Jeff Jordan on Building and Investing in Two-Sided Marketplaces

Jeff Jordan's route into technology began after working at Disney, when the internet prompted him to join a retail company exploring online video. He later joined eBay, where he says building a network-effect business without an established playbook taught him how to recognize similar potential elsewhere.

Moving from operator to investor required a change in posture. Jordan likens the investor to a passenger: the founder must steer. He shares what happened when he faced similar problems but does not want founders to follow his advice simply because he gave it.

That restraint does not mean withholding difficult feedback. If an investor sees a serious risk, Jordan believes they owe the founder an honest, empathetic conversation, while recognizing that the investor's judgment may be wrong.

One marketplace trait he values is fragmented supply. He contrasts a movie-ticket market with a handful of large chains, each with bargaining power, with OpenTable's many individual restaurants: gathering them is difficult, but an established aggregation can be harder to replicate.

OpenTable's early city-by-city build was slow. Its restaurant productivity tool helped it get started; once enough diners found a useful selection and restaurants gained customers, the network became more attractive to both sides.

Jordan recalls a tipping point at roughly 10% of target local supply, when enough diners valued the convenience despite incomplete selection. This is his retrospective estimate for OpenTable, not a universal marketplace threshold.

When selection becomes large, filters by neighborhood, cuisine and previous favorites can help customers decide. The bigger early challenge is often a strong enough supplier value proposition: demand is unlikely to arrive without useful supply, yet suppliers need a reason to join.

A marketplace creating its own supply can face channel conflict. Jordan says Instacart's partnership with grocers would be strained if it operated competing grocery warehouses; he speaks from his position as an investor and board member, not as an independent observer.

He also values marketplaces that deliver genuinely new customers. A hair salon may not want to pay a commission for an existing regular who could book through a free tool, whereas restaurants can value introductions to diners seeking new places and cuisines.

Acquisition channels change continually. Jordan says founders should watch emerging discovery behavior and experiment, citing a colleague's view that short-form video could change restaurant discovery; he does not treat a particular platform's future as settled.

He is cautious about businesses that must buy most of their users indefinitely. Some successful services spend heavily on acquisition, but he prefers a proposition compelling enough to generate organic demand and recommends understanding unit economics before expanding paid spend.

He recalls investing in direct-to-consumer e-commerce as an example of a thesis only partly right: online share grew, but many similar sellers competed for the same customers through the same channels. In his view that competition raised acquisition costs and made durable profits difficult.

When a heavily funded competitor spends aggressively, the answer depends on the capital environment. In the period of the interview Jordan says he was advising companies to extend their runway without damaging the business, contrasting it with an earlier period of plentiful funding.

He highlights advance payment as one rare source of marketplace capital efficiency. He says Airbnb received booking cash before a stay and that Incredible Health received some hospital payments in advance; he has investment and board ties to both companies.

Jordan recalls Incredible Health going several years after its earlier round without another financing and later raising with more cash than it had immediately after the prior round. He presents its later fundraising as a choice to add safety at a high valuation, not evidence that all marketplaces can follow that path.

Turning to demand, Jordan praises Brian Chesky and Airbnb's team for refining their message. He argues that reassuring people about stays in strangers' homes requires a story that makes an initially unfamiliar experience attractive, including themes of community and belonging.

Stebbings asks whether an initially narrow audience makes a market look too small. Jordan says his firm tries not to base decisions only on current market size because strong products can create or expand markets.

He recalls eBay's original focus on collectibles. Users began listing other goods, including cars, and the company learned to follow and support those emerging uses rather than assuming its initial category defined the eventual opportunity.

Jordan says investors must take market-timing risk. He contrasts earlier online grocery attempts with Instacart's later approach, arguing that mobile phones helped coordinate shoppers and made the consumer proposition more practical; this is his interpretation, not a controlled comparison.

He tries to revisit decisions by separating decision quality from the final outcome. His experience with direct-to-consumer investments led him to place more weight on sustainable acquisition economics, while he says investors still need to remain open to changed circumstances.

Early demand building involves experimentation with both sides and identifying the users who value the product most. Jordan supports beginning with a narrow, enthusiastic community and expanding use cases once a team understands what resonates.

When inspecting cohorts, he looks at how sharply users decline and how newer cohorts compare with older ones. He says Instacart once lost roughly three quarters of users between an initial purchase and a later point around a year out, but remaining users spent three or four times more, leaving revenue for the cohort close to its original level; those are his recollections and specific metrics, not current company results.

Improving cohorts can be consistent with a network effect, Jordan says, though they need interpretation. Stebbings raises the opposite problem: more demand can overwhelm supply and worsen service, making growth itself a source of deterioration.

Jordan says delivery speed must work operationally and economically at sufficient scale. He recalls declining to invest in rapid-commerce models because some reminded him of an earlier failed delivery generation, in his view without a sufficiently new solution.

He describes Instacart's early unit economics as deeply negative: from memory it billed around $10 to $12 per order while losing around $20 to $25 per order. Rather than assuming scale alone would fix it, he asked founder Apoorva Mehta for a specific improvement roadmap.

That roadmap included advertising, commercial arrangements with grocers and operational improvements. Jordan recalls dedicated checkout arrangements and software that let shoppers scan baskets and pay stores directly, reducing the minutes spent on each order; his claim about later profitability reflects his account at the time.

On customer acquisition cost, he expects costs often to rise as a company scales or competes for the same audiences. A narrow margin between lifetime value and acquisition cost therefore concerns him; his numerical examples are judgment calls, not guaranteed benchmarks.

He expects channel conditions to remain volatile and recommends a separate experimental budget so that the need to optimize established channels does not prevent trying new ones. Stebbings cautions against diversifying before a working channel has been developed; Jordan agrees to keep using what works while testing alternatives.

Maintaining balance between supply and demand is the whole game, Jordan says. In his experience, healthy demand is especially important because suppliers stay where the activity produces meaningful value; an Airbnb listing with very few bookings may not justify the effort to its owner.

He remains interested in home-services marketplaces, even though previous attempts have struggled. He recognizes problems including infrequent transactions, difficulty remembering a brand and leakage when customers and providers take future transactions off-platform.

Jordan describes publishing investment theses as a signal to founders about areas he wants to explore. He recalls Instacart's founder contacting him after reading a post about combining convenience with workable economics, illustrating how an investor's public writing can attract relevant opportunities.

As a board member, Jordan tries to focus on issues where the board can add value, rather than designing the product in a meeting. Having been an operator, he aims to offer specific input with empathy and avoid speaking merely to take up time.

He urges founders to select board members carefully and assess shared values, because an investor may be part of the company for a decade. Both acknowledge that competitive financing timelines can make it hard to build that understanding before a decision.

Jordan identifies a basic alignment problem: venture investors hold portfolios while a founder has one company. An investor may prefer the higher-risk, higher-upside route; he believes that difference should be explicit when giving advice.

He advises new board members to observe experienced colleagues, including negative examples, and ask who adds value and how. A good discussion need not be dominated by the person who speaks most.

In quick-fire answers, Jordan recommends books about sport and race, notes both the opportunity and complexity created by a16z's expansion, and says the firm had grown to more than 400 people at the time. For the 2023 outlook he favors planning for worse and hoping for better.

He says investing requires accepting a lower apparent hit rate than operating and considerable time meeting companies that do not become investments. He declines to disclose his own results in the interview.

Jordan recalls passing on an early DoorDash investment because he believed food delivery lacked durable barriers and restaurants would work with many providers. He says he underestimated later industry consolidation and praises Tony Xu's execution; this is his retrospective assessment, not proof that the original decision process was necessarily correct.

His closing investment advice is to consider a future market's potential rather than only its present size. He also discusses a recently announced commercial real-estate marketplace investment as an example of flexible space; his interest in that company is financial.

Stebbings thanks Jordan for the conversation, and Jordan closes the interview.