What this covers

Patrick talks with Sarah Tavel, a Benchmark investor, about how to judge consumer products and marketplaces. She opens with a philosophical frame from Christine Korsgaard: like a friend, an early-stage investor adopts the founder's ends as their own. The rest is mostly worked cases. Grubhub defined its supply too narrowly, and Postmates treated delivery as a third side of the marketplace to onboard restaurants that didn't deliver. Airbnb found a different atomic unit of supply than Vrbo. Booking.com charged hotels roughly 10 to 15 percent against Expedia's 30. GOAT's authenticity guarantee solved a trust problem eBay had left open.

The range then widens. On the product side, Tavel discusses power users versus the "next 100 million," and how reputation systems such as Airbnb's Superhost badge and review loops improve the quality of liquidity. She covers vertical marketplace opportunities that LinkedIn's generic identity layer misses, such as RigUp and Instawork. She also looks at companies that make opaque information legible, such as REKKI, Zillow, Glassdoor, and OpenTable. A restaurant-economics aside explains why supply chain costs matter, down to ordering six lemons instead of five. She also describes phases of data use, from first-principles research to decisions made on data.

Several claims are contested. One is a Schibsted analysis tying marketplace profitability to size relative to the number two player. Another is the argument that public incumbents cannot match privately funded challengers who ignore contribution margin. A third is the reported Facebook finding that passive use leaves people feeling worse, which Tavel connects to paid, members-only social networks. She closes on founders who tie their ego to the wrong metric, and on the doubt that even apparently successful companies go through.

Sharpest takeaway

Tavel argues that marketplace and consumer winners are built by focusing narrowly on a constrained market, maximizing one core action that drives liquidity, and designing products whose value compounds, rather than by racing for growth or GMV, which can hide a lack of liquidity.

  • Incumbents like Grubhub were vulnerable because they defined supply too narrowly, and challengers won by redefining the atomic unit of supply and cost structure
  • Marketplaces tip only when they build liquidity and network effects in a constrained market, not when they simply accumulate GMV across many cities
  • Sticky consumer products have accruing benefits and mounting losses, and their growth depends on a single core action correlated with retention

The argument · threads3 threads · 64 claims
0.71

Vertical marketplaces can build professional identity and access for underserved or niche workforces.

2 pointscentrality 4/5
  • RigUp built rich profiles for its workers, effectively a LinkedIn for the trades, and Instawork does a similar job for people who lack a professional profile, showing vertical professional identity can be built for underserved workforces.

    “They created rich profiles actually for all of the workers. So it became their LinkedIn.”

  • Hipcamp is building a marketplace described as Airbnb for land, where landowners post land and campers, glampers, and lodge operators list accommodations, with specialized search such as lodging near hot springs.

    “people will describe it as Airbnb for land”

0.30

Marketplaces that set opaque information free form a distinct, non-transactional category.

2 pointscentrality 1/5
  • Rich Barton, a venture partner at Benchmark, is associated with a generation of marketplaces that are not necessarily transactional but instead set opaque information free, including Zillow, Glassdoor, and OpenTable.

    “Rich Barton from Zillow who was a venture partner with us at Benchmark talks about how there's a some generation of marketplaces”

  • Tavel studied Kant and was advised by Christine Korsgaard, and Korsgaard's essay Creating the Kingdom of Ends first drew her to philosophy after she had studied economics.

    “Christine Korsgaard was my thesis advisor. And she has this essay.”

0.29

Facebook's model fails young users because lean-back watching lacks active participation.

1 pointcentrality 2/5
  • Facebook's experience is unlikely to translate well to young people, which is why they have migrated to Snapchat, Discord, and gaming, because most incumbent social products are lean-back experiences of watching others rather than active participation.

    “the Facebook experience first of all isn't going to translate as well to young people”