Latest / Marketplace Businesses with Fexingo: Two-Sided Networks, Liquidity, and Take Rates / How Marketplace Subsidies Evolve After the Cold Start
Transcript
- Lucas: So a couple of weeks back, DoorDash reported another quarter where they spent more on incentives than most investors wanted to see. And the stock took a hit for it. But here's the thing — that's not a bug in the marketplace model. That's the model working exactly how it was designed. Luna: Right, because we always talk about the cold-start problem — how marketplaces burn cash early to get both sides on board. But what happens after that? The subsidies don't just disappear. Lucas: Exactly. And that's what we're digging into today — the lifecycle of marketplace subsidies. How platforms use them to boot up, then how they try to wean themselves off without breaking the liquidity they worked so hard to build. It's the messy middle that nobody puts on the pitch deck. Luna: So let's start with the beginning. The classic play is Uber — they basically bought both sides of the market in the early 2010s. Lucas: Yeah, Uber's early strategy was almost comically aggressive. They were charging riders less than the cost of the ride, and paying drivers more than the fare was bringing in. Every single trip lost money. But what they were buying was density. In a city like San Francisco, you needed a driver around the corner when someone opened the app, otherwise the rider would just call a cab. Luna: And the driver side — they needed to see that there was enough demand so they didn't sit idle. It's a chicken and egg problem, and money was the egg. Lucas: Exactly. And it worked. By 2014, Uber was in dozens of cities and had effectively trained a whole generation of riders to expect a car in five minutes. The subsidy was the tuition fee for that education. Luna: But then the tuition bill comes due. So how do you start pulling back without the whole thing collapsing? Lucas: That's the million-dollar question. Uber started by slowly cutting driver incentives — they could do it because there was now enough demand that drivers didn't need the bonus to stay on the road. And on the rider side, they introduced surge pricing and started raising the base fare. The key was doing it incrementally, so you never hit a cliff where suddenly both sides feel the platform is worse than the alternative. Luna: But not every marketplace has the luxury of that kind of demand stability. Think about Airbnb — they don't have surge pricing in the same way. The liquidity is much more seasonal. Lucas: Right, Airbnb's subsidy was more about the supply side in the early days. They sent professional photographers to hosts' homes for free, which made listings look way better and built trust. That was a one-time cost, not a per-transaction subsidy, but it was still a subsidy — it lowered the barrier for hosts to list. Luna: And it worked because homes that got photographed got booked more, so hosts saw the value. But they eventually stopped that program. Lucas: Yeah, they did — and it was fine, because by then the network effect had kicked in. Travelers expect a huge selection of homes in any city, and hosts know that listing on Airbnb is the only way to reach that global demand. The subsidy had done its job. Luna: So the ideal is to use subsidies to reach a tipping point, then pull them back. But some marketplaces never really reach that point. Or they do, but the market dynamics force them to keep subsidizing. Lucas: And that's where DoorDash comes back in. Food delivery is a high-frequency, low-margin business. The liquidity is real, but it's also incredibly competitive. If DoorDash stops offering discounts, a user just opens Uber Eats. And if they stop paying drivers enough, the drivers switch to a different app. So the subsidy is not just for the cold start — it's a permanent feature of the market. Luna: Which is why investors get nervous. They see the subsidy line in the financials and think it's a sign of weakness, but it's actually a structural cost of doing business in that category. Lucas: Right. And that's a really important distinction. There are two types of marketplace subsidies: the ones you can eventually turn off, like Uber's early driver guarantees, and the ones that are effectively permanent, like DoorDash's constant promotions. Knowing which type you're dealing with is crucial for modeling the business. Luna: So let's get more tactical. How does a marketplace operator actually decide when to cut a subsidy? Lucas: The honest answer is, it's a constant experiment. You look at metrics like the percentage of gross merchandise value that comes from subsidized transactions. And you look at repeat rates. If a rider who used a promo code comes back and orders at full price within sixty days, that subsidy was an acquisition cost. If they never come back, you just lit money on fire. Luna: And there's a whole science to it now. I remember reading about how some marketplaces use holdout groups — they randomly assign some users to a no-discount control group and watch how their behavior differs. Lucas: Exactly. Incrementality testing. You're not asking 'do discount users order more than no-discount users?' You're asking 'would this specific user have ordered anyway, even without the discount?' That's the hard question, and a lot of marketplaces get it wrong because they just look at the average. Luna: Which is why you see some platforms panic when they pull a promotion and retention drops. They overreact and re-instate the promo, making the subsidy permanent when it shouldn't be. Lucas: Yeah, and that's a classic mistake. But there's also the opposite — cutting too fast and too deep, and destroying the liquidity that made the marketplace valuable in the first place. That's what happened to some of the early ride-hailing competitors in Southeast Asia, where they pulled back too quickly and lost the market to Grab. Luna: So it's a delicate dance. And it's not just about the consumer — you have to think about the supply side too. Cutting what you pay to the supply side too fast can be even more damaging. Lucas: Totally. On the supply side, the metric is often the earnings per hour for a driver or a delivery worker. If that dips below what they can make elsewhere, they churn, and then service quality drops, and then demand drops — it's a death spiral. So you have to manage both sides in tandem. Luna: And that's why the take rate is such a tricky lever. Raising it is the obvious way to improve unit economics, but if you raise it faster than liquidity allows, you hurt the network. Lucas: Yeah, we've talked about take rate a lot on this show, but it's worth connecting the dots here. When a marketplace raises its take rate, it's effectively pulling back a subsidy — either from the consumer by raising prices or from the supply side by paying them less. The question is whether the marketplace's value proposition can absorb that change. Luna: Which brings us to the idea of the 'take rate ceiling'. There's a point where you can't raise it anymore without breaking liquidity. And the ceiling is different for every marketplace depending on how strong the network effects are. Lucas: Right. Let's use a concrete example. A marketplace like StockX — which we've covered before — authenticates sneakers and other collectibles. They can charge a higher take rate because they're providing real value: trust. A buyer pays a premium for the certainty that the sneakers are real. That's a different kind of subsidy calculus than a generic classifieds site. Luna: So the take rate ceiling is higher when the marketplace solves a bigger problem. Makes sense. Lucas: And that's really the key insight for operators: the subsidy isn't just a cost to be minimized. It's an investment in the trust and liquidity that earns you the right to charge a higher take rate later. The art is timing the withdrawal so you capture that value without killing the goose. Luna: I want to come back to something you said earlier, about how some subsidies are structural. What's a marketplace doing when it's stuck in that position? DoorDash, for example — are they just permanently locked into low margins? Lucas: That's the billion-dollar question. They're trying to offset it with things like advertising — restaurants pay to rank higher — and more efficient routing. But the core category has a natural take rate ceiling. So one strategy is to expand into adjacent categories where the ceiling is higher or the subsidy need is lower. Luna: Like when Instacart started offering ads to consumer brands. Or when Amazon built a whole ad business on top of their marketplace. Lucas: Exactly. Advertising is another way to monetize without raising the take rate on the core transaction. It's a way to extract more value from the network without breaking liquidity. Luna: Okay, so let's talk about the elephant in the room. As we're recording this in August 2026, the funding environment has been tougher than it was a few years ago. How does that affect the subsidy game? Lucas: It makes it a lot harder. In the cheap-money era, marketplaces could subsidize aggressively for years, betting that they'd figure out the economics later. Now, with capital more expensive, investors are demanding a clearer path to profitability. So you're seeing marketplaces either cut subsidies earlier — sometimes too early — or avoid getting into subsidy-heavy categories altogether. Luna: That's why you're seeing so many startups pivot to B2B marketplaces, where you can charge higher fees and the users are more rational — they're spending the company's money, not their own, so they're less price-sensitive. Lucas: Good point. B2B marketplaces tend to have lower subsidy needs because the value proposition is clearer — a manufacturer will pay for a platform that reliably finds them hard to source components, because the cost of a wrong purchase is huge. So the take rate ceiling is higher and the subsidy requirement is lower. That's a much easier equation. Luna: So if someone's listening and thinking about building a marketplace, what's the one lesson you'd want them to take away about subsidies? Lucas: I think it's this: don't think of a subsidy as a finite cost to be spent down. Think of it as a dynamic tool that you're constantly tuning. The goal isn't to eliminate subsidies — it's to make sure every dollar of subsidy is buying liquidity that creates real long-term value. And be honest with yourself about whether your category allows you to ever wean off it. Luna: And be honest with your investors too, right? The market has shown that they punish surprise subsidy persistence. Lucas: Absolutely. Set the expectation early. If you're building a food delivery app, don't tell investors you'll hit 20 percent margins in two years. Show them the unit economics of the peer group and be clear about which subsidies are structural. That transparency will pay off in trust and valuation multiple. Luna: Before we wrap up, I want to be upfront about something on our side. This show is completely ad-free, and that's only possible because a small group of listeners chips in a few bucks a month through buy me a coffee dot com slash fexingo. If you get value from these conversations, that's what keeps them coming. Lucas: Yeah, it genuinely does. We're not backed by a big network or sponsors, so listener support is the whole engine. And it means we can keep doing deep dives like this without worrying about advertisers. Luna: Right. So that's the honest ask — no pressure, but if you can, it helps keep this independent thing running. Lucas: And speaking of the engine — that's actually a good segue back to marketplaces. Because the same logic applies: the subsidy from loyal listeners is like the early liquidity that helps you reach a tipping point. Once you're there, the network effect kicks in and the value compounds. Luna: Ha, we're a two-sided marketplace of listeners and hosts. But okay, so let's land this. What's the takeaway for someone running a marketplace right now? Lucas: The biggest takeaway is to know your subsidy type. Are you in a category where you can logically reach a point where the network is self-sustaining and you can pull back? Or are you in a category like food delivery where the subsidy is a permanent economic feature? That distinction should drive your entire business model — how you price, how you raise money, how you talk to investors. Luna: And if you're in the second category, you need to find other ways to extract value from the network without raising the take rate. Lucas: Right. Advertising, adjacent services, data products — those are the ways to make a permanent-subsidy marketplace work. The ones that figure that out are the ones that survive the long run. Luna: So, next time you see a marketplace report a big subsidy line, pause before you judge. It might be a cold-start investment, or it might be the cost of doing business in a competitive category. The smart investor asks which one it is. Lucas: Exactly. And the smart operator is the one who knows the difference. Thanks for listening, and we'll be back soon with more on how marketplaces actually work.