Marketplace Assign vs. Marketplace Select Model
A mental model for predicting whether a two-sided marketplace becomes winner-take-all
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 5
- Confidence
- 75%
This mental model explains why some marketplaces (like Airbnb) consolidate into a single global winner while others (like ride-sharing) support multiple durable competitors, including locally. The mechanism: in 'marketplace-assign' models, supply is undifferentiated and the customer only needs fulfillment within a few minutes at a similar price, so density wins over brand loyalty; in 'marketplace-select' models, customers actively pick among differentiated supply options, which builds habits that carry across markets. Combined with whether the network effect is local (city-by-city) versus global, and whether supply and demand can freely multihome across competing apps, this model predicts market structure. It explains how Lyft survived being outraised 30-to-1 by Uber in 2014: ride-sharing's local, undifferentiated, multihomed nature made it structurally not winner-take-all.
Origin
Articulated by the hosts drawing on analysis from journalist Brad Stone comparing Airbnb's cross-market network effects to ride-sharing's local ones, applied to explain why Uber never fully eliminated Lyft despite a massive capital advantage.
Core principles
- 01In 'marketplace-assign' models, supply is undifferentiated and demand doesn't care who fulfills the request, so local density beats brand loyalty
- 02In 'marketplace-select' models, demand actively picks among differentiated supply, building cross-market habits and stronger network effects
- 03Local network effects don't compound across geographies the way global ones do
- 04When both supply and demand can multihome across competing platforms, no single player can achieve winner-take-all dominance
How to run it
- 1
Determine Supply Differentiation
Ask whether the customer cares which specific supplier fulfills the transaction, or only that it's fulfilled quickly and cheaply.
- 2
Map the Network Effect's Geography
Identify whether the value of the network compounds within a single local market or travels with the user across markets.
- 3
Test for Multihoming
Check whether suppliers and customers can easily and cheaply use multiple competing platforms at once without real switching costs.
Pro tip Look at actual driver and rider behavior data, not stated brand preference, to test multihoming.
- 4
Predict the Resulting Market Structure
Combine the signals: undifferentiated supply, local-only network effects, and easy multihoming together predict a fragmented, multi-player market rather than winner-take-all.
Watch out Don't extrapolate from adjacent markets that look similar on the surface — Airbnb and ride-sharing look alike but have opposite structures.
- 5
Set Capital Strategy Accordingly
In multi-player markets, competitors can survive by winning enough local density in enough markets to stay capitalized, rather than needing to dominate globally on day one.
In the wild
Airbnb's network effect is strong across markets — travelers who used it on one trip default to it on the next — making home-sharing winner-take-all. Ride-sharing's network effect, by contrast, is almost entirely local: the vast majority of usage happens in one city, and both riders and drivers freely multihome between Lyft and Uber.
→ Despite Uber outraising Lyft roughly 30-to-1 in 2014, Lyft survived and later grew from 20% to nearly 40% U.S. market share, confirming ride-sharing was not structurally winner-take-all.
Common mistakes
Assuming all marketplaces are winner-take-all
In 2014 the prevailing view in Silicon Valley was that Uber had already won and investing in any competitor was a fool's errand — a costly misreading of the market's true structure.
Ignoring multihoming when sizing competitive risk
Underestimating how easily both drivers and riders could use two apps at once caused investors to misjudge how durable a well-capitalized #2 player could be.
Is it for you?
Best for
Founders and investors evaluating competitive dynamics in two-sided marketplace startups.
Not ideal for
Single-sided products or markets without clear supply/demand matching dynamics.
From the transcript
“Ride sharing dwarfs the market for Airbnb. It's just not a winner-take-all market.”
“I don't care who picks me up, I just need to get there, versus marketplace assist, which is the platform provides a bunch of different…”
From the episode
The Lyft IPO