Latest / Marketplace Businesses with Fexingo: Two-Sided Networks, Liquidity, and Take Rates / How Marketplaces Use Reviews to Build Trust and Raise Take Rates
Transcript
- Lucas: So earlier this week I was booking an Airbnb for a trip in September, and I caught myself doing something I almost never think about. I was ignoring the star rating. Luna: Interesting — because most of us have been trained to treat those five stars as gospel. What made you ignore them? Lucas: I realized the average was 4.93 — and so was the next listing, and the one after that. At that point, the signal is completely compressed. You can't differentiate. So I started reading the written reviews instead, and that's where I found the useful stuff. Luna: Right, the classic problem of rating inflation. And that's actually a huge strategic issue for marketplaces, because if reviews don't help users trust a listing, then the whole marketplace's value proposition weakens. Lucas: Exactly. And this episode is about how marketplaces use reviews to build trust — and then, crucially, how that trust enables them to raise their take rate without scaring users away. Because a review system is not just a feature. It's pricing infrastructure. Luna: Let's anchor on that. Airbnb is the obvious case, but maybe we start with Uber — because they had a very clean, very early feedback loop that actually shaped the entire product. Lucas: Great call. Uber launched with a simple five-star rating for drivers, and riders could also rate passengers. The driver's rating directly affected their ability to keep driving — below a certain threshold, they got deactivated. That created a strong incentive for drivers to be polite, show up on time, keep the car clean. And for passengers, it meant they knew the driver was vetted by thousands of prior riders. Luna: But here's the thing — Uber's ratings also suffer from inflation. Most drivers are between 4.8 and 4.9. Anything below 4.6 is basically a red flag. So the system still works, but it's not very granular. Lucas: Right. And Uber hasn't really changed that system in a decade. But Airbnb, to their credit, recognized that rating inflation was eroding trust. In 2022, they overhauled their review system. Previously, they used a simple average of stars across categories like cleanliness, location, value. But they found that most listings clustered between 4.5 and 5.0 stars. Luna: So what did they change? Lucas: They moved to a system where the overall rating is still a star average, but they also show the distribution of ratings — how many 5-star, 4-star, etc. And they prominently display a summary of the most recent written reviews. They also removed the exact decimal average from the search results page — so you see a badge like '4.9 ★' but not the precise number. That small change forces users to click in and read. Luna: And did that help? I remember reading that they saw a meaningful increase in booking conversion for listings with strong written reviews, even if the star rating was similar to competitors. Lucas: Yeah, internal data — which they shared at a conference — showed that listings with at least ten recent written reviews had a 7% higher booking conversion than those with only star ratings. That's massive in a marketplace that processes billions of dollars in bookings. Luna: So the written reviews became a trust signal that the stars alone couldn't provide. And that trust allows hosts to charge more — because guests are willing to pay a premium for a place they feel confident about. Lucas: Exactly. And that's where the take rate angle comes in. Airbnb's take rate — the commission they charge hosts — is around 14 to 15 percent on average. But for listings with very strong reviews and high trust, they could theoretically raise that take rate because the host is earning a premium anyway. The host might grumble, but if the platform is delivering high-converting guests, the host's incremental revenue outweighs the fee increase. Luna: Do we have any evidence that Airbnb actually does that? I mean, do they charge different take rates based on review quality? Lucas: Not explicitly, no. Airbnb doesn't have a publicly disclosed variable take rate based on ratings. But what they do is use reviews to power a premium tier — the Airbnb Plus program, which featured verified quality listings. Those hosts paid a slightly higher fee in exchange for a badge and better placement. The reviews helped identify which listings qualified. Luna: So reviews become a gatekeeper for premium features. That's a smart way to increase take rate without calling it a fee hike. Lucas: Let's contrast that with a B2B marketplace like Faire, which connects retailers with brands. Faire's challenge is that retailers are buying inventory they'll resell — so trust is even more critical. They rely heavily on reviews from other retailers, and they display them prominently. A brand with dozens of positive reviews can charge higher wholesale prices, and Faire takes a commission on that. Luna: And Faire's take rate is higher than Airbnb's — I think it's around 25 percent, right? Partly because they offer payment terms and returns. But the reviews justify that fee because they reduce the risk for the retailer. Lucas: Exactly. The review system is essentially a substitute for the retailer having to touch and feel the product. Without reviews, Faire would have to offer even more expensive trust mechanisms — like free returns or samples — which would eat into their margin. Reviews are a low-cost way to build trust, and that lets Faire keep a higher take rate. Luna: So the pattern is: a good review system increases trust, which increases willingness to transact, which lets the marketplace charge more. But there's a risk — if reviews are gamed or inflated, trust collapses. Lucas: That's the dark side. Amazon faced this — fake reviews became so pervasive that they lost credibility for a while. They had to invest heavily in AI and human moderators to clean up the system. And that cost money. So maintaining review quality is an ongoing operational expense. But it's worth it, because without reviews, Amazon's marketplace would probably have a much lower take rate — or would need to offer huge return windows to compensate. Luna: So if I'm building a marketplace today, the lesson is clear: invest in your review system early. Don't just copy the five-star template. Think about what signals actually build trust in your specific vertical. Lucas: Right. For a service marketplace like Thumbtack, the review might need to include timeliness, quality of work, communication — not just a star. For a luxury goods marketplace, maybe the review needs to verify authenticity. The more tailored the review, the more trust it builds, and the more pricing power the marketplace has. Luna: And that pricing power translates directly into a higher take rate. So reviews aren't just a feature — they're a lever for monetization. Lucas: Yeah, and if you get it right, you can even use reviews to dynamically adjust your take rate. Imagine a marketplace that charges a lower fee for new sellers with no reviews, and a higher fee for established sellers with a proven track record. The seller pays for the trust that the marketplace has helped them build. Luna: That's a fascinating idea. It would align incentives — the marketplace invests in helping the seller get good reviews, and then shares in the upside. Lucas: Honestly, if today was worth a coffee to you, that's the link — buy me a coffee dot com slash fexingo. Keeps these conversations ad-free and independent. Luna: Yeah, it's a small way to support the work we do. And we really appreciate it. Lucas: So back to that idea — there's a startup called Provenance that's trying exactly that for food marketplaces. They verify supplier claims through reviews and audits, and then the marketplace can charge suppliers a higher commission because buyers trust the verification. It's early, but the logic is solid. Luna: I want to see that work at scale. Because trust is the hardest thing to build in a marketplace, and the easiest to lose. If you can codify it into a review system that is both informative and trustworthy, you've built a real moat. Lucas: That's the takeaway. Reviews are not a feature you add after launch. They are a strategic asset that directly affects your ability to charge fees. And the better your review system, the higher your take rate can go — without losing users. Luna: And the higher the take rate, the more resources you have to invest in trust and safety. It becomes a virtuous cycle. Lucas: Right. So next time you're on a marketplace, pay attention to how they handle reviews. It's telling you a lot about their business model.