This episode breaks down how to build and invest in marketplace businesses (like Airbnb or eBay). Jeff Jordan from a16z says the key is network effects—the more users, the harder to beat. He likes Airbnb and Instacart because their supply is fragmented (many small hosts or stores) and they create new transactions. He warns about 'leaky bucket' problems: if users don't stick, you'll burn cash on ads forever. Three key holdings: Airbnb (CEO spent 5 years perfecting the brand), Instacart (lost $18 per order initially, now profitable via ads), and DoorDash (his biggest miss—he feared too many competitors).
At a Glance Andreessen Horowitz partner Jeff Jordan shares operational experience from his roles as eBay’s general manager, PayPal’s president, and OpenTable’s CEO, along with insights from investing in companies such as Airbnb, Instacart, and Pinterest. Core thesis: marketplace platforms must focus
Jeff Jordan is a General Partner at Andreessen Horowitz. He has served as General Manager of eBay.com, President of PayPal, and CEO of OpenTable, and currently sits on the boards of Airbnb, Instacart, and Pinterest. This episode's main thread: deconstructing the logic of building marketplaces, investment frameworks, and common pitfalls from the dual perspective of an operator and an investor. Jeff Jordan's core judgment: The most critical defensibility of a marketplace comes from network effects, and to determine whether a marketplace is worth investing in, one must first assess whether the supply side is sufficiently fragmented—if any single supplier holds market power (e.g., AMC for Fandango), the platform will struggle to build a lasting moat.
Jeff Jordan believes that evaluating whether a marketplace platform is worth investing in first requires examining two characteristics: the degree of supply-side fragmentation and whether the platform possesses genuine lead generation capabilities.
Jordan uses a comparison between OpenTable and Fandango to illustrate the importance of fragmentation. Restaurants on OpenTable, on average, each operate only a single location, so aggregating them is labor-intensive, but once accomplished, the value is immense. In contrast, Fandango partners with just five or six major theater chains, and the departure of any one (such as AMC) would significantly weaken the platform's value. Jordan recalls explaining this theory to a board member, who gave him an odd look — only later did he learn that the person was Michael Klein, the founder of Fandango.
The second characteristic is lead generation — whether the platform can facilitate transactions that would not otherwise occur. Jordan points out that the most dangerous scenario is when users already have an established relationship with a service provider (e.g., auto mechanics, barbers), making such platforms difficult to succeed, because service providers are only willing to pay for acquiring new customers, not for the convenience of existing ones. Airbnb is an ideal case: hosts are introduced to guests they have never met, and the average host owns only one property, resulting in extremely high fragmentation.
Jordan emphasizes that the best marketplace platforms rely almost entirely on organic customer acquisition, not paid channels. The OpenTable case is a classic example: the team developed a widget that restaurants could place on their own websites to offer online reservation functionality. When users clicked on it, they were redirected to the restaurant's page on OpenTable, where they discovered they could book all restaurants on OpenTable, and thus returned directly to OpenTable instead of Google. Jordan says: "During my time at OpenTable, I didn't spend a single dollar on demand acquisition, and the business grew very well."
Jordan points out that the two signals market platform investors should be most wary of are: fundamentally flawed unit economics, and a persistently declining user retention curve (the "leaky bucket" problem).
Jordan uses the case of Instacart to illustrate the importance of unit economics. At the time of investment, Instacart's revenue per order was about $12, but its cost per order was roughly $30, resulting in significant losses. Jordan's "leap of faith" was not about whether people wanted grocery delivery, but whether founder Apoorva Mehta could make the unit economics work. Mehta presented a detailed list of 19 improvements, including securing better pricing agreements with grocers, achieving price parity, and building an advertising business. Jordan said: "He listed these 19 things, said he had already completed half of them... and basically achieved them all later." Today, Instacart's advertising business generates over $1 billion in annual revenue, and high-margin ad revenue subsidizes delivery costs, preventing consumers from bearing the full delivery expense.
Regarding the "leaky bucket" problem, Jordan cites analysis from partner Olivia Moore: If the user retention curve never bottoms out and keeps declining, it means the platform constantly needs "fresh blood" to grow, and will eventually hit a growth ceiling. This is one of the most dangerous signals.
Jordan also shared an important lesson: "Whenever I thought I was getting a bargain, I should have run — that was a sign of no heat. And every time I was forced to pay a high price, those companies have performed well so far." Instacart's Series B round is a typical example: Jordan initially offered $90 million (about 100x GMV), but Apoorva told him that was less than one-third of other term sheets, and a decision was needed by the next day. Jordan ultimately raised his offer to $300–400 million. He said: "Many of the best deals come with this kind of pricing pressure — the price is set by the market, not by the metrics."
At the core of Jordan's investment framework is the "defensiveness theory": he invests almost exclusively in companies with network effect potential, yet DoorDash stands as his biggest mistake — a case where he should have invested but missed out due to excessive concern over competition.
Jordan admits that his investment career has been "almost a one-trick pony": seeking out companies with network effects, typically realized through digital marketplace dynamics. He once tried investing in D2C (direct-to-consumer) companies but found that such businesses lack defensiveness — when six "mattress-in-a-box" companies appear on the market, competition quickly erodes profits. Thus, he stopped investing in D2C seven or eight years ago.
DoorDash is the exception to this rule. Jordan recalls knowing Tony Xu as early as when Tony was still at Stanford Business School (at the time, Tony was personally delivering meals), but when he examined the food delivery market, he saw 15 competitors including Uber, Caviar, and OrderAhead, and concluded that this market could not be profitable — "I foresaw a future where you go to a restaurant, and they would have 15 iPads or printers taking orders simultaneously." In the end, however, Uber's capital retreated, Tony executed brilliantly, and DoorDash emerged as the winner. Jordan says: "For me, this is the biggest exception to this rule. Kudos to Tony."
Jordan also shared the lesson of how eBay was "hollowed out." He believes eBay faced two major threats: first, being picked apart by vertical marketplaces one by one (such as Goat and StockX in sneakers, StubHub in ticketing); second, being overtaken by competitors from the flanks — Google became the entry point for product searches, with sellers preferring to transact through Google because it allowed them to own the customer relationship; Amazon replicated the half.com model and launched its marketplace. Jordan says: "You don't see the people chasing you in the rearview mirror; you have to look at the side mirrors — they come at you from the side."
Jordan introduces the concept of "layers of growth": marketplace growth stems from continuous product improvements — eliminating friction, adding use cases, and expanding functionality, with each layer accelerating growth.
Using eBay as a case study, Jordan breaks down the growth layers in detail:
1. Fixed-price format: Initially, eBay was 100% auction-based, and the community feared that introducing fixed prices would "kill the golden goose." However, Jordan's team allowed both formats to coexist, serving different user groups and product types, which significantly accelerated growth.
2. Storefronts: Providing sellers with a more structured way to display their offerings.
3. Payments: This was the most critical layer. Jordan describes eBay's early transaction process — after winning an auction, the buyer negotiated shipping costs with the seller, then went to a bank or post office to buy a money order and mail it to the seller, who then went to the post office to ship the item after receiving payment. The entire process took two weeks. PayPal reduced the payment time to two minutes, completely eliminating friction. Jordan states: "Every layer of product improvement accelerated growth."
Jordan emphasizes that the core driver of growth is product improvement, not paid customer acquisition. In eBay's early days, growth came from observing the community's spontaneous behaviors and amplifying them — for example, discovering that people were selling real cars led to the creation of eBay Motors. Airbnb is doing something similar, adding a new layer to travel through its "Experiences" business.
Jordan believes that the most successful marketplace platform founders possess two key traits: a lifelong learning mindset and a certain "superpower" – often overdeveloped, with shortcomings in other areas, which they compensate for with their teams.
Jordan notes that consumer-oriented founders are typically very young, digital natives who can spot opportunities overlooked by older industry veterans. He cites several founder examples:
Jordan emphasizes that all successful founders go through a phase of "walking in the desert." Pinterest's Ben Silbermann tried 49 approaches before the 50th worked; Airbnb sold Obama and McCain-themed cereal boxes to stay afloat. Jordan said: "I give these entrepreneurs incredible credit – they walk in the desert, drilling holes to find water. Sometimes they find it."
Jordan compares a16z’s model to CAA’s disruption of Hollywood: shifting from a loose partnership of individuals to an institutionalized model centered on a platform of services.
Jordan explains that traditional venture capital operates as a "loose partnership of individuals," where each partner maintains their own client relationships and business. Mark Andreessen and Ben Horowitz drew inspiration from Michael Ovitz’s transformation of CAA: when a client walked into a CAA meeting room, a dozen experts were present—covering domestic distribution, international distribution, publishing, music, and other fields. Similarly, a16z systematically built a network to support founders.
The core of a16z’s culture is "We Win as a Team." Jordan states: "Each of us has our own superpowers. Andrew excels at growth, Alex knows fintech inside out, Ben is strong in management, and I am better at marketplace platforms. But we win as a team." This means any partner is willing to help any entrepreneur.
a16z currently has over 350 employees, which Jordan says is more than the combined total of its top ten U.S. competitors. These investments are funded by the partners’ management fees—other firms would treat these as direct income, but a16z reinvests them into the platform that serves founders. Jordan notes: "We have aligned interests among LPs, portfolio companies, and ourselves—if the portfolio companies do well, we all benefit."
Regarding partner hiring, Jordan shares a key lesson: Initially, they tended to hire people with long operational experience, but this led to two problems—first, those with extensive operational backgrounds were often older, and their peers in their networks were no longer starting companies; second, this almost inevitably led to a homogeneous partner group (white males). Therefore, they relaxed the requirements and recruited younger, more centrally networked, and more diverse talent.
| Position | Guest Stance | Key Data |
|---|---|---|
| Airbnb | Bullish (invested and serves on board) | Supply side long-term constrained; Brian Chesky spent five years refining brand positioning |
| Instacart | Bullish (invested and serves on board) | At investment time, revenue per order was $12, cost per order was $30; advertising business generates over $1 billion in annual revenue |
| Bullish (invested and serves on board) | Ben Silbermann tried 49 different approaches before finding product-market fit | |
| OpenTable | Bullish (former CEO) | Spent zero on demand acquisition; acquired customers through restaurant website widgets |
| eBay | Bullish (former General Manager) | 100% auctions → introduced fixed prices; after introducing payments, transaction time shortened from two weeks to two minutes |
| DoorDash | Not invested (biggest miss) | Market once had 15 competitors; Tony Xu personally delivered meals |
| PayPal | Bullish (former President) | Defeated eBay's internal payment product, Billpoint |
| Fandango | Neutral (as a cautionary case) | Relied on 5-6 theater chains; any one exiting would weaken the platform |
| Goat / StockX | Bullish (as a vertical marketplace case) | "Hollowed out" a large business from eBay in the sneaker vertical |
| StubHub | Bullish (as a vertical marketplace case) | "Hollowed out" a large business from eBay in the ticketing vertical |
| Incredible Health | Bullish (a16z investment) | Nurse recruitment vertical marketplace platform |
| Belong | Bullish (a16z investment) | Rental housing marketplace platform |
| Thumbtack | Neutral (under observation) | Home services sector; has not yet built a "definitive" marketplace platform |
1. Jordan on the two key characteristics of marketplace investments: Supply must be fragmented (each supplier holds, on average, only one unit), and the platform must be able to create transactions that would not have occurred otherwise (lead generation). Airbnb satisfies both conditions.
2. Jordan on the paradox of "bargain hunting": "Whenever I see a bargain, I should run — that's a signal of no heat. Every time I was forced to pay a high price, those companies have performed well to this day." Instacart's Series B round is a classic example — Jordan initially offered $90 million but ultimately closed the deal at $300–400 million.
3. Jordan's warning on the "leaky bucket" problem: If the user retention curve never bottoms out and keeps declining, the platform needs "fresh blood" to grow, eventually hitting a ceiling. This is the most dangerous signal.
4. Jordan on the exception to his defensibility theory: DoorDash was his biggest mistake. He passed on the investment after seeing 15 competitors, but Tony Xu executed brilliantly, and after market consolidation, DoorDash emerged as the winner.
5. Jordan's framework on layers of growth: Marketplace growth comes from continuous product improvements — eliminating friction, adding use cases, and expanding features. eBay moved from auctions to fixed prices to payments, with each layer accelerating growth.
6. Jordan's observation on founders' "superpowers": The most successful founders all have some overdeveloped superpower, along with other weaknesses, but they compensate with their teams. Brian Chesky's superpower is brand and communication; Ben Silbermann's is continuous learning and relationship networks.
7. Jordan's analogy for the a16z model: Drawing from CAA's disruption of Hollywood — shifting from loose individual partnerships to an institutionalized model centered on platform services. a16z has over 350 employees, more than the combined total of its top ten competitors.
8. Jordan's principle of "always bias toward the buyer": Although the seller is usually the paying party, marketplaces should always optimize the buyer experience — because sellers come to the platform for the largest pool of buyers. Both eBay and OpenTable faced conflicts with sellers over this, but it ultimately proved to be the right choice.