Latest / Marketplace Businesses with Fexingo: Two-Sided Networks, Liquidity, and Take Rates / How OpenTable Uses Secondary Inventory to Fix Liquidity
Transcript
- Lucas: A few weeks ago I tried to book a Friday night dinner at a popular spot in Austin. OpenTable showed nothing between 7 and 9. But I knew that restaurant. I knew they held tables back. So the inventory you see in the search results isn't the full picture. Luna: Right, and that gap between what's available and what's actually possible is the central puzzle for any marketplace with perishable supply. Lucas: Exactly. And the reason I bring up OpenTable is that it's a perfect case study in how to solve that gap using what I'll call secondary inventory. Not the main feed of bookable slots, but a parallel layer that captures demand when the primary inventory looks empty. Luna: So we're talking waitlists, standby, that kind of thing? Lucas: Precisely. For OpenTable, the primary inventory is the set of time slots a restaurant chooses to make bookable online. But any restaurant manager knows you never release all your tables. You hold some for walk-ins, for regulars, for VIPs. So the supply visible to OpenTable is always a subset. Luna: Which means the marketplace looks less liquid than it actually is. Fewer matches happen, and OpenTable's take-rate revenue suffers. Lucas: That's the problem. OpenTable makes money per seated diner, not per reservation made. If a diner searches, sees nothing, and walks away, OpenTable gets zero. But the restaurant might still have a table at 7:45 if the 7:00 party finishes early. Or a cancellation could open up right when someone walks in. Luna: So how do you capture that without annoying diners with false hope? Lucas: You create a waitlist. OpenTable's waitlist isn't just a FIFO queue. It's a dynamic matching engine. When a diner joins the waitlist for a specific time window, OpenTable monitors cancellations and table turn times in real time. If a table opens up, it notifies the first person in line with a match. Luna: And the restaurant has control too, right? They can set how many waitlist slots to accept per time block. Lucas: They can. And that's key. The restaurant is effectively creating a second, flexible inventory pool. They decide how much of their hidden capacity to expose through the waitlist. OpenTable's algorithm then optimizes for the likelihood of a seat actually materializing, so you're not getting false alerts. Luna: So the waitlist acts like a shock absorber for the primary inventory. Lucas: Exactly. And this matters beyond restaurants. Think about hotels. Expedia sells rooms, but a hotel always holds a few for walk-ins. A marketplace like HotelTonight or the opaque channel on Priceline is essentially a secondary inventory layer—selling the rooms that the hotel wouldn't put on the main shelf. Luna: Or think about service marketplaces like Thumbtack. A plumber might not show all available time slots, but if a customer sends a request, the plumber can still accept if it fits. That's secondary inventory too. Lucas: Right, and the economic insight is this: primary inventory is priced and listed in advance. Secondary inventory is discovered and matched close to real time. The friction is higher, but the yield is better because you're monetizing capacity that would otherwise be wasted. Luna: And for the marketplace, secondary inventory directly increases liquidity. More matches happen, take rate grows, and the network gets more valuable. Lucas: Now, the interesting thing about OpenTable's specific implementation is that the waitlist also feeds back into the primary inventory. If a restaurant sees that a certain time slot gets a lot of waitlist demand, they might decide to release more tables to OpenTable in that slot next week. So secondary inventory becomes a signal for supply expansion. Luna: So it's not just a band-aid. It's a data channel. Lucas: Exactly. And that's a lesson for any marketplace founder. If your core inventory looks thin, don't just try to recruit more suppliers. Ask whether you're seeing all the supply that actually exists. Maybe you need a secondary inventory layer to unlock it. Luna: Before we go further, I want to tie this back to something you said earlier about take rate. You mentioned that OpenTable only gets paid when a diner is seated. So the waitlist doesn't just improve liquidity—it directly improves revenue per search. Lucas: That's right. Without the waitlist, a diner who sees no availability leaves. OpenTable earns zero. With the waitlist, that same diner might still get seated 30 percent of the time, and OpenTable collects its fee. It's a pure incremental revenue stream with no incremental cost to the restaurant. Luna: And if you think about the take rate itself, OpenTable typically charges around two to three dollars per seated diner. The waitlist doesn't change the fee, but it increases the volume of fee-generating events. Lucas: So it's a volume play. And the beauty is that the restaurant loves it because they fill more seats. The diner loves it because they get a table they thought they couldn't have. OpenTable loves it because they make more money. That's the kind of three-way alignment great marketplaces create. Luna: It reminds me of something I read about the early days of Uber. They didn't just show available cars. They had a 'request' button that essentially created a secondary inventory of drivers who were close to being available but weren't yet online. Lucas: Right, same concept. And Uber's surge pricing is another form of secondary inventory management—it dynamically prices demand to match available supply. But waitlists are lower friction because they don't require a price change. Luna: Okay, so OpenTable's waitlist is a specific feature. But the broader lesson is that marketplaces should design for the gap between listed and actual supply. Lucas: And if these kinds of structural insights are useful for what you're building or running—if you're thinking about how to design a marketplace or improve liquidity in your own platform—then you might appreciate that this show stays ad-free precisely because listeners support it directly. It's a small gesture, but it means we can keep digging into mechanics like this without sponsors shaping the conversation. Luna: And we really do mean small. It's the digital equivalent of buying your coworker a coffee after they share a great tip. Lucas: If that feels right to you, you can head to buy me a coffee dot com slash fexingo. All lowercase. And thank you to everyone who already has—it genuinely helps. Luna: Alright, back to secondary inventory. One more example I want to hear your take on: Airbnb. Do they have a secondary inventory problem? Lucas: Interesting question. Airbnb's primary inventory is the set of listings that are bookable for specific dates. But hosts also have hidden capacity—maybe they block off nights for family or for last-minute cleaning. Some superhosts use 'instant book' with strict cancellation policies, which acts like a primary inventory. Others manually approve each booking, which is more like a waitlist. Luna: But Airbnb doesn't have a formal waitlist for a specific property. They have 'save' and 'wishlist' which are more like bookmarking. Lucas: Actually, they do have a version of it with their 'contact host' feature. If a listing shows as unavailable, you can still message the host. That's a secondary inventory channel. The host might say, 'Actually, I can open that weekend if you book three nights.' So it's less automated than OpenTable, but the principle is the same. Luna: So the question becomes: how do you make that secondary channel efficient without adding too much friction for the user? Lucas: That's the design challenge. OpenTable's waitlist is almost frictionless—you set your preference and get a notification. Airbnb's message-based system requires active back and forth. But the trade-off is that the host has more control and can offer custom deals. Different marketplaces will optimize differently. Luna: And the liquidity impact? For OpenTable, the waitlist probably increases total seated diners by 10 to 20 percent for popular restaurants. That's a huge lever. Lucas: I've seen some analyses suggesting it's even higher for prime-time slots. And crucially, it doesn't require the restaurant to change its behavior. They already had those tables. The waitlist just makes the inventory visible at the moment it becomes available. It's a timing fix, not a supply fix. Luna: So for anyone building a marketplace, the takeaway is clear: look at your 'no results' page. That's where your hidden liquidity lives. Lucas: And we're not saying every marketplace needs a waitlist. But you need a mechanism to capture demand when primary inventory is thin. It could be a waitlist, a request form, a callback request, a predictive availability badge—something that keeps the user engaged instead of bouncing. Luna: Next time, we'll look at how marketplaces use 'deposit to hold' features to guarantee that a matched booking actually converts—because even secondary inventory doesn't help if the diner doesn't show up. Lucas: Good preview. For now, if you want to dig deeper into OpenTable's specific data on waitlist conversion, I'll drop a link in the show notes to a case study from their engineering blog. Thanks for listening, and we'll be back tomorrow.