking system, a concept tree and a quiz type could all agree on it. Rebuilding the architecture for mobile made the whole system more disciplined, and the biggest lesson stuck: motivation design isn’t cosmetic. It’s why a student comes back tomorrow.