The Pattern: Violate the Mental Model
In 2010, Nithin Kamath saw a problem that nobody else saw as a problem. Indian brokers charged 0.1% commission on trades. That sounds small until you realize: 90% of retail traders lost money anyway. Adding friction (commission) didn't change the outcome.
So he proposed removing it. VCs asked: "How do you make money?" He didn't have a good answer. That's why they passed.
Razorpay faced the same rejection in 2014. Harshil Mathur and Shashank Kumar wanted to build payment infrastructure for India. But payments were dominated by banks, and India's digital adoption was minimal. Why would developers choose a startup over HDFC Bank?
Every assumption was wrong. That's why it worked.
Why Investors Miss These Ideas
Paul Graham calls this "idea-market fit failure." The idea is good, but it violates what investors believe about the market. They extrapolate from existing data, not emerging trends.
In 2014, digital payments looked niche. Zerodha's commission-free model looked unsustainable. But founders saw something: behavior was changing faster than business models.
Indian investors operate with mental models from the previous decade. A 2010 VC couldn't imagine 1.5 crore Indians trading stocks online. A 2014 VC couldn't imagine merchants accepting digital payments on a phone.
The best founders don't challenge the investor. They simply move forward. Zerodha bootstrapped. Razorpay raised from international investors who understood infrastructure plays.
The Terrible Idea Framework
Scott Belsky's "The Messy Middle" explains why execution beats prediction. The idea phase is about spotting patterns others miss, not about perfect clarity.
Here's how to spot a terrible idea that might work:
1. Does it solve a problem that large players actively ignore?
Zerodha: Brokers ignored retail traders because the unit economics looked bad. But Zerodha saw scale in volume, not margin. Zoho: Enterprise software companies ignored small businesses. But Zoho saw a 10x larger addressable market if you served them well.
2. Does early traction come from an unexpected segment?
Zerodha's first power users were millennials. Razorpay's first adopters were developers building SaaS companies. These weren't the "obvious" users. That's the signal.
When your early adopters surprise you, keep going. That's where investors won't follow—and where you'll build moat.
3. Does the unit economics only work at scale?
This is the trap. Most founders see this and quit. "We need too much volume." But if volume is feasible (and it was in India), then it's not a problem—it's a distribution challenge.
Zerodha needed 10 lakh+ users to break even on trading infrastructure. That was insane in 2010. But the Indian market had 1 crore+ retail traders. The math worked. You just had to be patient.
Why This Matters for Indian Founders
India's market is still mispriced. VCs are optimizing for categories that worked in the US: B2B SaaS, consumer apps, logistics. But India's real opportunities are in:
- Fintech for the 40 crore Indians who lack banking
- Infrastructure for the 10 crore+ businesses without digital tools
- Vertical SaaS for India-specific industries (textiles, agriculture, logistics)
These ideas sound unsexy. They don't have network effects. They can't scale to $1B in 5 years. But they can bootstrap. They can become massive. They can be profitable.
The Actionable Test
Before you kill an idea because "VCs will say no," ask three things:
1. Is there a behavior shift in India that capital hasn't caught? (Zerodha: retail trading moving online. Razorpay: SaaS startups building in India.)
2. Does your early traction come from an unexpected user? (If your first 100 customers surprise everyone, you're onto something.)
3. Can you bootstrap or raise from contrarian investors? (Zerodha did. Razorpay raised from Y Combinator and early-stage VC funds. Zoho bootstrapped for years.)
If you answer yes to all three, stop optimizing for investor conviction. Start optimizing for founder conviction and early user obsession.
VCs are followers, not leaders. The best Indian startup ideas will look wrong to them for 2-3 years. That's the feature, not the bug.