Swiggy Founder on the Winning Pitch Deck, Early Days & Logistics Moat

ELI5/TLDR

In 2014, three Pilani grads built a 12-page pitch deck in two or three days — all function, no design — and used it to raise Swiggy’s seed round. Eleven years later, founder Sriharsha Majety and one of his first backers, Anand Daniel of Accel, sit down with the original slides and check the math. Most of it held: the 15% restaurant commission, the four-orders-a-month repeat rate, the bet that owning the delivery fleet was the whole game. The big thing they got wrong was being too conservative — the investors literally told them to add a zero to their order projections.

The Full Story

The deck nobody agonized over

The first surprise is how unfussy the whole thing was. The deck took two or three days, came together only because they’d finally landed a meeting, and had no template behind it.

“We didn’t have any frames of reference for our deck per se… but from our friends who had raised — few of our friends from Pilani had raised money by then — they gave us the basics of the deck structure.”

That basic structure is the entire pitch-deck syllabus, and it has not changed: what’s the problem, how are you solving it, why is your solution different, what are the unit economics, what’s next. Majety would barely touch it today. The slides were ugly on purpose — “almost all function and no form,” made before any design tools existed.

The order of the slides didn’t matter much to him either. The conventional wisdom is to lead with the team, but Majety shrugged: they had no special claim to this business, no “street cred,” so why pretend the founders were the headline. They led with the problem instead.

Logistics first, food second

The origin story runs backwards from how you’d expect. They didn’t start by deciding food ordering was broken. They started by wanting to be in logistics — they’d been running an e-commerce aggregation logistics business called Bundl — and on-demand mobility was exploding around them. Uber and Ola were making cars appear at the push of a button. The leap was simple:

“If you can push a button and make cars move on the road super fast, you must be able to push a button and make things move super fast on the road.”

The consumer insight that fell out of that: every existing platform — Just Eat, Foodpanda, TastyKhana, Tiny Owl — called itself a tech platform connecting restaurants and consumers. But that’s not what people actually wanted.

“Consumers asked for my favorite restaurants delivered fast.”

The good restaurants weren’t on those platforms because those platforms refused to own delivery. They’d hand the order to the restaurant and hope. The result, in Daniel’s memory, was that ordering food could mean waiting “anywhere from 45 minutes to 3 hours.” The pitch slides named the fix directly: build a Domino’s layer over every restaurant. Domino’s was the only place in India where ordering already worked frictionlessly, because Domino’s owned its own delivery. Swiggy’s whole bet was to rent that capability to restaurants that didn’t have it.

That bet — owning the logistics rather than just brokering the order — is what the title calls the moat. It was also the thing almost nobody believed could be made profitable.

The unit economics that mostly came true

The most striking part of the conversation is watching them grade an eleven-year-old spreadsheet. The 15% restaurant commission they projected (rising to 20%) is, Majety says, still close to right today. The delivery charge held. The repeat rate is the real headline:

“Our repeats are to this day close to four times a month after 100 million users have tried out the service.”

The early adopters in 2015 ordered about four times a month. After scaling to a hundred million users, the average is still about four times a month — but for a different reason. The core users turned out to order more than budgeted, while a long tail of low-frequency users dragged the average back down. Two errors cancelling out to the same number.

Cost per delivery is where reality beat the forecast. They originally delivered within a 3-4 km radius (there’s a fond anecdote about users nudging their map pin to unlock restaurants just out of range). Today they deliver up to 18 km in big cities. Hold the distance constant, and the cost per delivery is far better than they’d ever projected, even after years of inflation. The operation got more efficient than they dared assume.

The one number they got wrong, they got wrong by being timid. When Accel’s Anand Daniel and others reviewed the model:

“They said, ‘Boss, I think you’re just under-billing, why don’t you just put one zero extra,’ and I was like, okay, that seems wild, but okay, who knows.”

The founders thought maybe 10,000 orders a day. Daniel underwrote 100,000. The real number is in the millions.

The slides that didn’t age well

Two future-facing bits missed. One was a slide listing phone calls as a future ordering channel — Majety’s co-founder Nandan wasn’t keen on it even then, and it “thankfully” never happened; Swiggy never took an order by phone. The other irony: a 15-minute delivery slide that sat dormant for a decade until they relaunched the idea as Bolt in 2024. They’d sketched the future correctly and then waited ten years for the world to catch up.

There’s a smaller engineering story that captures the “consumer-back” instinct Daniel kept praising. In 2015 neither customers nor delivery riders had smartphones — riders had feature phones. So Swiggy gave each rider four phone numbers and used missed calls as a free signalling system: a missed call to number one meant order confirmed, number two meant reached restaurant, and so on. A jury-rigged live-tracking system built from the only technology available. And when too many orders came in for too few riders, a yellow banner appeared saying they couldn’t take orders — rather than accept the order and fail.

“It’s easy to say I’ll take the order and then figure it out. But they said no, let’s be transparent with the consumer.”

What an investor actually buys

Daniel’s reflection on his own decision is unusually honest. He’d done a “prepared mind” thesis on food, met more than fifteen companies in the space, and the deck’s clean numbers — 20% week-on-week growth, the retention, the repeats — moved him at the time. But in hindsight he’s almost dismissive of those metrics:

“These are all early numbers… I’ve seen startups have this early on and then not hold later on. We were doing 70 orders a day.”

You cannot underwrite a company on 70 orders a day. What he actually bought was the founders’ clarity of thought and refusal to dilute it. When he pushed them to add more restaurants, add reviews, they said no — Swiggy was about speed, not discovery; the customer already knows what they want. The “I’ll figure out the unit economics, but we have to solve this” conviction was the real signal. The deck was the filter that earned the meeting; the people earned the cheque.

On whether founders still need a deck at all, Majety is relaxed. The deck is just “one form factor” for a single job to be done: explain to someone who’s never heard of your business why it’s exciting, in the least possible time. A Q&A memo, a Notion page, even an AI-generated podcast — any of them works. His one rule: the shorter the better, 10 to 12 pages, because for you it’s your only company and for the investor it’s one of thousands a year.

Key Takeaways

The pitch-deck skeleton is permanent. Problem → how it works → how you’re different → unit economics → what’s next. Five sections, unchanged in a decade. Lead with the problem, not the team, unless the team is genuinely your edge.
Shorter is the whole strategy. 10-12 pages. The deck’s only job is to win the next meeting in the least time. The investor is reading thousands; you are selling one.
The deck wins the meeting, the founders win the cheque. Daniel calls early traction metrics nearly worthless predictively — 70 orders a day proves nothing. What he underwrote was clarity of thought and conviction held under pushback.
Owning the delivery fleet was the actual moat. Every rival positioned as a neutral “tech platform” and outsourced the hard part. Swiggy’s “Domino’s layer over every restaurant” — owning logistics — is what made fast, reliable delivery possible and what made it defensible.
Founders underestimate the long term. The team modelled 10,000 orders a day; the investor pushed them to add a zero; reality blew past both. The conservative founder is the common case, not the hype-merchant.
Solve backwards from the customer, build with the tech you have. Missed-call status updates on feature phones, a yellow “can’t take orders” banner instead of a silent failure. Transparency over the easy lie.
Strong opinions, weakly held. Majety’s phrase for how they made design calls — say no to reviews and extra restaurants to protect speed, but stay data-driven enough to adjust.
An idea can wait a decade. The 15-minute delivery slide sat dormant from 2015 until Bolt launched in 2024. Being early and being wrong look identical until the world catches up.

Claude’s Take

This is a good genre — the founder and the investor grading the original deck together, eleven years on, with the actual slides on screen. It avoids the usual founder-interview problem where everything is reverse-engineered into genius. Majety is refreshingly willing to say large parts were luck, intuition, and beers (“let there be Swiggy… not that much logic after a point”), and Daniel admits the early metrics he found persuasive were, predictively, close to noise.

The most useful idea is the cleanest one: the deck wins the meeting, the people win the cheque. It reframes the whole pitch-deck-industrial-complex away from polish and toward the one job of buying a founder thirty more minutes of an investor’s attention. The runner-up is the quiet point that founders systematically lowball — the “add a zero” anecdote is a nice corrective to the assumption that startups oversell.

What keeps this from an 8 is that it’s a single SeedToScale episode and a touch promotional — it exists partly to drive deck downloads, the survivorship bias is total (we’re hearing the one deck that worked), and the back half drifts into pleasant-but-soft territory about “level five leadership” and the future of quick-commerce. The first twenty minutes, on the deck and the unit economics, are genuinely worth the time. Seven.