Latest / Marketplace Businesses with Fexingo: Two-Sided Networks, Liquidity, and Take Rates / How Marketplaces Use Referral Programs to Solve Cold Start
Transcript
- Lucas: If these conversations are useful for what you're building or running — today I want to talk about one of the oldest growth tricks in the marketplace playbook, and it's something almost every successful two-sided network has used at some point: the referral program. Luna: Referral programs feel almost too simple to be a real strategy. But I remember reading that Uber's early driver referral bonus was one of the single biggest levers for getting supply in new cities. Lucas: Exactly. And it's not just Uber. Airbnb, TaskRabbit, Etsy — they all leaned heavily on referrals at different stages. The cold start problem — how do you get enough supply and demand on a platform when neither side wants to show up unless the other is already there — referrals address that in a very specific way. Luna: Because you're not just asking anyone to join. You're asking someone who already has a relationship with a current user. Lucas: That's the key insight. A referral brings in someone who already has some trust in the platform through the referrer. It's a warmer lead than any Facebook ad or Google search term. And for marketplaces, trust is the currency that makes liquidity happen. Luna: So let's talk about the mechanics. Uber's early driver referral — I think it was something like $500 for referring a new driver who completed a certain number of rides. Lucas: It varied by city, but yeah, that ballpark. The structure mattered: it wasn't just 'sign up, get cash.' The new driver had to complete, say, 20 rides within a month. That ensured the referral was actually adding meaningful supply, not just someone who signed up and never drove. Luna: And Uber paid the referring driver in stages, right? Part when the new driver completed their first ride, the rest after they hit the threshold. Lucas: Right. That kept both parties engaged. The referrer had an incentive to coach the new driver to stick with it, because their full bonus depended on it. That created a kind of informal mentor network, which helped retention. Luna: Airbnb took a different approach. Instead of cash, they offered travel credits — $25 for the referrer and $75 for the new user, but only after the new user completed their first booking. Lucas: And that's genius for a marketplace. The credit can only be spent on the platform, so it's self-reinforcing. The referrer gets a future booking discount, which creates more demand. The new user gets a discount on their first trip, which makes them more likely to become a host later. It's a liquidity loop in microcosm. Luna: TaskRabbit used a tiered system — I think they gave the referrer a $50 credit after the new tasker completed their first task, and another $50 after ten tasks. Lucas: That two-stage approach is smart because it incentivizes the referrer to care about long-term quality, not just a quick sign-up. TaskRabbit's problem was always about having enough reliable taskers in each city. A one-time referral might bring in someone who does one job and quits. But if the referrer gets a bonus for the new person completing ten tasks, they'll help onboard and train them. Luna: There's a downside though. Referral fraud. People gaming the system by referring fake accounts or buying referrals. Lucas: Absolutely. Lyft had a well-known issue where some drivers were creating fake rider accounts to refer themselves. They had to cap referral earnings and add verification steps. Uber also had to tighten its program over time, requiring phone number verification and real ID checks. Luna: So the design of the referral program needs to balance generosity with safeguards. Too loose, and you get fraud. Too tight, and nobody bothers. Lucas: Right. And the timing matters almost as much as the amount. The best programs launch when the marketplace is in a specific growth phase — not too early, when there's barely any liquidity to show off, and not too late, when word of mouth is already happening organically. Luna: When do you think is the sweet spot? Lucas: I'd say after you've achieved product-market fit in one or two dense markets. You have proof that the marketplace works. You have users who've had a good experience. Then you ask them to bring their friends. That's when the referral program amplifies existing momentum instead of trying to create it from nothing. Luna: That makes sense. The referral is an amplifier, not a starter pistol. Lucas: Exactly. And the data from referral programs can tell you a lot about your best users. Who refers the most? What kind of users do they bring in? Do those referred users stick around longer? Marketplaces like Etsy have used referral data to identify power users and then target them for other initiatives, like becoming community moderators or beta testers. Luna: So the referral program isn't just a growth channel — it's a diagnostic tool. Lucas: For a good operator, yeah. You learn which segments have the highest network value. And you can adjust the incentive structure accordingly. Maybe you find that referrals from hosts in a particular city have a higher retention rate, so you double down there. Luna: This is exactly the kind of thing that makes me appreciate how much thought goes into what looks like a simple 'refer a friend' button. Lucas: Quick honest thing — a handful of listeners chip in monthly through buy me a coffee dot com slash fexingo, and that's literally what funds making this many of these episodes possible. Luna: Yeah, it's a small group of people, but it keeps the show ad-free and lets us drill into exactly these kinds of mechanics. Lucas: So back to referrals — one thing we haven't talked about is how marketplaces decide between a single-sided vs. double-sided referral program. Should you only reward the referrer, or the new user too? Luna: I think Airbnb's model is a good example of double-sided. Both sides get a credit, which makes both feel valued. Lucas: And it reduces friction for the new user. If only the referrer gets a reward, the new user might feel like they're being used. But if they also get something, it feels like a shared benefit. Luna: Uber did double-sided too, right? The new rider got a free ride, and the referring rider got a credit. Lucas: Yes, and that was crucial for the demand side. When Uber was entering a new city, they needed riders to try the service. A free ride was a powerful incentive. And the referring rider got a credit for their next trip, which created repeat usage. Luna: So the structure is: reward both sides, but with different currencies or amounts depending on which side you're trying to grow. Lucas: Exactly. If you need more supply, tilt the reward heavier toward the referrer on the supply side. If you need more demand, tilt toward the new user. And you can change that balance over time as the marketplace evolves. Luna: I've also seen marketplaces use referral programs to expand into new geographic areas. Like, if you refer someone in a new city, you get a bigger bonus. Lucas: That's a smart tactic. TaskRabbit did something like that when they expanded to new metro areas. They offered a higher referral bonus for the first 100 taskers in a new city. That created a rush of supply on day one. Luna: But you have to be careful not to cannibalize existing supply. If you pay more for referrals in a new city, existing taskers in your core city might feel undervalued. Lucas: That's a real risk. You have to frame it as a growth initiative, not a permanent differential. And you can cap it — first 100 referrals get the bonus, then it reverts to the standard rate. Luna: One more angle: how do you measure the success of a referral program? Is it just cost per acquisition? Lucas: Cost per acquisition is a start, but the better metric is lifetime value of referred users vs. non-referred users. Studies have shown that referred customers often have higher retention and higher spend. If that's true for your marketplace, you can afford to pay more per referral. Luna: And you should track the referral multiplier — how many new users each referrer brings in. If you have a handful of super-referrers, maybe you create a special program for them. Lucas: Exactly. Some marketplaces have created ambassador programs for their top referrers, giving them early access to new features or a higher reward tier. That turns a transactional referral program into a community-building tool. Luna: So the referral program is not a one-size-fits-all tactic. It's a system that needs to be tuned constantly. Lucas: And that's why it's such a good lens for understanding marketplace dynamics. The challenges of cold start, liquidity, trust, and growth all intersect in the design of a good referral program.