Latest / Marketplace Businesses with Fexingo: Two-Sided Networks, Liquidity, and Take Rates / How Marketplaces Monetize Data Without Charging Users
Transcript
- Lucas: If you run a two-sided marketplace, the instinct is to make money on the transaction — take rate, listing fees, something per match. But there's an alternative revenue stream that doesn't touch the core exchange at all: selling data. Luna: Data you're already generating anyway, from the activity on your platform. Lucas: Exactly. And the canonical example here is OpenTable. They don't charge restaurants per reservation — they charge a monthly subscription for the software, but a big chunk of their revenue actually comes from selling anonymized data to companies that serve the restaurant industry. Luna: Who buys that data? I assume it's not consumers. Lucas: Right — it's suppliers. Think point of sale providers, reservation management software vendors, even kitchen equipment manufacturers. OpenTable packages things like average table turn times, peak dining hours, menu popularity trends, and sells it as a product called 'Diner Insights' and a broader 'Restaurant Revenue Growth' report. According to their investor materials from a few years ago, data products were generating roughly $80 million annually, which was around 12 percent of total revenue. And that's nearly pure margin — the data is a byproduct of the core business. Luna: So the marginal cost of producing that data product is essentially zero. The data is already flowing through the system. Lucas: That's the beauty of it. You've already paid for the infrastructure to collect the data — the reservation system, the diner check-in, the payment processing. Packaging it into a report costs a bit of engineering and maybe some data scientist time, but nothing like the cost of acquiring a new restaurant customer or building a new feature. Luna: And the buyers have strong incentives. A POS company wants to know which markets are growing so they can target their sales efforts. A reservation software vendor wants to know what features restaurants actually use. OpenTable has the best signal because they sit on both sides of the transaction. Lucas: That's the data network effect in action. The more restaurants and diners use OpenTable, the richer the data set becomes, and the more valuable the data product is to suppliers. It's a virtuous cycle that doesn't depend on raising the take rate. Luna: Let me push back a little though. Restaurants are giving OpenTable their data — reservation times, menu choices, party sizes. If they find out OpenTable is selling that data, even anonymized, doesn't that create a trust problem? Lucas: It can. OpenTable is very careful to anonymize and aggregate — they don't sell individual restaurant data, they sell trends across neighborhoods or cities. And they've been doing it for over a decade, so restaurants have largely accepted it. But trust is a real risk. If a restaurant feels like their data is being used against them — say, to help a competitor — you lose them. Luna: Grubhub does something similar, right? They sell order data to restaurant chains — what's selling where, peak times, delivery radius insights. Lucas: Yeah, Grubhub has a whole data services arm. They sell to both restaurants and food manufacturers. A big chain might use that data to decide which menu items to test in a new market. The key difference with Grubhub is that the data comes from delivery orders, which are a much more granular signal — they know exactly what was ordered, what time, what the tip was, how long the delivery took. Luna: And the restaurant is paying Grubhub a commission on each order already. So the data sale is just an additional revenue stream, not the main one. Lucas: Right. And that's actually the strategic sweet spot. You don't want data monetization to be your primary business model, because then you're incentivized to extract as much data as possible, which can hurt the user experience. But as a supplement, it's incredibly attractive. Luna: Another example I think of is CoStar in commercial real estate. They aggregate property listings, lease data, sales transactions — and they sell that data to brokers, appraisers, investors. But unlike OpenTable, CoStar actually charges for access to the data. Their whole business is data. Lucas: Yeah, CoStar is a pure data marketplace. They're different because the data itself is the product — they don't have a transaction-based business underneath. For OpenTable, the data product is a sidecar. The core business is still the reservation platform. Luna: So what's the lesson for a marketplace operator who's thinking about adding a data product? When does it make sense? Lucas: First, your core marketplace needs to be mature enough that the data is statistically meaningful. If you have only a hundred transactions a day, the data isn't valuable to anyone. You need volume. Second, you need to be able to anonymize effectively — if you can't guarantee that individual users or merchants can't be identified, the trust risk is too high. Third, you need a clear buyer — someone who will pay for insights they can't get elsewhere. Luna: And the buyer is almost never the end consumer of the marketplace. It's a B2B sale. Lucas: Exactly. You're selling to suppliers, to vendors, to people who serve your merchants. OpenTable sells to restaurant tech companies. Grubhub sells to food manufacturers. Uber sells data through Uber Movement to city planners and transportation researchers — that's a different buyer, but same principle. Luna: I remember Uber Movement launched in 2017 as a way to share anonymized trip data with cities. It was partly a good-faith gesture after all the regulatory battles, but it also builds goodwill and — I assume — creates a channel for selling premium data products later. Lucas: Right. Uber doesn't charge for Movement data publicly, but they've explored premium tiers for consultants and developers. The point is, even the free version generates value — it strengthens relationships with regulators, it demonstrates transparency, and it positions Uber as a partner, not just a disruptor. Luna: That's smart. It's data as a relationship tool, not just a revenue tool. Lucas: Yeah, and I think that's the broader insight. Data monetization in marketplaces is often more about reinforcing the core business than about creating a separate profit center. If you can use data to make your merchants more successful — by giving them insights they can act on — they'll be more loyal, they'll transact more, and your take rate revenue grows anyway. Luna: So it's a double win. But there's a risk of over-monetizing. If you start selling data in a way that feels extractive, you can kill the trust that makes the marketplace work. Lucas: Absolutely. And that's why most marketplaces start with the core transaction model and only layer data products once they have scale and trust. You don't see early-stage marketplaces trying to sell data — they don't have enough data, and they can't afford the trust hit. Luna: For listeners who are building or operating a marketplace, the question is: what data are you already sitting on that could be valuable to someone else — without hurting your users? Lucas: And that's a really practical exercise. Map out every data point you collect, ask who would find it valuable, and then think about whether you can package it in a way that's anonymized and aggregated. If you can, you might have a revenue stream that doesn't require raising your take rate or adding fees. Luna: We deliberately don't run ads on these episodes. If you want to support that choice, the link is buy me a coffee dot com slash fexingo. Lucas: Yeah, keeping this ad-free is a conscious decision. Listener support is what makes it possible. So if these conversations are useful for what you're building, that link is there. Luna: Alright, back to data products. One thing I want to dig into is the economics of switching costs. When a restaurant buys OpenTable's data dashboard, they build workflows around it. That creates lock-in. Lucas: Absolutely. The data product itself becomes a sticky layer. If you're a regional restaurant group and you use OpenTable's Diner Insights to decide which menu items to feature on weekends, you're not going to switch to a competing reservation system just because of a feature difference — you'd lose your historical data and the continuity of the trends. Luna: So the data product increases the moat around the core marketplace. It's a retention mechanism disguised as a product. Lucas: Exactly. That's a theme we've seen in other marketplaces too. Etsy, for example, offers sellers data on search trends and popular categories. It's part of their seller dashboard, and it keeps sellers invested in the platform. They're not selling the data externally — they're giving it to sellers for free — but the effect is the same: higher engagement and lower churn. Luna: So there's a spectrum. On one end, you give data away to users to improve their experience. On the other, you sell it to third parties. OpenTable does both — they give restaurants their own data for free, and sell aggregated trends. Lucas: Right. And the key is designing the data product so that it doesn't compete with your users. If you sell data that helps a supplier raise prices on your merchants, that's a problem. But if you sell data that helps suppliers serve your merchants better, everyone benefits. Luna: Let's talk about a specific case where data monetization went wrong. Any examples come to mind? Lucas: One that's often cited is the Facebook-Cambridge Analytica scandal, though that's not a marketplace per se. In marketplace land, there was controversy around how some real estate platforms sold lead data to agents without clear disclosure. If a home buyer enters their info on Zillow and then gets bombarded by agents who bought that data, that erodes trust. Zillow has since tightened its practices, but it shows the risk. Luna: So transparency is critical. The best data products are ones where the user understands what's being shared and either opts in or gets clear value in return. Lucas: Exactly. And that's why OpenTable's model works — restaurants know their data is being used to generate industry reports, and they get the reports themselves for free. It's a reciprocal relationship. Luna: For a marketplace operator listening, what's the first step? Should they start talking to potential data buyers, or build the product first? Lucas: I'd start with internal data audit and a quick validation with potential buyers. Talk to three to five companies that serve your merchants — ask them what insights they wish they had. Then see if your data can provide that. If the answer is yes, build a minimal version, maybe a monthly PDF report, and test whether they'll pay for it. The engineering cost is low, and the upside is a new revenue stream that only grows as your marketplace grows. Luna: That's a concrete takeaway. Thanks, Lucas. Next episode we'll look at how marketplaces manage fraud without killing liquidity. Lucas: Looking forward to it. For now, if this made you think of a data opportunity in your own business, that's exactly the kind of conversation we want to start.