The Indian EdTech Opportunity
India has 250 million school-going students. CBSE alone has 26,000+ affiliated schools. Yet the digital learning tools available for these students are, with a few exceptions, underwhelming - PDFs uploaded to a portal and called e-learning.
We've built a virtual science lab platform for CBSE classes 8–12, two mobile learning apps for early childhood, and have been deeply embedded in the Indian EdTech space long enough to know what works and what doesn't.
This is what we'd tell a founder starting an EdTech product for the Indian market today.
The CBSE Curriculum Is Your Architecture Blueprint
The CBSE NCERT curriculum is published publicly on ncert.nic.in. It's the single most important document for any K-12 EdTech product targeting India.
Every piece of content - every video, quiz, simulation, or worksheet - must map to a specific chapter and learning outcome in the NCERT syllabus. Indian students and parents evaluate EdTech products by one question: "Will this help with exams?" If your content can't answer that question with a chapter reference, you will not retain users.
For our virtual lab platform (labs.mindgridlabs.org), we mapped every experiment to the exact NCERT chapter, competency, and learning objective. Teachers could filter labs by class, subject, and chapter. This was the single biggest driver of adoption.
Technical Architecture for EdTech at Scale
Offline-First is Non-Negotiable
Internet connectivity in India is unreliable outside Tier 1 cities. A school in a Tier 3 city may have wifi that drops every 20 minutes. Your app must work - at least for content consumption - without an active internet connection.
For mobile apps, this means:
- Content download for offline use (videos, interactive elements)
- Progress sync when connection is restored
- Graceful degradation (show cached content, queue quizzes for later submission)
We use Room database on Android for offline data persistence, with WorkManager handling background sync. For web, Service Workers with a cache-first strategy.
Video Delivery
Video is the backbone of most EdTech products. Hosting video on YouTube is tempting (free, fast) but creates two problems:
1. YouTube's algorithm will recommend non-educational content to students
2. CBSE schools often block YouTube
We recommend AWS MediaConvert + CloudFront for video delivery. Upload the master video, transcode to multiple bitrates (360p, 480p, 720p), and deliver the appropriate bitrate based on network speed detection. Cost for 1TB of video served per month via CloudFront: approximately $85.
Assessment Engine
For any curriculum-mapped product, the assessment engine is as important as the content. You need:
- MCQ, short answer, and fill-in-the-blank question types (matching NCERT exam patterns)
- Adaptive difficulty (harder questions after correct answers, easier after wrong)
- Timer-based exam mode for competitive exam practice
- Detailed result breakdowns by chapter and topic
- Parent/teacher reports showing progress over time
We build assessment engines on PostgreSQL with a flexible question schema that supports all question types through a polymorphic data model.
Interactive Simulations for Science
For our virtual lab platform, we built chemistry, physics, and biology simulations. The technology decision here:
- Simple 2D simulations (titration, circuit diagrams): Canvas API or SVG, fully custom
- Complex 3D simulations (molecular structure, optics): Three.js
- Physics simulations (projectile motion, pendulum): Matter.js
Avoid Unity WebGL for web-based simulations - load times are too high for the average Indian school's internet connection. Native Canvas + WebAssembly for physics calculations gives comparable results at 10% of the load time.
Content Localization
Hindi is spoken by 500M+ Indians. But India has 22 official languages and hundreds of regional languages. For CBSE specifically, the medium of instruction (English vs Hindi) splits roughly 60/40.
For a national product, plan for at minimum:
- English interface and content
- Hindi interface and content
- Regional language audio dubbing (not full translation - audio dubbing on English content is 80% of the value at 20% of the cost)
For our early childhood apps (Animal Kingdom Hub, KiddLearn Hub), we supported English, Hindi, and Assamese - reflecting our initial target markets.
The Regulatory Layer: COPPA and DPDP
If your app serves children under 13, two frameworks apply:
COPPA (Children's Online Privacy Protection Act): Required if you plan to distribute in the US via the App Store or Play Store. No data collection from under-13s without verified parental consent. No third-party advertising. No social features without moderation.
India's DPDP Act (Digital Personal Data Protection Act, 2023): Requires verifiable parental consent for processing data of children. Prohibits tracking or behavioural advertising targeting children. Platforms must take "reasonable security safeguards" to protect children's data.
In practice for app development:
- Remove all analytics SDKs (Firebase Analytics, Mixpanel) from the kids version
- Remove all advertising SDKs
- Implement a parental gate before any purchase or social feature
- Build a data deletion flow (parental request to delete account and all data)
Monetization Models That Work in India
Freemium with a subscription does not convert as well in India as in Western markets. Price sensitivity is high, especially outside Tier 1 cities.
Models that work:
- One-time purchase (₹299–₹999): Works well for focused tools (a single-subject app, a specific exam prep tool). Low friction, no billing fatigue.
- Annual subscription at low price point (₹999–₹1,999/year): More compelling than monthly at ₹99–₹199, which feels like an ongoing commitment.
- Institutional licensing: Schools and coaching institutes pay per-student or per-institution. B2B revenue in EdTech is more predictable than B2C.
- Government schemes: DIKSHA, PM eVIDYA, and state-level digital education programs often procure EdTech products. Slower sales cycle, but high volume.
What We'd Do Differently
If we were starting an Indian EdTech product today, we'd:
1. Build mobile-first (75%+ of Indian internet usage is mobile)
2. Start with a single class and subject, not a full curriculum
3. Partner with one school for the first 6 months as a design partner
4. Make offline-first a day-one requirement, not a retrofit
5. Launch in English and Hindi simultaneously - not English first, Hindi later