Jeff Jordan on Building Two-Sided Marketplaces

Harry Stebbings asks why fragmented supply can make a marketplace attractive. Jeff Jordan contrasts movie-ticket service Fandango, which could aggregate a handful of major cinema chains, with OpenTable, whose target was tens of thousands of mostly independent restaurants.

A handful of large suppliers can each exercise power. Jordan recalls going directly to a cinema chain’s website when it was absent from Fandango. OpenTable had a much harder job signing restaurants individually, but an aggregated network was also harder for a competitor to reproduce, and no single restaurant had comparable leverage.

That advantage took a long time to build. OpenTable initially sold restaurants a productivity tool: sales were slow in each city until there was enough supply to attract diners. Once diners arrived and the service could bring restaurants revenue, restaurant sales accelerated.

The diners, not OpenTable, ultimately determined what counted as enough supply. Five available restaurants might not interest them; an assortment containing recognized names gave the service credibility. Jordan recalls a tipping point around 10% of a target market’s supply, when convenience could outweigh an incomplete selection; he presents that as experience, not a universal threshold.

Once there are many options, the platform can help users decide with filters such as neighborhood, cuisine and past favorites. Jordan calls too much supply a high-quality problem compared with the earlier challenge of gathering it.

Supply acquisition depends on the depth of the value proposition. Even successful marketplaces can be supply constrained, Jordan says; Airbnb, for example, kept seeking hosts to improve the experience. Some marketplaces must start with supply because demand has no reason to arrive without it.

Stebbings asks whether a marketplace should create or own supply itself. Jordan sees potential channel conflict: if Instacart opened its own grocery operation, it could compete with the grocers it sought to help. He contrasts that with Netflix’s move into creating content, an consequential decision that worked for that business.

The quality of leads matters as much as the scheduling tool. A hairdresser whose customers return regularly may resist paying a marketplace for access to people already coming to the salon. Restaurants, by contrast, benefit when diners seeking variety discover a new place through OpenTable; that incremental customer made the service more valuable.

Discovery channels keep changing. Jordan describes debates about short-form video and restaurant discovery, and argues that restaurants and marketplaces must watch emerging channels. His discussion reflects the landscape at the time of the interview, rather than a fixed prediction about any platform.

Jordan is wary of companies whose growth depends heavily on buying users from major advertising platforms. He says compelling concepts can grow without much paid marketing, but does not treat this as an absolute rule: he notes that TikTok itself bought users at scale.

The time to increase paid acquisition is when the unit economics are understood and attractive. Jordan recalls his early enthusiasm for e-commerce, followed by the realization that similar products could face many competitors bidding for the same customers. Acquisition costs rose while price competition limited returns.

A heavily funded competitor can force a difficult choice. Jordan says the answer depends partly on the capital environment: at the time, he was advising companies to extend runway without damaging the business, whereas a few years earlier abundant capital had encouraged much faster spending.

He offers Airbnb as an example of a more capital-efficient marketplace. Guests paid in advance, giving the company cash before the stay; Jordan recalls that it became cash-flow positive early, before operating profitability. The timing of cash flows can help, although it does not by itself establish profitability.

Incredible Health, he says, had a similar feature: hospitals paid in advance for nurse hiring. After its earlier financing, the company went years without another raise and had more cash when it next raised, according to Jordan. He values this kind of advance payment because it can reduce the need for external capital.

Jordan recalls that Incredible Health’s founder chose to raise a subsequent round even though the business had momentum and strong cash generation. At the valuation available, he says, the safety of additional cash outweighed the extra dilution. His point is about that financing decision, not a general prescription to raise or avoid funding.

Stebbings turns to demand-side messaging. Jordan says he learned by watching Brian Chesky and the Airbnb founders repeatedly refine the essence of the brand. They had to make the unfamiliar idea of staying in a stranger’s home feel attractive and trustworthy, emphasizing community and belonging.

A founder can begin with a very specific audience even when the eventual opportunity is broad. Jordan says his firm tries not to decide investments solely on the size of the existing market, because some successful companies create markets as they grow.

When Jordan joined eBay, he saw an early plan focused on collectibles. Users found the marketplace useful and began listing other kinds of goods. Rather than assume the initial category defined the ceiling, the team watched what people did and made new categories easier to discover.

He recalls early employee Simon Rothman searching for a Lamborghini and finding real cars listed on eBay, an early signal behind eBay Motors. The same approach helped the business support other categories: observe unexpected use and build around it.

Market timing is another risk. Jordan compares Webvan and Instacart: both addressed online grocery shopping, but their execution and the technologies available differed. He argues that mobile phones made the coordination of shoppers, picking and delivery far more practical for Instacart.

Stebbings asks how investors stay open-minded after failures. Jordan says frequent failure is humbling, and advocates examining the quality of a decision rather than treating the outcome alone as proof. The point is to identify what was understood correctly, what was missed and what can change in a future decision.

His experience with direct-to-consumer e-commerce strengthened his attention to lifetime value relative to customer acquisition cost. Even investments he considers among the better companies produced disappointing results in his telling, prompting him to reduce his enthusiasm for businesses reliant on heavy paid acquisition.

Early marketplace demand requires experimentation on both sides. Jordan likes starting with a narrow group that feels strongly about the product and serving it well. He recalls Airbnb initially appealing to younger travelers seeking community and a cheaper alternative before expanding its use cases.

When examining cohorts, Jordan looks at the rate at which usage declines and whether newer cohorts improve. He recalls an Instacart example in which roughly three-quarters of users had dropped out between the first and tenth purchase, but those remaining spent three to four times as much, leaving cohort revenue close to its initial level.

Improving cohorts can be consistent with a network effect: as more people use a marketplace, it may become more valuable. Stebbings also raises the opposite problem, where demand overwhelms supply and delivery quality deteriorates. Jordan notes that quick delivery needs both reliable service and viable economics.

Jordan says his firm did not invest in the fast-commerce sector, partly because internet veterans remembered earlier attempts such as Kozmo. Scale alone was not enough for him; the team needed to see what had changed and how the business could eventually work economically.

At the time of the Instacart investment, Jordan recalls the company billing roughly $10–12 per order while losing $20–25 on each. He asked founder Apoorva Mehta how the business could close the gap. Mehta, a former Amazon fulfillment engineer, laid out a path of continuous operational improvements.

The proposed improvements included advertising for consumer packaged-goods brands, agreements with grocers, scale and better fulfillment. Early shoppers waited in ordinary checkout lines; later, dedicated lanes and arrangements allowing shoppers to scan products as they picked could remove minutes from an order.

Jordan says those incremental savings mattered because time was a major cost. He describes Instacart as having become significantly profitable by the time of the conversation; that is his historical assessment, not a claim about the company’s current finances.

Acquisition costs are volatile, but Jordan’s experience is that they often rise as a company scales and competes for the same audience. He therefore wants a substantial margin between lifetime value and acquisition cost. His examples of ratios express his judgment and tolerance for erosion, not a universal cutoff.

Privacy changes, regulation and new advertising platforms also make channel economics uncertain. Jordan likes keeping an experimental budget separate from the core acquisition budget: a team can continue investing in a working channel while testing alternatives that may reduce dependence over time.

Balancing supply and demand is the central marketplace task. Jordan sees demand as especially strategic because suppliers tend to follow it: a host who gets substantial bookings has reason to stay, while one who receives only a couple of bookings may not.

Home services illustrates the difficulty of creating a marketplace. Stebbings points to leakage when customers and providers move their relationship off-platform. Jordan agrees and adds that many services are infrequent, making it hard for a marketplace brand to become memorable.

Even so, Jordan sees an unmet need for a trusted way to find high-quality contractors, particularly for jobs people do not routinely need. He remains interested in the category while acknowledging why prior efforts have struggled.

To find founders, Jordan combines a thesis about promising markets with openness to entrepreneurs’ approaches. He describes publishing as a way to send a “bat signal”: a short post looking for businesses that combined convenience and sound economics prompted Instacart’s founder to contact him.