Jeff Jordan - Building & Investing in Marketplaces - [Invest Like the Best, EP. 276]
What this covers
Patrick O'Shaughnessy talks with Jeff Jordan, a general partner at Andreessen Horowitz who earlier ran eBay.com, served as president of PayPal, and was CEO of OpenTable. Marketplaces are the through-line, and the conversation alternates between his operating years and his investing. At eBay he describes an economy that behaved like perfect competition: sellers who found outsized yields from scarcity were copied until prices regressed. Growth came from layered product enhancements such as fixed-price listings, stores, and electronic payments, and from amplifying behavior users had already started, as with eBay Motors. He also covers how Google pulled traffic away from eBay and how specialized vertical marketplaces can hollow out a horizontal one.
The investing half turns these experiences into screens. Jordan favors fragmented supply, contrasting OpenTable's individual restaurants with Fandango's handful of theater chains, and demand generated by distribution hacks like OpenTable's embeddable widget rather than paid acquisition. He names cohort retention curves that never flatten as a red flag. He is skeptical of companies that hit $100 million in GMV quickly but have no defensibility. Airbnb, which he first thought the stupidest idea he had heard, is the case for supply motivated by non-economic satisfactions. On Instacart he explains betting on Apoorva Mehta's plan to fix roughly $12 of revenue against $30 of cost per order, and paying up when the round's check tripled from about $90 million to $300 million. Bargains, he says, signal missing demand.
The last stretch covers Andreessen Horowitz's platform model, borrowed from CAA, and its changed approach to hiring partners. It also covers founders' "superpowers," with Brian Chesky and Ben Silbermann as examples, and why CEOs should be candid with investors about problems.
Jordan argues that marketplace success depends on defensibility, mainly network effects, and that the best marketplaces grow by fragmenting supply, generating demand without paid acquisition, and layering product improvements, while investors should avoid pricing that lacks a defensible theory.
- Marketplaces need a network effect or other defensibility, since growth charts alone do not show whether competitors can copy the model
- The best marketplaces acquire demand through product hacks and distribution rather than paid marketing, and fragmented supply gives the platform leverage
- Operators grow marketplaces through layered product enhancements that remove friction, and must balance buyer and seller interests
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
If a marketplace's cohort retention curve never bottoms out and keeps declining, the business is dependent on constantly acquiring fresh users, which is a red flag because it will hit a growth constraint, whereas cohorts that stabilize signal a bullish, sticky base.
“If the cohort curve never bottoms out and just keeps going down, that's just plain smashing it at the ground. I get very worried on that metric. If you just need fresh meat to grow, you'll reach a growth constraint at some point.”
In a two-sided marketplace, when an entrepreneur finds a supply or demand inefficiency and earns outsized yields from scarcity, other sellers observe this, fill the gap, and prices regress to their rational level, so the platform's job is to keep both sides in equilibrium.
“They were chewing off the very top of the price curve and then other sellers would observe that activity, fill the gap and the pricing then would regress to where it rationally should be.”
Fandango's dependence on five or six theater chains gives any one chain market power, because a chain that pulls out can send its customers directly to its own website instead of staying on the platform.
“Fandango basically has deals with the five or six major theater chains and any one of them can have market power because if AMC pulls out of Fandango, I am motivated to go to amc.com”
Early eBay initiatives designed in-house performed worse than watching what the community was already doing and amplifying it, so the best growth action was to identify nascent user behavior and build product around it, as with eBay Motors after sellers began listing cars.
“we found out the leverage was way more to watch what the nascent behavior the community was doing and seek to amplify it.”
Many successful consumer companies, including Pinterest and Airbnb, flailed for product-market fit, and their success came from trying many experiments, like Airbnb selling cereal boxes of political conventions, with persistence eventually surfacing the working idea.
“On a ton of them, they had to flail for product market fit. Pinterest was clearly one of those. Airbnb was clearly one of those.”
A marketplace is more valuable when its supply is highly fragmented, because individual suppliers like single-restaurant owners have no bargaining power and are hard to aggregate, whereas concentrated suppliers like Fandango's five or six theater chains can exert market power over the platform.
“the average restaurant owner on OpenTable owns one restaurant. And so aggregating them is a pain in the ass. But once you've aggregated them, it's a very valuable thing.”
The strongest marketplace models rely little on paid acquisition and instead find distribution hacks that generate user demand at scale, such as OpenTable's embeddable reservation widget for restaurant websites, which routed diners to OpenTable and made them return to it rather than Google.
“The best models are ones that don't really rely much on paid acquisition. The best entrepreneurs have figured out hacks to get user demand at scale through a user proposition.”
In a marketplace with buyers and sellers, a robust buyer base attracts sellers, so growth requires optimizing for buyers even when sellers are the ones paying, as eBay showed when sellers fled to Yahoo and Amazon's free auction listings lacked buyers and returned.
“why the sellers came is we had a robust buyer base, and so then growing the business requires optimizing the buyers' base.”
A horizontal marketplace like eBay can be hollowed out vertical by vertical by specialized marketplaces, because consumers prefer tailored experiences for categories like tickets or collectibles, so the platform must tailor enough per vertical to remain useful while keeping one coherent user experience.
“One threat was we get hollowed out vertical by vertical by specialized marketplaces... So we'd spend a lot of time trying to figure out, 'Okay, how do I go part way?'”
Airbnb has been supply constrained since Jeff became involved, and even though he initially thought the concept was the stupidest idea he had heard, the supply side was driven by people who enjoyed meeting strangers, showing that supply can be motivated by non-economic satisfactions.
“Airbnb's been supply constrained almost since I got involved in the company.”
In eBay's early years, the marketplace behaved like a perfect, self-contained economy in which microeconomic theory played out predictably, so pricing and supply-acquisition decisions could be made by applying economic theory and the results matched predictions.
“It is a perfect, at least it was a perfect economy where it was self-contained and all of the things in economic theory of microeconomics were playing out on eBay.”
When Jordan sees a bargain in a funding round, he should walk away, because bargains signal a lack of market demand, and his bargain deals consistently disappointed while the companies he had to pay up for performed well.
“One was whenever I saw a bargain, I should run. It's a sign of no heat. Whenever I did a bargain, I regretted it later. Whenever I was forced to pay up, to date that has been a very good basket of companies.”
The best deals often involve pricing pressure set by the market rather than by metrics, so an investor must decide whether to climb the price ladder, which Jordan did for Instacart when Apoorva Mehta insisted that his allocation be under a third of other term sheets, forcing the check to triple from about $90 million to $300 million.
“A lot of the very best deals have that kind of pricing pressure, and the pricing's set by the market. It's not set by metrics. So you have to figure out, 'Okay, do you climb?' And I tend to climb if I think it's legitimate heat.”
Airbnb has been supply-constrained for most of Jordan's involvement, and the supply side is harder for unconventional marketplaces because the concept of strangers staying in homes seemed counterintuitive, so supply depends on evangelists who value human connection.
“When it's that counterintuitive, the supply side, evangelical people to kind of say, 'I see it, and I enjoy it.'”
Instacart initially earned about $12 per order against roughly $30 in cost, but Apoorva Mehta's plan to secure grocer deals with better pricing and revenue sharing, combined with a high-margin advertising business, made the unit economics work, so the investment bet was on his ability to execute rather than on demand for grocery delivery.
“he was earning something like $12 a order in money to Instacart, and he was spending about $30.”
Google became an alternative to buying on eBay by helping users find anything, and sellers preferred Google transactions because they owned the customer relationship and could remarket directly, so Google drew traffic away from eBay even though eBay was cheaper per transaction.
“And the sellers actually liked having a Google transaction more than an eBay transaction because they owned the user.”
Andreessen Horowitz borrowed from Michael Ovitz's CAA model, which replaced individual partners with a firm-level set of experts, by offering founders operating guidance and networks, since founders like Zuckerberg arrived without networks and needed introductions.
“Marc and men kind of applied that logic early on and said, 'Okay, we're entrepreneurs, what do we lack?' We lack some guidance and definitely lack networks.”
Andreessen Horowitz, with over 350 employees, invests more in its platform than its ten biggest US competitors combined, funding that investment with partner fees that other firms would take as income, creating an alignment in which LPs, companies, and the firm all do well when companies do well.
“I think we're now a little over 350 employees at Andreesen Horowitz. That's probably more than our 10 biggest competitors in the US combined.”
Andreessen Horowitz relaxed its requirement for long operating experience when hiring general partners because old operators' networks are dominated by peers who no longer start companies, and the old rule skewed toward white men who had opportunities two decades earlier, so younger investors embedded in current networks are more useful and more diverse.
“One is it's a young person's business and that really comes through particularly in networks... my network is of my peers and they're not starting businesses anymore.”
Companies that reach $100 million in GMV in a few minutes, such as mattress-in-a-box, jewelry-finding, and meal-kit companies, look attractive on growth charts, but without defensibility the question of six or more competitors doing the same thing undermines the returns.
“Companies go $100 million in about eight minutes. Tiles to find things, $100 million in seven minutes. Meal kits, $100 million in six minutes... But then you say, 'What happens when there's six mattress in a box companies and six...'”
Successful founders each have a superpower that is often overdeveloped alongside underdeveloped areas, and the best ones lean into their superpower while supplementing gaps with their team, as Brian Chesky did with brand communication that eventually produced the 'Live like a local' positioning.
“Now, they're often the superpowers are overdeveloped and the non-superpowers are underdeveloped... The ones who really figure it out lean into their superpower and then supplement them with their team.”
The worst thing a CEO can do is withhold real problems and concerns from investors, and treating investors like owners means candidly sharing both successes and struggles, including unresolved issues like international expansion, which Jordan argues signals opportunity rather than weakness.
“The worst thing for me is when the operator's not talking to you about their real problems and their real concerns. ... Treating the investor like an owner for me is a philosophy I try to use on both sides of the table.”
Perpetual learning is a key attribute of founders who can both start and scale a company, and Jordan assesses it by asking whether founders recognize their own gaps, because founders who seek out expertise, like Ben Silbermann taking 51 meetings at a conference, tend to succeed.
“I had 51 meetings with different people at the conference... Everyone there has some special power and I wanted to know what they are.”
eBay grew by successively layering product enhancements, including a fixed-price format alongside auctions, stores, and electronic payments, with each layer removing friction or adding use cases, so growth came from continuous product innovation rather than a single launch.
“I quickly became convinced early on at eBay that the key way to grow these businesses was product enhancements. Giving users additional use cases, functionality, take away friction, do whatever you can.”
Before PayPal, eBay buyers had to negotiate shipping, mail money orders, and wait weeks for shipment, and adding electronic payments that enabled payment in two minutes made transactions far easier, which accelerated the eBay marketplace and was more important than the payments business itself.
“Layering in electronic payments, and the ability to pay in two minutes took the transaction from hellishly hard to pretty darn easy.”
The two features Jeff screens for first in a new marketplace are supply fragmentation, where each supplier is small and aggregation is hard but valuable once achieved, and lead generation, where the platform creates customer relationships that otherwise would not exist.
“Two main ones. One is fragmentation of the marketplace.”
Relationship-based marketplaces where existing service providers pay a little for convenience do poorly, whereas marketplaces that create new relationships are attractive because providers will pay heavily to acquire new customers, so lead generation is a key second feature to look for.
“Those don't do well because the service provider, they'll pay a little bit for convenience. They'll pay a whole lot for a new customer.”
Direct-to-consumer categories like mattress-in-a-box, tiles, and meal kits produced companies that reached $100 million in revenue in minutes, but without defensibility, many competitors could enter, so Jordan stopped investing in direct-to-consumer about seven or eight years earlier.
“Companies go $100 million in about eight minutes... But then you say, 'What happens when there's six mattress in a box companies and six...'”
Jordan's investment approach rests on requiring a clear theory of defensibility, which for consumer businesses has typically meant a network effect, because high growth alone does not indicate whether a company can resist copycats.
“I think it's critical you have a theory and defensibility. And for me, that defensibility typically has been network effect.”
Rentals, blue-collar jobs, and home services remain largely on Craigslist or fragmented sites, because Craigslist's network effect has not been broken and no definitive marketplace has yet been built in those segments, though someone eventually will.
“There are three of the key remaining businesses on Craigslist. One is rentals... Craigslist has a network effect and that one hasn't ripped up.”
Three segments still on Craigslist are ripe for marketplaces: rentals, where listings are often already rented and there is no information; blue-collar jobs, where hiring managers bear all the effort; and home services, where Thumbtack is attempting to flip from lead gen to marketplace.
“There are three of the key remaining businesses on Craigslist. One is rentals.”
As a hyper-growth business scales, the operator's role shifts from being the player making every call to being a coach who hires people to decide similarly, then a general manager, and finally a commissioner who orchestrates rather than acts, so the operator becomes progressively further from the action.
“And you become a coach. You hire a bunch of people... They become the coach and you become a general manager and you're further and further from the action in the field each time.”
OpenTable's internal research showed that when a diner opened a review, they were twice as likely to make a reservation, which justified publishing reviews even though restaurants resisted, with an opt-out offered as a fallback.
“We did research. If a customer opened a review, they were twice as likely to make a reservation.”
A high-margin advertising business, which Amazon has shown to be very profitable, lets a marketplace like Instacart subsidize delivery fees for consumers, making the consumer value proposition compelling without charging the full fare.
“the advertising business means he can provide a very compelling value prop to the consumer because they don't have to pay the full fare for the delivery.”
Horizontal platforms like LinkedIn fail for hiring nurses because licensing, qualifications, and experience are central, so vertical marketplaces like Incredible Health, which tailor hiring to specific professions, can succeed where horizontal ones cannot.
“LinkedIn does not work for nurses. ... So incredible health is a vertical play trying to do healthcare well, healthcare hiring well and rip it out.”
Jordan built a twice-weekly Stanford basketball community with the rules 'no assholes' and 'don't get hurt', and this convening created a mentoring network where senior members help junior players get their first jobs and grow careers.
“Number one rule is no assholes. Number two rule is we don't want to get hurt... There's a lot of mentoring going on.”
When eBay introduced a fixed-price format alongside auctions, many in the community feared it would kill the golden goose, but making both formats coexist for different audiences and goods pushed growth dramatically once fixed price was right.
“that accelerated growth pretty dramatically when we got fixed price right.”
Jeff tends to raise his offer when he sees legitimate market heat, since many of the best deals have pricing pressure set by the market rather than by metrics, so investors have to decide whether to climb.
“A lot of the very best deals have that kind of pricing pressure, and the pricing's set by the market. It's not set by metrics.”
In a risky IPO, placing the offering in a concentrated group of large, long-holding institutional investors creates a stable shareholder base that lets management focus on running the business rather than managing the stock.
“We ended up highly concentrating the offering into a handful of the best internet investors... They bought big. They wanted it in and they held for a good chunk of time, which makes very stable investor base and let me focus on running the business.”
Jordan's biggest investing miss was passing on DoorDash, which broke his defensibility rule because the food delivery market had many competitors, yet the company won because Tony Xu executed well and competitors eventually dried up.
“The one it didn't, ironically, was DoorDash. That's probably my biggest miss.”
While Jordan ran OpenTable, the company spent no money on demand acquisition and still grew well, relying entirely on its restaurant-website widget distribution.
“While I was there, we didn't spend a penny on demand acquisition and we're growing very nicely based on that.”
The consumer founders who break out are often young digital natives who notice things about the digital world that more experienced workers miss, and many had almost no job experience before starting their companies.
“they're digital natives, and they've made observations about the state of the digital world that typically more senior workers with more experience don't make.”
Airbnb generates lead gen because hosts are introduced to guests they would never have known, and the host base is spectacularly fragmented, with the average host owning one property.
“Hosts are being introduced to guests they never would've known. It's spectacularly fragmented. The average host owns one property.”
When Jeff realized that a 7 p.m. line of decisions still awaited him after an early start, he concluded that hyper-growth was not scaling and that he had to hand off decisions to build a team.
“there's still a line outside my door waiting for me to make decisions. I go, 'This is not scaling. I got to change something.'”
OpenTable's IPO came in May 2009 at the depth of the financial crisis, when the NASDAQ was off about 70 percent, and it was the first venture-backed tech company to go public in about two years.
“My first financing was the OpenTable IPO at the depth of the financial crisis in May 2009.”
Moving from division head to public-company CEO dials up the level of worry and final quality control, so although OpenTable was much smaller than PayPal, Jordan was more stressed because the buck stopped with him rather than with Meg Whitman.
“OpenTable is way smaller than PayPal, but I was more stressed because it's on me. Before it was on Meg, now it's on me.”
Jeff feels obligated to pay forward the mentorship he received by investing time in people who impress him, but he has not figured out how to scale this beyond one-on-one relationships.
“How do you do it at scale? How do you have the bigger impact and doing it one at a time?”
Jordan says Sheryl Sandberg told him that Facebook's main early competitor was eBay, not a search engine or Yahoo, because of the commerce angle, which surprised him.
“And she goes, 'Oh, we knew early on that our main competitor was eBay.' And I'm just sitting there like, 'Holy cow.'”
Mark Zuckerberg's early experience illustrated how founders with no business-running experience may lack networks and operating knowledge that former operators provide, which is the gap a16z's model was built to fill.
“they had old operating people like me as the early founders.”
Jordan's first job, cooking at a restaurant at 15 after his father died, taught him valuable operating skills, as a chef who worked 13 shifts a week trusted him to take on more responsibility until he became weekend brunch chef and later an award-winning chef.
“Within a couple of months, I was brunch chef... I ended up being an award winning chef. We won best brunch in Philadelphia.”
Specialized vertical marketplaces such as StockX in sneakers and StubHub in tickets built large businesses in competition with eBay, showing that a horizontal platform can lose vertical categories and that eBay is now trying to win back by authenticating sneakers and trading cards.
“in shoes, go in StockX. Done a heck of a job... StubHub was another that built a really big business in competition with eBay.”
Jordan's mother gave him credible ambition while his father's intelligence was not inherited, and he observes that several high-performing executives who were essentially orphans by their twenties develop strong drive because they have no backstop.
“I know a couple very high performing execs who were essentially orphans by the time they were in their 20s. You get drive, if you're on your own and there's no backstop, it's pretty good for providing motivation.”
Building a strong community requires finessing new members through a trial, such as playing a pickup game to assess character, and enforcing norms through direct conversations, because running businesses teaches leaders to have tough conversations.
“One former NFL player who has a pretty good temper scared the shit out of one of these 6'4, 200 pound guys. And so I get the very fun conversation to go up to said hothead and say, 'Dude, you can't do that.'”
In the Instacart round, Jeff was offered a late position with a weekend deadline, the founder said Jeff's check had to be a third of the others, and Jeff had to decide whether to triple his proposed amount.
“you've got to know you're less than a third of any of my other term sheets. And by the way, I'm deciding tomorrow.”
John Donahoe's back-channel referral led to Jeff's interview with Andreessen and Horowitz, and his first deal was Airbnb, which he suggested while the firm was raising, so his first investment came almost on arrival.
“So my first deal was partnering with Mark on Airbnb almost as I was walking in the door at Andreessen Horowitz.”
Half.com was a hybrid of eBay and Amazon listing items under UPC codes, and after eBay acquired it, Amazon built its own marketplace to knock off the model, which made Half.com a defeated competitor and exposed eBay to Amazon.
“I've been told by sources that Amazon, in building Marketplace, had called it the Half-killer. So eBay bought Half but then Amazon knocked it off.”
eBay grew as new use cases emerged beyond collectibles, such as overstock merchandise and hard-to-find Christmas gifts, and in some Christmas seasons Jordan says eBay sold 80 percent of the units Microsoft produced for a new PlayStation console.
“I need the new PlayStation console. We would sell 80% of the units that Microsoft produced would sell on eBay during a Christmas season.”
Benchmark's Bob Kagle invested in eBay after seeing collectibles that he had previously driven across Michigan to find suddenly appear on eBay, which illustrated how online marketplaces solve search friction for niche goods, and eBay never even cashed his check because it kept the money as security.
“He's used to driving all over Michigan trying to unearth a duck decoy and all of a sudden five pop up on eBay.”
Jordan met Tony Xu roughly a decade earlier when Stanford sent Palo Alto Eats to him, and he assessed the food-delivery market at the time as crowded with about 15 competitors, which led him to pass.
“Stanford Business School asked me to sit down with some of their entrepreneurs 10 years ago. And one they sent in was Tony Xu with Palo Alto Eat.”
Mark Zuckerberg contacted Jordan when Facebook had just relocated to California, and he said he had never run a business before and had never worked at one, which surprised Jordan given how much operating knowledge he assumed employees gain.
“He goes, 'Well, people who know me know that I've never run a business before.'”
Jordan left operating for venture capital because he was seeking a new, steeper learning curve, which he had already found enjoyable through angel investing and advising startups on the side while working at PayPal and OpenTable.
“It was seeking a new learning curve. In my spare time at PayPal and at OpenTable, I'd meet with entrepreneurs... I was enjoying that side hustle way more than I was enjoying my day job.”
When mentoring young people, the core objective is to get them started by helping them identify interests, skills, and networks, because once they land a first job they can progressively figure out the rest, so launching is the high-satisfaction outcome.
“If I can help them launch, very high satisfaction for them.”
After repeatedly moving to successively larger growth-stage companies, Jordan got good at that stage, but his learning curve flattened, and the combination of constant stress and serving on a board became toxic.
“The bad news is the learning curve just shallowed out like crazy. When I'm operating, I'm always on, always stressed, always tired.”
Andreessen Horowitz embeds a written culture into daily operations, with the principle that anyone at the firm will help any entrepreneur and that partners' complementary superpowers make the firm stronger than individuals.
“Anyone at the firm will help any entrepreneur at any time. We all have our individual superpowers.”
Jordan mentors people because several formative figures unusually invested in him after his parents died, and he finds the satisfaction high but has not figured out how to scale mentorship beyond one person at a time.
“The frustration is I haven't figured out how to scale it. It is still one at a time.”
The head of admissions at Stanford Business School called Jordan the next day after he said he could not afford $70,000 in debt, saying she could not sleep and that the school would figure it out, which he considers the kindest act someone has done for him.
“And she called me back the next day and goes, 'I couldn't sleep at night. You have to come. We'll figure it out. You'll be out of debt in two years.'”
Airbnb sold collectible cereal boxes, including Obama O's and Captain McCains, as a revenue source during its search for product-market fit, an example of the creativity it took to survive.
“selling serial boxes of Republican and Democratic Convention to make money called Obama O's and Captain McCains”
Craig Robinson used a basketball game with Barack Obama to assess character, and Jeff uses a similar approach to read people's character through their play.
“Craig Robinson played basketball with Barack Obama to try to assess his character when he was dating Michelle.”
Disney under Michael Eisner and Frank Wells attracted remarkable young talent, and a company's ten-year run was driven by roughly 30-year-old executives, which ran counter to Hollywood norms and taught Jordan that talent is the biggest business lesson.
“It's the amount of young talent that Michael Eisner and Frank Wells attracted to that company was just remarkable.”
eBay's acquired Billpoint payments product was run by someone who bypassed Jordan's management layer and reported directly to Meg Whitman, and this was ultimately fortunate because Billpoint became roadkill against PayPal, whose team executed far better.
“Billpoint became absolute roadkill to PayPal. The PayPal guys did a phenomenal job and the eBay people did not.”
Andrew Chen's book The Cold Start Problem addresses the hard side of a marketplace, reflecting that marketplaces are constantly challenged to stay in equilibrium.
“My partner, Andrew Chen just wrote a book called The Cold Start Problem.”
Patrick and Jeff agree that the key investor target is an innovation that removes friction and convenience barriers, and that growth comes from constantly anticipating the next friction point, a paranoid posture.
“Shows you how much real innovation there was. There wasn't an alternative.”
Jordan got the OpenTable CEO role when Benchmark's Bob Kagle suggested him after Thomas Layton said the company was ready to go public but did not want to run a public company, because Jordan understood network effects and restaurants from his chef background.
“Oh you should talk to Jeff. He knows network effects and he knows the restaurant business because he was a chef.”
Jordan's first investment at Andreessen Horowitz was Airbnb, which he pursued after seeing Brian Chesky present at the Allen conference and learning in an interview that the company was raising.
“I'd just seen Brian Chesky present Airbnb at Allen Conference, and I go, 'Airbnb.' And Mark looks at me quizzically and goes, 'You know they're raising, right?'”
Jeff's fragmentation theory was explained to a board member who turned out to be the founder of Fandango, which illustrates how concentrated suppliers can see through and exploit platform dependence.
“You do know I'm the founder of Fandango, right?”
Patrick notes that a fragmented supply base is straightforward to aggregate, which prompts the question of which attributes distinguish teams that do lead generation well.
“What about on that other side, what's shared in common amongst the most talented people that you've seen thinking through this problem of lead gen?”