Increasing Returns Marketplace Test
Identify markets where each new participant makes the platform stronger
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 98%
The test starts by mapping every participant group and the value exchanged between them. A genuine increasing-returns business becomes more useful as participation rises: more supply attracts more demand, more demand motivates additional supply, and the reinforcing cycle improves the platform's position. The crucial question is not whether the initial operating model looks attractive in isolation, but whether the right pieces can cause the company to accelerate toward a winner-take-all or winner-take-most outcome. Founders and investors should verify that the loop changes customer behavior, creates reasons to converge on one network, and becomes harder for competitors to reproduce at scale. Once validated in one setting, the same structure can be used as a search lens for other industries with fragmented supply and demand.
Origin
Gurley derived this investment lens from Brian Arthur's work on increasing returns, encountered through Mitchell Waldrop's Complexity, and first applied it at Benchmark to OpenTable.
Core principles
- 01Participant growth can improve the product itself
- 02Supply attracts demand and demand attracts supply
- 03Strong network effects can produce winner-take-all outcomes
- 04A difficult initial business may become attractive at network scale
- 05The mechanism should transfer across industries
How to run it
- 1
Map the Sides
List the distinct participant groups and the transaction or interaction connecting them. Clarify what each side needs from the platform.
Pro tip Use behaviorally specific labels such as diners and restaurants rather than broad labels such as users.
Watch out A business with customers and vendors is not automatically a network-effects marketplace.
- 2
Trace the Reinforcing Loop
Explain how growth on one side increases value for the other and prompts further growth in return. The loop should improve the user proposition, not just company revenue.
Pro tip Write the cycle as a sequence of causal arrows and challenge every link.
Watch out Marketing-driven growth without rising participant utility is not increasing returns.
- 3
Test Convergence
Ask whether participants benefit from gathering on one dominant network rather than dividing attention across several. Look for convenience, liquidity, selection, or reliability advantages.
Pro tip Measure multi-homing behavior to see whether users actually need competing platforms.
Watch out Do not claim winner-take-all dynamics where switching and multi-homing remain effortless.
- 4
Account for the Cold Start
Identify what must be subsidized, sold, or manually assembled before the loop becomes self-reinforcing. Evaluate whether the company can survive this phase.
Pro tip Start in a narrow geography or category where density is attainable.
Watch out Long-run network effects do not erase short-run distribution and operating risks.
- 5
Search for Transferable Markets
After validating the mechanism, scan other industries where fragmented participants face similar discovery or coordination problems. Reapply the model rather than copying surface features.
Pro tip Compare market structure and participant incentives, not just product appearance.
Watch out A loop that works for restaurant reservations may fail where transactions are rare or trust requirements differ.
In the wild
Benchmark initially faced objections that selling computer hardware to small restaurants was unattractive. Gurley's thesis was that comprehensive restaurant supply would make OpenTable the preferred destination for diners, while concentrated diner demand would make participation increasingly necessary for restaurants.
→ The reinforcing supply-and-demand loop produced network effects that helped OpenTable become an indispensable reservation platform.
A transportation marketplace attracts riders with faster pickup and attracts drivers with denser demand. As both populations grow locally, waiting times fall and earning opportunities improve, strengthening the service relative to fragmented alternatives.
→ Local density creates a reinforcing loop and a defensible consumer marketplace.
Common mistakes
Rejecting the Ugly Starting Model
Early hardware sales or manual onboarding may look unattractive while still enabling a valuable network. Analyze the mature reinforcing system as well as the initial operation.
Calling Any Scale a Network Effect
Economies of scale lower costs, while network effects increase participant value. Confusing the two can produce an exaggerated competitive thesis.
Ignoring the Cold Start
A compelling end state does not explain how the first participants receive value. The company still needs a credible route to initial density.
Is it for you?
Best for
It is best for evaluating two-sided marketplaces, social platforms, and other products whose utility may grow with participation.
Not ideal for
It is not ideal for products where user growth creates congestion, participants receive no added utility, or customers readily use many interchangeable platforms.
From the transcript
“increasing returns was this argument that if you have the right pieces in place, your company will accelerate towards winner take all.”
“if you've got more, if you got all the restaurants on, the consumers would only want to go there, and if you got all the…”
“And then from there I started thinking about what other industries would that apply to, and that's what led to all these other things.”
From the episode
BONUS: Bill Gurley on Investing Early in Tech Disruptors & 'Runnin' Down a Dream'
Bill Gurley