We built an EdTech platform for a coaching institute in Pune that serves 15,000 students. The first version had every feature — live classes, recorded videos, quizzes, doubt resolution, progress tracking. Completion rate? 12%. We stripped it down, added a structured learning path with daily targets, push notifications for streak maintenance, and peer discussion groups. Completion rate jumped to 47%. Features don't drive learning — engagement systems do.
What We'll Cover
Types of EdTech Platforms
| Type | Examples | Key Features | MVP Timeline |
|---|---|---|---|
| Course marketplace | Udemy model — multiple instructors, many courses | Instructor onboarding, course creation tools, reviews, search/discovery | 4-6 months |
| LMS (Learning Management System) | School/corporate training platform | Cohort management, assignments, grading, attendance, reports | 3-5 months |
| Live tutoring | 1-on-1 or small group live sessions | Video calling, whiteboard, scheduling, tutor matching, payment per session | 3-4 months |
| Test prep | JEE/NEET/CAT prep platforms | Question banks, mock tests, analytics, adaptive difficulty, peer comparison | 3-4 months |
| Skill-based learning | Coding bootcamps, design courses | Project-based curriculum, code playgrounds, peer review, portfolio building | 4-6 months |
Core Feature Priorities
Don't build everything. Prioritize based on what actually drives learning outcomes and retention.
| Feature | Priority | Impact on Retention | Build Complexity |
|---|---|---|---|
| Structured learning paths | Must-have (Day 1) | High — students need direction, not a content dump | Low |
| Progress tracking + streaks | Must-have (Day 1) | Very high — daily streaks increase 30-day retention by 2-3x | Low |
| Video lessons (recorded) | Must-have (Day 1) | Medium — content is expected, but content alone doesn't retain | Medium (hosting, streaming, DRM) |
| Quizzes and assessments | Must-have (Sprint 2) | High — active recall is the most effective learning technique | Medium |
| Discussion forums/doubt resolution | Important (Sprint 3) | High — peer learning and quick doubt resolution reduce drop-off | Medium |
| Live classes | Nice-to-have (unless core to model) | Medium — high production cost, low attendance rates (typically 20-30%) | High (WebRTC, scaling, recording) |
| Certificates | Nice-to-have | Low on learning, high on perceived value for professional courses | Low |
Video Streaming: The Expensive Part
Video is the biggest cost center for EdTech platforms. An hour of HD video costs ₹500-2,000/month to serve to 1,000 students, depending on your CDN and encoding choices.
Build vs Buy
| Approach | Cost | Best For |
|---|---|---|
| YouTube (unlisted) | Free | MVP, free courses. No DRM, no analytics, minimal branding control |
| Vimeo OTT / Mux | $0.005-0.02 per min watched | Premium courses. DRM support, analytics, customizable player |
| AWS MediaConvert + CloudFront | $0.002-0.01 per min watched | Scale. Full control over encoding, DRM (Widevine/FairPlay), custom player |
| Bunny.net Stream | $0.005 per min stored + bandwidth | Budget-friendly with decent features. Good for Indian market pricing |
Video DRM: Preventing Piracy
In India, course piracy is a massive problem. Free Telegram groups share recorded courses within hours of release. DRM isn't perfect, but it raises the bar significantly:
- Widevine L1 (Google) + FairPlay (Apple) for browser/mobile encryption
- Watermarking: Invisible per-user watermarks in the video. If leaked, you can trace the source
- Screen recording detection: Pause playback when screen recording is detected (mobile apps)
- Token-based playback: Video URLs expire after a short time. Can't be shared directly
Engagement and Gamification
The biggest problem in EdTech isn't acquisition — it's completion. Average course completion rates are 5-15%. Gamification done right can push this to 30-50%.
What Works
- Daily streaks (Duolingo-style). The fear of breaking a streak is a powerful motivator. We saw 2.5x increase in daily active users
- XP points + leaderboards. Per-cohort leaderboards, not global. Students compete with their batch, not the entire platform
- Milestone celebrations. "You finished Module 3!" with confetti animation. Sounds trivial, feels great
- Spaced repetition. Automatically resurface concepts the student struggled with, 3 days and 7 days later. Science-backed retention boost
What Doesn't Work
- Badges for everything. Too many badges dilute meaning. 5-7 meaningful badges beat 50 trivial ones
- Global leaderboards. Demotivating for beginners. Use cohort-based or self-improvement metrics instead
- Forced gamification. If students must earn points to access content, it feels like a gate, not a reward
AI in EdTech (What's Real vs. Hype)
| AI Application | Readiness | Value |
|---|---|---|
| AI doubt resolution (chatbot) | Ready now (GPT-4, Claude) | High — instant answers at 2 AM when tutors aren't available. Reduces tutor workload by 60% |
| Adaptive learning paths | Ready now | Medium-high — adjust difficulty based on quiz performance. Skip content the student already knows |
| Automated essay/answer grading | Ready for structured answers. Partial for essays | Medium — good for practice quizzes. Still needs human review for high-stakes assessments |
| AI-generated content | Ready for summaries, quizzes. Not for full courses | Medium — generate practice questions from lecture content. Don't generate the course itself |
Monetization Models
| Model | Works When | Typical Pricing (India) |
|---|---|---|
| One-time course purchase | Self-paced, content-heavy courses with clear outcome | ₹500-5,000 per course |
| Monthly/annual subscription | Large catalog, continuous learning (like a library) | ₹200-2,000/month |
| Cohort-based (live) | Instructor-led, time-bound, with community | ₹5,000-50,000 per cohort |
| Freemium | Large user base needed. Free content + premium features/certificates | Free → ₹500-2,000/month for premium |
India-Specific Challenges
- Bandwidth constraints. Tier 2-3 cities have unreliable internet. Support video download for offline viewing. Compress aggressively — 480p is fine for mobile learning
- Vernacular content. English-only platforms miss 70% of India's students. Hindi, Tamil, Telugu, Bengali, Marathi support dramatically expands your addressable market
- Price sensitivity. Indian students compare against free YouTube content. Your platform's value must be clearly above "I could learn this free on YouTube"
- Parent-as-buyer. For K-12 and test prep, the parent pays but the student uses. Marketing and UX need to address both audiences
- Piracy. Course recordings appear on Telegram within days. DRM, watermarking, and community-driven value (live doubt sessions, peer groups) are your defenses
Frequently Asked Questions
How much does it cost to build an EdTech platform?
Basic LMS: ₹15-30 lakh (3-4 months). Platform with live classes + video streaming: ₹30-60 lakh (4-6 months). Full-featured marketplace with AI features: ₹60 lakh-1.2 crore (6-9 months). Ongoing costs: video hosting (₹1-5 lakh/month at scale) and AI API costs.
Should I use an existing LMS like Moodle or build custom?
Moodle (or Teachable/Thinkific) for MVP validation — launch in 2-4 weeks. Build custom when you need unique features that differentiate your platform: custom gamification, AI-driven learning paths, or specific assessment types. Most successful EdTech startups start on existing platforms and migrate to custom as they grow.
How do I prevent course piracy?
DRM (Widevine + FairPlay) for video encryption. Per-user watermarking to trace leaks. Token-based playback URLs that expire. But technology alone isn't enough — build value that can't be pirated: live doubt sessions, community access, personalized feedback, and certificates. The community is harder to replicate than the content.