Latest / Marketplace Businesses with Fexingo: Two-Sided Networks, Liquidity, and Take Rates / How Marketplace Subsidies Fix the Cold-Start Problem
Transcript
- Lucas: So here's a number that's been stuck in my head: the first time Uber launched in a new city, they were spending roughly forty percent of gross bookings on driver incentives. That's not a typo. Forty percent. Luna: Forty percent? That's a massive subsidy just to get drivers on the road. But I guess it makes sense — without drivers, there's no ride to sell. Lucas: Exactly. And what's interesting is that this is the cold-start problem in its purest form. You need supply to attract demand, but supply won't show up until there's demand. So someone has to break the loop. Luna: And that someone, historically, is the marketplace itself, writing a check to whichever side is more reluctant. Lucas: Right. Today I want to dig into how marketplaces use subsidies to build liquidity, and how they do it without torching their economics forever. Because there's a real art to knowing which side to subsidize, how much, and for how long. Luna: And we've covered a lot of marketplace angles on this show — take rates, disintermediation, the chicken and egg — but I don't think we've ever really drilled into the subsidy playbook. Good call. Lucas: Let's start with the classic framework. For any marketplace, you've got two sides: supply and demand. In the early days, one side typically needs a nudge more than the other. The side that faces more friction — higher cost, higher risk, more effort — is usually the one you subsidize. Luna: So for ride-hailing, that's the drivers. They're asked to sit in their cars, burn gas, and wait for pings. That's a real cost with no guaranteed income. Lucas: And it's why Uber and Lyft both leaned so heavily on driver guarantees. 'Make at least twenty-five dollars an hour during peak times.' That kind of promise gets drivers on the road, and once riders see short wait times, they come back. Luna: So the subsidy is really buying reliability on the supply side, which then sells to demand. Lucas: Precisely. But here's the thing: subsidies are expensive, and they only work if you can eventually pull back. The goal is to get to a liquidity threshold where the marketplace works without the crutch. And that threshold is usually around a certain density of supply per demand in a given geography. Luna: Is there a number for that? Like, rides available within five minutes, or restaurants with a table open at eight PM? Lucas: It varies by category, but the principle is: demand-side users need a near-certainty that they'll get what they want quickly. For ride-hailing, it's about a five-minute pickup time. For food delivery, it's about a thirty-minute delivery window. Once you hit that reliability, you can start cutting subsidies and raise take rates. Luna: But what about the side that seems like it would be more willing to show up? Like, riders are pretty eager to get a ride. So why do some marketplaces subsidize demand instead? Lucas: Good question. Sometimes the friction is on the demand side — especially when the product is new or unfamiliar. Take OpenTable, back in the late nineties. They had to convince restaurants to install these clunky terminals and pay a subscription, but diners weren't exactly clamoring to book online yet. Luna: So they subsidized the restaurants? Or the diners? Lucas: Actually, they did both, but the clever part was that they guaranteed restaurants they'd fill seats. OpenTable would buy unsold inventory and resell it at a discount, effectively subsidizing demand by guaranteeing the restaurant a minimum number of covers. Luna: So they were buying demand, essentially. And that gave restaurants the confidence to sign up. Lucas: And once the network got dense enough, they could drop the guarantee and move to a per-cover fee. That's the take-rate eventually. But without that early guarantee, they might never have gotten the restaurants on board. Luna: So the key is to subsidize the side that has the most friction, and the friction isn't always on the supply side. Lucas: The right framework is, which side is more price-sensitive or more risk-averse? In ride-hailing, drivers bear the cost of waiting. In restaurant reservations, the restaurant bears the risk of empty tables. So you subsidize that risk away. Luna: That makes sense. But what about the delivery apps? They seem to subsidize both sides heavily — free delivery, discounts on meals, and still pay drivers. How does that ever turn profitable? Lucas: That's the billion-dollar question. And the answer is, it takes a long time, and some never make it. But the ones that do follow a pattern: subsidize heavily to build habit, then slowly reduce incentives as the habit becomes ingrained. The take rate climbs, and the subsidy as a percentage of gross bookings falls. Luna: But there's a risk of subsidy addiction, right? If you train users to only order when there's a promo code, they won't pay full price. Lucas: That's the trap. If you cut subsidies too early, demand collapses. If you keep them too long, you're just buying revenue. The trick is to use subsidies to get to a density where the network effects kick in — where the value of the marketplace itself is the draw. Luna: And network effects are the real endgame. Once you have the largest driver pool, drivers get more rides, riders get shorter wait times, and that's a moat no one can subsidize away. Lucas: Exactly. But getting there requires a subsidy strategy that's disciplined and data-driven. You have to know your unit economics cold, and you have to be willing to walk away from certain geographies or categories where you can't reach that density. Luna: Before we go deeper into the mechanics, I want to pause for a second. If this conversation is helping you think about your own marketplace or your investment thesis, that's exactly why we make this show. And the way we keep it ad-free is listener support. If today was useful, you can buy me a coffee at buy me a coffee dot com slash fexingo. It's a simple way to keep these episodes coming. Lucas: Yeah, and it genuinely makes a difference. Every bit helps us cover the research and the guests. So if you're getting value, that's the way to keep it going. Now, back to the strategy. Luna: So we've talked about subsidizing supply and demand. But what about the timing? When do you pull the trigger on cutting subsidies? Lucas: The conventional wisdom is to cut subsidies once you hit that liquidity threshold in a given market. But there's a nuance: you don't cut across the board. You cut where the network is dense, and you keep subsidies where it's still thin. Luna: So it's hyperlocal. In downtown San Francisco, you can cut driver incentives, but in the suburbs, you might need to keep them. Lucas: That's exactly what ride-hailing companies did. They rolled back guarantees in dense urban cores first, and kept them in less dense areas until they hit critical mass there. It's a surgical approach. Luna: And how does that affect take rate? Because the endgame is to raise take rate, right? Lucas: Right. Once subsidies drop, the cost side of the ledger improves, and you can start charging more — either to consumers or to suppliers. But you have to be careful not to raise take rate so much that you create an incentive to disintermediate. Luna: And that's a topic we've covered, but it's worth repeating — if you squeeze too hard, your best suppliers will take their business elsewhere, and then you've lost your liquidity. Lucas: So the ideal path is: subsidize to build liquidity, then gradually raise take rate as the marketplace becomes more valuable to both sides. The value you've created through network effects justifies the fee. Luna: Is there a case where a marketplace never needs to subsidize at all? Like, if the product is so compelling that both sides show up naturally? Lucas: Rare, but possible. Think of a marketplace for a scarce, high-demand item — say, collectible sneakers. The demand is already there, and the supply is inherently limited. StockX didn't need to subsidize either side; they just needed to provide authentication and trust. But that's an exception. Luna: So for most marketplaces, subsidies are just the price of entry. The real skill is knowing when to stop. Lucas: And that's the takeaway. Subsidies aren't a sign of weakness; they're a strategic investment. But they're only worth it if you have a clear path to network effects and a plan to unwind them. Luna: So for anyone building a marketplace today, what's the first step? How do you decide which side to subsidize? Lucas: Start by mapping the friction. Talk to potential users on both sides and ask, 'What's the biggest risk or cost for you?' The side that says 'I'm worried I'll waste my time' or 'I'm worried I'll lose money' is your subsidy target. Then design a subsidy that removes that specific fear. Luna: And measure everything. You need to know your cost to acquire each side, your activation rate, and your long-term value per user, so you know what you can afford to subsidize. Lucas: Right. Because the worst outcome isn't spending too much on subsidies — it's spending too much without ever reaching escape velocity. That's how marketplaces die. Luna: It's a fine line. But when it works, you get something like a ride-hailing network where the value is so obvious that you can eventually take a healthy cut. Lucas: And that's the ultimate goal: a marketplace where both sides are there for the network, not the incentive. When you hit that, you've solved the cold-start problem for good.