TOP Secrets To Build A Billion Dollar Business ft. Zepto's Founder Kaivalya Vohra | FO33 Raj Shamani
with Kaivalya Vohra
Executive Summary
Kaivalya Vohra, 19-year-old co-founder of Zepto, tells Raj Shamani how two bachelors in lockdown Mumbai turned a broken grocery run into what he calls India's fastest-growing e-commerce company of its time: start from a problem you are living, find product-market fit by cutting retention data until a signal appears (for Zepto it was delivery time), then scale with relentless focus. He contrasts easy-to-scale marketplaces with tight unit economics against Zepto's capital-intensive but proven model, explains why the best opportunities are the least sexy businesses, and lays out his two-layer objective: maximize shareholder value professionally while building a generational organization personally. All claims and views are the speakers' own.
Key Takeaways
The best businesses are the least sexy ones
Vohra's filter: software for cattle trading in rural India, construction supply (Infra.Market), contract software for lawyers - boring markets with one-crore downloads and insane growth. Unsexy deters a hundred copycats, and Zepto's purple brand only looks sexy from outside; the operational pain inside is the moat.
Start from a problem you are living
No entrepreneurship eureka: two bachelors in first-wave-lockdown Mumbai, online groceries taking five to six days, local stores under stress, habits spoiled by Instacart in the US. The mentality was 'this sucks, let's figure something out' - the company and the ambition came later.
The 10-minute model came from one data cut
Delivering from local stores in 45-60 minutes produced mediocre retention and NPS. In the YC product-market-fit mentality they cut the data every way - basket, price point - and the signal appeared when they segmented by delivery time: customers who lived closer retained far better on every metric, so they built the infrastructure to make that speed deliberate.
Fast growth plus proven unit economics is the valuation formula
Vohra's own math for the then-$600M valuation: revenue doubling every month, a low revenue multiple he calls public-market-like despite hypergrowth, and mature micro-markets already profitable on every operating metric within five to six months of launch.
Every model trades scale-ease against unit economics
His comparison: ride-sharing and food delivery are light and fast to scale but stay tight on unit economics; Zepto's vertically integrated model is capital-intensive and complex to scale but its unit economics were provable quickly. Neither is wrong - know which side of the trade you are on.
Ambition without a number
No unicorn or decacorn target: the professional objective is maximizing shareholder value; the personal one is looking back in 10, 20, 30 years proud of a solid, enduring, generational organization - the RK Damani and DMart test.
Relentless focus beats option-chasing
Asked about white-labeling or the next big product line, Vohra's 'boring answer' is the strategy itself: keep scaling the machine - more geographies, fresher products, more consistent delivery - and do not lose sight of the ball in a very large market.
The Conversation Breakdown
Two bachelors, a lockdown and a broken grocery run
⏱ Untimed topicThe origin story: first pandemic wave in Mumbai, online grocery slots taking five to six days, the fastest option three days, and two unprepared bachelors who had been spoiled by US products like Instacart. There was no 'this is the moment we become entrepreneurs' moment - they were bored of bad TV and annoyed that groceries were painful, so they started delivering from local stores.
The data cut that invented 10-minute delivery
⏱ Untimed topicScaling 45-60 minute delivery from local stores felt flat: app problems, unexciting retention and NPS. Following YC's product-market-fit drills, they sliced the data by basket and price without luck, then segmented by delivery time and found that customers who serendipitously lived close to delivery points - getting orders in 15-30 minutes - had dramatically better retention, NPS and frequency. That signal became the 10-minute model.
The $600M valuation logic: growth plus proven unit economics
⏱ Untimed topicVohra calls Zepto one of the fastest-growing companies in Indian history and the fastest-growing e-commerce company of its time, with revenue doubling monthly. His case for feeling undervalued: a low revenue multiple for that growth rate, and a clear picture of profitability with mature micro-markets already profitable on every operating metric - all attributed as his own claims.
Scale-ease versus unit economics: every model's trade-off
⏱ Untimed topicA clean framework: ride-sharing and food delivery onboard supply quickly and scale light, but end-state unit economics stay tight; Zepto's model needed far more infrastructure and capital, yet proved full unit economics in five to six months across multiple micro markets. The pro of this model is clarity of economics; the con is the capital needed to scale it.
Building at 19: early projects, the Snapchat book and college as a Bay Area excuse
⏱ Untimed topicBoth founders built things young - a carpooling project among them - and Aadit was working on ideas in high school. Vohra read the Snapchat book about Evan Spiegel building at Stanford and applied to college largely as an excuse to enter the Silicon Valley ecosystem; he says the bulk of company-building learning happened on the job, not in class.
Unsexy markets, and what he would do if not Zepto
⏱ Untimed topicHis opportunity filter: the least sexy businesses are the best - cattle-trading software with a crore of downloads, construction supply, contract software - because boring markets attract no copycats and hidden operational pain becomes a moat. If not quick commerce he says he would probably build in EdTech, definitely not crypto, and he remains a fan of 'atoms' businesses that touch the real world.
Focus as strategy: no numbers, no side quests
⏱ Untimed topicPressed for the endgame number, Vohra refuses: professionally, maximize shareholder value; personally, build a generational organization to be proud of decades later, the way RK Damani might look back at DMart. The near-term plan is deliberately boring - scale the machine into more geographies and keep improving freshness, assortment and delivery consistency - which Raj reads back as the episode's theme: relentless focus.
Core Frameworks & Ideas
Delivery-time segmentation: the PMF signal hunt
When aggregate retention looks mediocre, keep cutting the data by new dimensions - basket, price point, and finally delivery time. Zepto's 10-minute model was discovered, not invented: a segment getting faster deliveries by accident outperformed on every metric, so they rebuilt the infrastructure to make that accident the product.
The unsexy-business filter
If a business is sexy, something is wrong with it. Boring markets - cattle trading, construction, legal contracts - come with less competition, because few founders want the pain; and a consumer brand that looks glamorous but is operationally brutal to run keeps imitators out. Glamour attracts competition; pain builds moats.
The scale-versus-economics trade-off map
Locate your model on two axes: how light it is to scale versus how strong its unit economics are. Asset-light marketplaces scale fast but stay tight on margins; vertically integrated models like quick commerce need capital and complexity but can prove clear unit economics quickly. Choose consciously rather than discovering your trade-off late.
The two-layer objective
Separate the professional objective from the personal one. Professional: maximize shareholder value - the answer a founder is supposed to give. Personal: build a large, enduring, generational organization you will be proud of in 20 or 30 years. Keeping both layers explicit prevents vanity targets like a valuation number from becoming the goal.
Actionable Next Steps
- ✓Take one problem you personally hit this week and write down who else faces it and what they do about it today; that is the seed of a real business.
- ✓Cut your retention or satisfaction data by one dimension you have not tried - speed, price, cohort - and look for a segment that outperforms for accidental reasons.
- ✓Run the sexy test on your next idea: if it sounds glamorous at a dinner party, ask what operational pain will keep a hundred copycats out.
- ✓Write your objective in two layers: the professional metric you are maximizing, and the personal 'proud of it in 20 years' test it must pass.
- ✓Name one shiny side quest you will say no to this quarter so the machine you are already scaling gets your full focus.
Notable Quotes
“what are the least sexy businesses? Those are always the best businesses.”
“I usually think that if a business is sexy there's something wrong with it.”
“you can figure out if this is a problem if you're facing it yourself”
“their retention rates were much better. Their NPS was much better. Their frequency of usage was much better.”
“fast growing revenue plus clear profitability equals like an undervalued business”
“we're trying to maximize shareholder value. That's the right answer.”
“I left nine to five and now I work 24/7”
The One Thing to Remember
Find your signal in the data, then refuse to be distracted: Zepto's 10-minute model came from one retention cut, and everything after that has been focus.