Latest / Marketplace Businesses with Fexingo: Two-Sided Networks, Liquidity, and Take Rates / How Marketplaces Use Referral Programs to Grow Supply
Transcript
- Lucas: So earlier this week I was looking at some old data from Uber's early driver-recruitment playbook, and one number jumped out: in 2014, Uber was spending roughly six hundred dollars per new driver acquired through their referral program. That's more than double what a typical lead cost from a Google search ad was back then for a ride-hailing service. Luna: Six hundred bucks per driver? That seems steep. Did it pay off? Lucas: Incredibly well, actually. Because that referral program wasn't just about getting any driver — it was about getting the drivers. The referred driver came pre-vetted by an existing driver who knew the job. And that existing driver had a financial incentive to recommend someone reliable, because their own earnings depended on the network's reputation. Luna: So it was a quality filter built into the incentive. Interesting. I remember Airbnb did something similar — but they offered travel credits instead of cash. Lucas: Right. And that distinction — cash versus credit — is actually a really important design choice. Uber's cash referral of five hundred dollars per new driver was expensive up front, but it created an immediate, tangible reward. Airbnb's credit-based referral meant the host got a future discount on a stay, which reinforced usage of the platform itself. Luna: So one is a direct cash incentive, the other is a platform-loyalty loop. Which one works better for supply acquisition? Lucas: It depends on your marketplace's unit economics and how urgently you need supply. Uber was in a land-grab: they needed drivers in every new city within weeks. Cash referrals moved fast because they appealed to drivers who had immediate financial needs — think part-time gig workers. Airbnb had more time to build its host base, so they could afford a slower, more engaged referral that deepened the host's relationship with the platform. Luna: And the speed of the payout matters too, right? Uber paid out within weeks. Airbnb's travel credits could take months to use. Lucas: Exactly. The timing of the reward changes who you attract. Immediate cash tends to attract more transactional, price-sensitive suppliers. Delayed credit attracts people who are already somewhat invested in the platform. Both can work, but they serve different stages of marketplace maturity. Luna: I want to talk about the viral coefficient here. For a marketplace, referral programs can create a flywheel — more supply brings more demand, which attracts more supply. How do you measure that effect? Lucas: You look at the K-factor, or viral coefficient. It's the number of new suppliers each existing supplier brings in. If it's above one, the supply side grows exponentially without paid acquisition. Uber's early driver referral program had a K-factor of about 1.2 in some cities — meaning every ten drivers brought in twelve new ones. That's powerful. Luna: But you can't keep that up forever, right? Eventually the pool of friends and acquaintances runs dry. Lucas: Right. The viral coefficient decays over time as the market saturates. That's why the best referral programs are timed to the early growth phase. Once a marketplace reaches a certain density, referral shifts from a primary growth engine to a retention and quality tool. Luna: Let's talk about referral program structure. Some marketplaces offer a two-sided reward: both the referrer and the new supplier get something. Others only reward the referrer. What's the trade-off? Lucas: Two-sided rewards tend to have higher conversion rates because the new supplier feels welcomed — they get a bonus just for signing up. But it's more expensive. Uber did a two-sided cash incentive: the existing driver got five hundred dollars, and the new driver got a smaller bonus after completing a certain number of trips. That helped retention too, because the new driver had a reason to stay through the first few weeks. Luna: I've seen some marketplaces like TaskRabbit use a different approach: they give the referrer a discount on their next service fee, not cash. That keeps the reward within the platform's ecosystem. Lucas: That's a variant of the credit model. It lowers your cash burn and increases the likelihood that the referrer will transact again. But it only works if the referrer's existing demand is high enough that they value the discount. For a casual user, a fifty-dollar credit is worth less than fifty dollars cash. Luna: So you have to know your user segments. A power supplier values platform-specific rewards; a marginal supplier wants cash. Lucas: Exactly. And that's why many marketplaces start with cash to bootstrap supply, then shift to credits or tiered rewards as the marketplace matures and they have more data on their suppliers. Luna: One more nuance: the referral program's distribution channel. Uber embedded the referral link directly in the driver app. Airbnb used email and social sharing. Does the channel affect the quality of referrals? Lucas: Massively. In-app referrals tend to come from more engaged users because they're actively using the platform. Email and social sharing can generate volume, but the quality is more variable. Uber's in-app referral had a higher conversion rate than any other channel because the driver was already in the mindset of earning. Luna: So the ideal referral program is two-sided, in-app, cash-based for early growth, then evolves into a credit-based loyalty program. That sounds like a formula. Lucas: It's a common pattern, but you have to adapt it to your specific marketplace dynamics. For example, in a high-stakes marketplace like Upwork where the supplier is a freelancer with a portfolio, a referral program might emphasize quality over quantity — lower reward, but only for referrals who complete their first project. Luna: Right, because a bad freelancer can damage the platform's reputation with clients. So you want to gate the reward on performance. Lucas: Exactly. That's a deferred reward tied to a quality milestone. It reduces the risk of low-quality supply flooding in. Some marketplace operators call this a 'quality-adjusted referral.' The reward is higher for referrals that stick around and transact. Luna: If these kinds of strategic conversations are useful for what you're building or running, consider supporting the show. Listener support is what keeps us ad-free and focused on real marketplace mechanics. Lucas: Yeah, it makes a difference. If you'd like to chip in, you can find us at buy me a coffee dot com slash fexingo. No pressure, just helps us keep the episodes coming. Luna: So back to referral programs: one thing I find fascinating is how some marketplaces use referral programs to solve the cold-start problem on the supply side. For example, when Uber launched in a new city, they'd often have no drivers at all. Lucas: Right. The referral program was a way to get the first few drivers on the road. They'd target existing drivers from other platforms or people who already drove for a living — taxi drivers, limo services — and offer a generous referral bonus if they brought in colleagues. Luna: That's essentially using referral to bootstrap liquidity from scratch. It's a classic chicken and egg solution: you pay the early suppliers to recruit others, so the network grows without you having to spend massively on brand advertising. Lucas: And the beauty of it is that the referral program itself creates a social proof loop. When a driver sees their friend earning money from Uber, they're more likely to sign up. That personal endorsement is far more credible than any ad. Luna: What about the risk of referral fraud? Some marketplaces have seen people gaming the system by referring fake accounts. Lucas: It's a real problem. Uber had to implement safeguards like requiring the new driver to complete a certain number of trips before the referrer got paid. Airbnb required the new host to complete a booking. Those friction points reduce fraud but also slow down the growth. It's a trade-off. Luna: So the best referral programs balance speed of growth with fraud prevention. That's a hard line to walk. Lucas: It is. But the marketplaces that get it right build a sustainable supply engine. I think the key insight is that referral programs aren't just about acquisition — they're about creating a community of advocates who have a stake in the platform's success. Luna: And that community effect is what ultimately drives the network effects that make a marketplace defensible. Lucas: Exactly. When your suppliers become your recruiters, you've turned your growth into a self-reinforcing loop. And that's the ultimate goal of any marketplace.