The Liquidity Paradox No One Talks About
Byju's raised $22B. Unacademy raised $2.8B. Both collapsed under the same pressure: teachers left. In ed-tech, the reverse two-sided network kills you. Students flee when teacher quality drops. Teachers flee when student attendance becomes unpredictable. This isn't like Uber, where drivers and riders both want transaction volume. Teachers are constrained suppliers. They have alternative income paths: private tuition, school jobs, corporate training. Students have no alternative when they're in exam prep mode. The asymmetry matters. You must solve teacher liquidity first.
The India Stack Made This Problem Visible
Before Aadhaar and UPI, ed-tech couldn't verify teacher credentials at scale. Payments were friction-heavy. Identity was unverifiable. The India Stack removed those frictions by 2020. Instantly, every startup could claim 50M students. But removing tech friction exposed human friction. Which teachers? Which qualifications? How do you retain them? None of the dashboards and APIs solve this. You need ground operations. You need vetting. You need relationship management. The tech stack amplified demand but did nothing for supply.
Why Teacher Side Matters More Than You Think
Consider Vedantu's trajectory. In 2022, they had 1.2M students. By 2024, they'd halved to 600K. Student churn accelerated because teacher consistency collapsed. Vedantu paid tutors per-session, no fixed commitment. Teachers taught when they needed money. Sessions were inconsistent. Parents noticed. The math is brutal: if 30% of students experience 2+ teacher changes per month, churn jumps above 60%. You can't acquire students fast enough to compensate. A typical SaaS platform needs 8% monthly churn to stay profitable at reasonable CAC. Ed-tech operates at 55-70% annual churn. That requires obsessive focus on teacher quality, not student scale.
The Two-Sided Threshold Nobody Measured
Work backward from profitability. A tier-1 city needs roughly 4,000-8,000 active paid students to support a sustainable tutor ecosystem. That's roughly 200-250 qualified teachers at 20-30 students each. For those teachers to stay, they need predictable income: ₹80K-120K per month. Across 250 teachers, that's ₹2-3 crore per month in teacher cost. At ₹500-1,000 per student monthly pricing, you need that 4,000-8,000 student base. If you fall below 3,000 students, teacher retention collapses. Teachers move to tutoring collectives or corporate training. Your network evaporates. This threshold is not a myth. It's geometry. Every major failure (Vedantu's pivot, Unacademy's layoffs) happened because they crossed this threshold in reverse: high student count, insufficient teacher supply.
The Sequencing Implication
Build teacher supply first, in one metro. Not national. Not "AI-enabled." Just strong, verified, persistent humans. Target 1,200-1,500 teachers in Bangalore or Delhi NCR. Ensure 75%+ retention over 24 months. Then acquire students to liquidity. Once proven at 6,000+ students with stable teachers, expand to metro 2. This takes 3 years, not 18 months. It means rejecting 95% of institutional money that wants hockey-stick unit economics. But it works. It's why Byjus's core cohorts (before the sprawl) had reasonable LTV. They'd built teacher stability first, locally.
What the India Stack Tells Us
Aadhaar and UPI made verification and payment instant. But they also made it obvious what the real friction was. It wasn't identity. It wasn't payments. It was human reliability at scale. You can't UPI your way around that. You need operations. You need people. The Stack removed the wrong friction first, which made founders chase scale on the wrong side.