The situation
An education startup with 4,000 enrolled students and no app to serve them
The business in this scenario — a Hyderabad-based test preparation company focused on competitive government examinations — had grown rapidly through offline coaching centres and a basic website. When they secured funding to expand online, they needed a mobile app that could deliver live classes, manage assignments, process payments, and support thousands of concurrent students. Their first batch for the new digital programme was starting in 4 months.
The founder had spoken to three development agencies before us. Two quoted 8–10 months. One quoted 6 months but required the client to have a dedicated technical co-founder to manage the project. We committed to 16 weeks and delivered in 15.
Core challenges
Four distinct problems compounding each other
Hard deadline — batch starts in 4 months
4,000 students had already enrolled expecting a digital product. Missing the batch start date wasn't a commercial option — it would trigger refunds, reputational damage, and potential legal exposure.
Live video at scale
Live classes with thousands of concurrent viewers require infrastructure most app developers aren't equipped to handle. Low-latency, high-reliability video streaming was non-negotiable for the core learning experience.
Complex payment scenarios
Students paid by instalment, by scholarship tier, and by subject bundle. The payment system needed to handle 6 fee structures, partial payments, and automatic access revocation for payment failures — without a single manual step.
Cross-platform requirement
Students were on Android (78%) and iOS (22%). Building two separate native apps would double development time and budget. A single Flutter codebase was the only viable path to hitting the deadline.
Our approach
A structured 6-phase SEO programme
We built a phased plan that addressed technical foundations first, then content, then authority — in that order. Skipping the technical audit would have wasted the content investment.
1
Architecture and technical design
Week 1 was entirely planning. Firebase for real-time data and auth, Agora for live video, Razorpay for payments, and Flutter for cross-platform UI. API contract designed before a single screen was built — preventing rework later.
FirebaseAgora video SDKRazorpayFlutter
2
Student authentication and onboarding
Student login via phone OTP (no passwords to forget). Course selection flow with scholarship code redemption. Profile completion with photo, subject selection, and exam target. Onboarding completion rate in beta testing: 94%.
Phone OTP authScholarship codesExam target selectionOnboarding completion tracked
3
Live class module
Agora integration for low-latency live video with up to 2,000 concurrent viewers per session. In-class chat, Q&A queue, hand raise, and poll features. Class recording stored to Firebase Storage with automatic availability 2 hours after class ends.
Agora live video2,000 concurrent viewersIn-class Q&AAuto-recorded classes
4
Assignment and progress tracking
Teachers upload assignments with deadline, marks, and file attachments. Students submit via app. Auto-grading for MCQ assignments. Manual grading interface for written work. Progress dashboard showing completion rate, scores, and rank within cohort.
Assignment submissionMCQ auto-gradingProgress dashboardCohort ranking
5
Payment module
Razorpay integration supporting UPI, cards, net banking, and EMI. 6 fee structures configured. Automatic access revocation at T+7 days for payment failures with reminder push notifications at T+1, T+3, T+5. Zero manual payment processing required.
Razorpay integration6 fee structuresAuto access revocationPayment reminders
6
Push notifications and go-live
Firebase Cloud Messaging for class reminders, assignment deadlines, result announcements, and promotional messages. Segmented notifications by subject and batch. Both iOS and Android submitted to app stores in week 14. App approved and live in week 15.
Firebase notificationsSegmented pushiOS App StoreGoogle Play Store
What we would measure
How success would be judged
We agree a baseline in the first week and report against it every month. We don't promise figures before we have seen your data.
Crash-free sessions
Measured from day one with crash reporting, fixed before new features.
Day-1, 7 and 30 retention
Cohort retention, so we know whether students come back, not just install.
Store conversion
Store page views to installs, tested with screenshots and copy.
Rating and reviews
Prompted at the right moment in the app, with every review answered.
