Every founder who tries to build a platform in India eventually hits the same wall: the two-sided marketplace cold start problem. Your app is worthless to buyers until it has sellers, and equally worthless to sellers until it has buyers. Nobody wants to be the first person standing in an empty room. This is the chicken-and-egg problem, and it has quietly killed more Indian marketplaces than bad code or thin funding ever have. The encouraging news is that it is solvable, and the founders who solved it, from Zomato and Swiggy to Urban Company and Meesho, did not do it with a big-bang national launch. They did it by sequencing one side before the other, concentrating liquidity in a tiny geography, and doing a great deal of manual work that did not scale. This playbook lays out that framework for an Indian operator.
What the cold start problem actually is
A marketplace is not really selling a product. It is selling liquidity: the near-certainty that a buyer who shows up will find what they want, and a seller who lists will get a transaction. Below a threshold of density, that certainty does not exist, both sides churn, and the marketplace collapses back to zero. The founder's first job is not growth. It is manufacturing enough concentrated liquidity in one narrow slice of the market that the network becomes self-sustaining, and only then expanding.
The second thing to internalise is that the two sides are almost never equally difficult. In most marketplaces there is a hard side, the side that is scarce, expensive to acquire, and slow to trust you. For ride-hailing and home services it is supply. For a B2B wholesale platform it can be either, depending on who has the leverage. Identify your hard side before you write a marketing plan, because everything downstream depends on it.
Deciding which side to seed first
The default answer, and it is a good default, is to seed the hard side first, and for most consumer marketplaces that means supply. Buyers arrive for selection: if the listings are there, demand follows. Both Uber and Ola leaned on driver incentives to seed supply and build network liquidity in a new city before demand could be reliably served. Ola, which started as a cab aggregator in 2010, three years before Uber entered India in 2013, poured driver incentives in to build the supply base, then tapered them as each city matured. When the two later squeezed those incentives and raised commissions, driver earnings and supply took a visible hit, which tells you how load-bearing that subsidy was in the early liquidity engine.
But supply-first is a default, not a law. The exception shows up most often in B2B, where sellers have margin to defend and view your platform as a new channel rather than a threat to their existing business. There, you can lead with demand, de-risk the buyer, and let the sellers' appetite for that demand pull them in. Udaan, founded in 2016 by former Flipkart operators Sujeet Kumar, Amod Malviya and Vaibhav Gupta, aimed at the roughly nine in ten Indian retail sales that flow through small mom-and-pop stores. The scarce, valuable side there was the kirana buyer with a genuine sourcing problem, so aggregating that retailer demand gave Udaan the leverage to bring manufacturers and wholesalers onto the network. Ask a simple question: which side, if I could conjure it from thin air, would make the other side desperate to join? Seed that side.
Building liquidity city by city, not nationwide
The single most common Indian failure mode is launching in fifteen cities at once and achieving critical mass in none. The counter-move is to concentrate to the point of near-absurdity. Swiggy launched in August 2014 with a small team of delivery riders and a modest set of partner restaurants, confined at first to the Koramangala neighbourhood of Bengaluru, chosen precisely because it packed restaurants and young professionals into a small radius. That density let Swiggy promise delivery well inside the sixty-to-ninety-minute norm of the time, and it made every rider's route efficient. Swiggy built its growth on delivery density, orders per square kilometre, rather than raw geographic spread.
Zomato tells the same story from the supply side. It began in 2008 as FoodieBay, and the founding work was unglamorous: manually listing restaurants and digitising their menus across Delhi-NCR, and saturating that home market before expanding to other Indian cities from 2011. Delhi was worked hard before the next city was touched. The lesson is that a marketplace has a smallest self-sustaining network, sometimes a single pincode, sometimes one professional category in one city. Get that atomic unit to real liquidity, prove the unit economics, then copy-paste the motion into the next micro-market. Density before geography, always.
Doing the unscalable work to seed supply
Early marketplace liquidity is manufactured by hand, not by product. Zomato's founders digitised restaurant menu cards that the restaurants themselves had no interest in putting online, because in 2008 the concept barely existed in India. That is a concierge approach: the founding team performs the marketplace's core function manually to create the first matches and generate the data an algorithm will later need.
Meesho, founded in 2015 by Vidit Aatrey and Sanjeev Barnwal, is the sharpest Indian example of meeting supply where it already lived. The founders noticed small merchants were already selling through WhatsApp, so instead of forcing them onto a new surface, Meesho built its reseller model around WhatsApp, Facebook and Instagram, letting people, overwhelmingly homemakers, resell products to their own social circles with no inventory and no upfront cost. By 2017 it had roughly a thousand suppliers and twenty thousand resellers, and it later scaled to millions of resellers, more than ten million of them women. The lever was removing every ounce of risk from the reseller, who carried no inventory, paid nothing upfront, and earned only when a sale actually closed. Urban Company, founded in 2014 in Gurugram, did the unscalable work on quality: it hand-picked and vetted service professionals, standardised pricing, trained them, and guaranteed the outcome, rather than just listing whoever signed up.
Constraining supply and owning quality
In India the deepest moat is not selection, it is trust. Buyers have been burned by unreliable service, fake listings and no recourse, so a marketplace that guarantees quality can win even with less supply. This is why Urban Company deliberately constrained and curated its supply instead of maximising it, taking ownership of vetting, pricing, training and payment so that both sides could trust the transaction. That full-stack control is expensive, but it is what let the company build a durable consumer brand: by its September 2025 IPO it reported serving more than fourteen million consumers and listed at a strong premium to its issue price of one hundred and three rupees per share.
Concretely, the quality mechanisms an Indian marketplace should design from day one include background and credential verification of the hard side, two-way ratings and reviews with real consequences for low scores, escrow or platform-held payments so neither party can be cheated, standardised pricing to remove haggling and information asymmetry, and clear dispute resolution with refunds. Under-supplying deliberately, keeping demand slightly ahead of supply, also keeps your best sellers busy and motivated, which is a feature, not a bug, in the seeding phase.
The India-specific wrinkles
Several realities make the Indian cold start different from a Silicon Valley template. First, the market is deeply tiered: Ola's early push into Tier 2 and Tier 3 towns, where Uber arrived later, built loyalty that a metro-only strategy would have missed, so plan your city sequence beyond the top eight. Second, language matters: Meesho's vernacular-first, regional-language approach was central to onboarding first-time internet sellers. Third, capital is a supply constraint of its own. Small Indian sellers and kirana buyers are often blocked by working capital, which is why Udaan pairs its marketplace with working-capital credit through its lending arm, Udaan Capital, and runs its own logistics network to control reliability. Fourth, India now has genuine digital public infrastructure you can build on: UPI removed much of the payments friction that once throttled every transaction, and the government-backed Open Network for Digital Commerce, incorporated at the end of 2021 and piloted from 2022, had by late 2024 onboarded several lakh sellers and was processing well over a crore transactions a month across hundreds of cities. ONDC will not solve your cold start for you, but it can lower the cost of acquiring the seller side, and it is worth evaluating rather than ignoring.
Metrics that prove liquidity is real
Gross merchandise value is a vanity number in the seeding phase, easily inflated by subsidy. The metrics that actually tell you whether you have crossed the cold start threshold are liquidity metrics: the share of listings or search sessions that end in a transaction, the median time to match a buyer with a seller, the fill rate on the demand side, and, above all, organic repeat rate once incentives are removed. If retention holds when you stop paying for it, you have real liquidity. If it evaporates, you have a subsidy, not a marketplace, and expanding to the next city will only multiply the leak.
A sequenced cold start playbook
- Name your hard side. Decide which side is scarce, expensive and slow to trust you, and make acquiring it the whole plan.
- Pick your atomic network. One neighbourhood, one city, or one category where the smallest self-sustaining network can exist. Resist launching wide.
- Seed that side by hand. Digitise, vet, train, and concierge-match manually. Buy or subsidise the hard side if you must, as Uber and Ola did with drivers.
- Give the hard side standalone value. Tools, earnings, or distribution that are useful even before the network is dense, so supply sticks while demand builds.
- Reach real liquidity in the unit, measured by match rate and post-incentive repeat rate, before touching the next micro-market.
- Engineer trust in. Verification, escrowed payments, standardised pricing, ratings and refunds, because in India quality certainty beats raw selection.
- Copy the motion outward, city by city, tier by tier, adapting for language, credit access and local density rather than assuming the metro playbook travels.
The cold start problem never disappears entirely: every new city, category and country reopens it in miniature. But it stops being existential once you accept the core discipline. Concentrate before you scale, seed the hard side first, do the work that does not scale, and treat liquidity, not headline growth, as the product you are actually building.