Latest / Marketplace Businesses with Fexingo: Two-Sided Networks, Liquidity, and Take Rates / How the Best Marketplaces Use Referral Programs to Jumpstart Liquidity
Transcript
- Lucas: Back in 2014, Uber was spending roughly 40 percent of its marketing budget on referral credits. Not billboards, not TV ads — just giving riders and drivers cash to bring in their friends. Luna: Forty percent is huge. That's basically betting the whole growth engine on word of mouth. Lucas: Exactly. And it worked. Uber's referral program generated something like a 10 percent lift in sign-ups per week during that period. But here's the thing — referral programs are one of those strategies that sound easy in theory but are incredibly hard to get right. Luna: Right, because if you just throw money at people, you can end up with low-quality users or even fraud. So how do the best marketplaces design referrals that actually build liquidity? Lucas: Let's start with the classic example — PayPal. In 1999, PayPal offered new users $10 to sign up and another $10 for referring a friend. That's a $20 combined incentive for two clicks. Back then, their customer acquisition cost was about $20 to $30 per user. So the referral bonus essentially paid for itself if the referring user stuck around. Luna: And they famously grew to 100 million users in just a few years. But that was also a time when online payments were new — people needed a reason to trust it, and the cash incentive helped. Lucas: Right. But the real genius wasn't just the dollar amount. It was the viral coefficient. PayPal calculated that each new user brought in an average of 0.7 additional users through referrals. That's below 1, so it's not purely viral, but combined with paid acquisition, it was enough to drive exponential growth. Luna: So the key metric is the viral coefficient — how many new users each existing user recruits. For a marketplace, you want that number as high as possible, but you also need to balance it with quality. Lucas: Exactly. Airbnb learned that the hard way. In 2011, they launched a referral program offering $50 in travel credit to both the referrer and the new user. But initially, they didn't do any targeting — anyone could refer anyone. They ended up with a lot of low-intent users who never booked a trip. Luna: So they had to tighten it up. What did they change? Lucas: They started restricting referrals to people who had actually completed a booking. That way, the referrer had firsthand experience with the platform. They also introduced a cap on how many credits you could earn in a year. And they A/B tested everything — the offer amount, the email copy, the landing page. Eventually, they found that $25 was the sweet spot for their margin. Luna: Interesting. So the incentive size matters a lot. Too low and nobody bothers. Too high and you attract fraud or people who churn after cashing out. Lucas: Right. Uber's early 'Give $20, Get $20' program had the same problem. In some markets, people would create fake accounts using prepaid credit cards to claim the bonus, then never drive or ride again. Uber had to implement fraud detection — things like requiring a verified phone number and a minimum number of trips before you could cash out. Luna: So fraud prevention is a core part of referral program design. What about timing? When should a marketplace even launch a referral program? Lucas: Most founders make the mistake of launching too early. If your product isn't good enough to retain users, a referral program just accelerates churn. You're paying to bring people in who then leave. I've seen data that suggests you need a retention rate of at least 30 percent after 90 days before referrals become net positive. Luna: That makes sense. Because if you're bleeding users, more acquisition just makes the leak bigger. So first fix retention, then turn on referrals. Lucas: Exactly. Another mistake is treating both sides of the marketplace the same. In a two-sided network, liquidity comes from having enough supply and enough demand. But the referral incentive might need to be different for each side. For Uber, driver referrals often had higher bonuses than rider referrals because driver supply was the harder constraint. Luna: That's a great point. In a marketplace, you need to figure out which side is harder to acquire and then tilt the referral incentive toward that side. Lucas: Right. And once you reach a certain density, you can actually reduce or even eliminate the referral bonus. Uber did that in many cities once they had enough drivers — they stopped the driver referral program entirely and focused only on rider referrals. The logic: once supply is abundant, demand growth becomes the priority. Luna: So the referral program is not static. It evolves as the marketplace matures. That's a key insight. Lucas: Let's talk about a less obvious example — Etsy. Etsy doesn't have a traditional referral program with cash bonuses. Instead, they rely on seller-generated word of mouth. Sellers tell other sellers about the platform because they want more buyers to come. But Etsy does run occasional referral campaigns where sellers get a small credit toward listing fees if they bring in a new seller. Luna: That's interesting because it's not about cash — it's about reducing friction for the referrer. The credit is useful only if you're already an active seller, so it reinforces the behavior they want. Lucas: Exactly. The best referral programs align the incentive with the platform's currency. Airbnb uses travel credits, Uber uses ride credits, PayPal used cash. Cash is the most flexible but also the most prone to fraud. Platform-specific credits keep users engaged within the ecosystem. Luna: So what's the takeaway for someone building a marketplace today? What should they test first? Lucas: Start with a simple question: what is the single most valuable action a new user can take on your platform? For a ride-hailing app, it's taking their first ride. For a freelance marketplace, it's completing their first project. Then design your referral incentive to reward that action — not just sign-up. Luna: So reward the 'aha moment,' not the account creation. That way you're paying for behavior that actually builds liquidity. Lucas: Right. And measure your viral coefficient religiously. If it's below 0.2, your referral program probably isn't working hard enough. If it's above 1, congratulations, you've got a self-sustaining loop. Most marketplaces hover around 0.3 to 0.5, meaning referrals are a meaningful but not dominant channel. Luna: One thing I've noticed is that some marketplaces now use multi-sided referrals — like, refer a friend who becomes a driver, and both you and the driver get a bonus, plus the first 10 riders they pick up get a discount. That's a chain reaction. Lucas: Yeah, that's a more sophisticated approach. Lyft experimented with that in 2015 — they called it the 'Lyft Line' referral. It worked well in dense urban areas but was harder to execute in suburbs. The key is to not overcomplicate the offer. Simplicity drives sharing. Luna: Before we wrap up, I'm curious — are there any marketplaces where referrals just don't work? Lucas: Absolutely. If your platform is extremely niche or has a very small total addressable market, referrals might not generate enough volume. Also, if your product has a long purchase cycle — like a marketplace for industrial equipment — referrals are less effective because the time between referral and conversion is too long. Luna: Right, so referrals work best when the product is frequently used and socially visible. Ride-sharing, short-term rentals, food delivery — those are perfect. Lucas: Exactly. And one last thing — always have a way to track referral attribution. Use unique links or codes. Without that, you can't measure ROI, and you'll never know if the program is actually creating liquidity or just burning cash. Luna: This has been really practical. If listeners found this breakdown useful, you know, these deep dives into marketplace strategy only stay ad-free because of listener support. Lucas: Yeah, it's a small thing but it makes a big difference. If today's episode gave you a concrete idea you can use, consider throwing a few dollars at buy me a coffee dot com slash fexingo. It's how we keep the show independent and focused on what works. Luna: Right, no ads, no sponsors — just the stuff we'd want to hear ourselves. So if you're into that, the link is there. Lucas: Okay, back to referrals. I want to leave you with one data point. In 2019, a study of over 100 marketplace startups found that those with a formal referral program grew their user base 3.5 times faster than those without, controlling for other growth channels. That's a massive multiplier. But only if you design it right — reward the right action, target the right side, and turn it off when it stops working. Luna: And don't forget fraud detection. That's the hidden cost that can kill your unit economics. Lucas: Absolutely. So next time you're building a marketplace, think about referrals not as a growth hack, but as a core part of your liquidity strategy. Get it right, and you'll build a network that grows itself.