Small-Team Incubation Ladder
Start with a tiny team, prove real behavior, then add resources after fit.
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 6
- Confidence
- 97%
Liu's incubation rule is to resist going full bore on an unproven idea. Begin with two to five people, often with leaders who also support other products, so the bet does not acquire the cost and scrutiny of a hundred-person commitment. Search for evidence that users already perform the desired behavior and study the difference between people who believe in it and those who cannot imagine it. Then identify the binding constraint to scale. For Facebook Marketplace, local sale groups already had supply, but demand was fragmented across closed communities, so broader liquidity became the early problem. Small teams test solutions, build the trust and safety groundwork, and look for product-market fit. Only after evidence appears does the organization double down. Resources become a response to learning rather than a substitute for it.
Origin
Liu used this approach while incubating Facebook Marketplace. Small teams studied organic buying and selling in local Facebook groups, diagnosed fragmented demand, and expanded only after learning from real behavior.
Core principles
- 01Resources should follow evidence of product-market fit.
- 02A large starting team creates cost, scrutiny, and a visible target.
- 03Existing user behavior can reveal a product before the company formalizes it.
- 04The first job is to identify and remove the binding constraint.
- 05Trust and safety are part of marketplace groundwork, not afterthoughts.
How to run it
- 1
Keep the initial team tiny
Assign two to five people and limit dedicated leadership overhead. Give the team enough capacity to test the behavior without declaring a massive organizational bet.
Pro tip Leaders can retain responsibilities on other products during the earliest stage.
Watch out A large starting team makes an uncertain bet expensive and politically conspicuous.
- 2
Find behavior before building
Look for users already approximating the product with groups, workarounds, or manual processes. Treat that behavior as evidence of a need, not proof that the final solution is known.
Pro tip Study both active users and people who cannot imagine adopting the behavior.
- 3
Name the binding constraint
Determine what prevents the organic behavior from becoming broadly useful. Marketplace already had listings, but closed local groups fragmented demand and limited liquidity for less common items.
Pro tip Distinguish a supply problem from a distribution or demand problem before adding features.
- 4
Test small solutions
Run focused experiments that relieve the constraint and watch whether real usage improves. Preserve the ability to change direction without defending a large sunk cost.
Watch out Executive enthusiasm is not product-market fit.
- 5
Build the operating foundations
Add the fraud, authenticity, trust, and safety mechanisms required for the behavior to work at broader scale. These systems are part of the product's viability.
Pro tip Use assets the parent company already owns when they create a genuine advantage.
- 6
Double down after fit
Increase resources when users repeatedly receive value and the next constraint is clear. Scale the evidence-backed product rather than the original pitch.
Pro tip Share concrete customer outcomes internally to make the demonstrated value legible.
Watch out Do not confuse a single success story with enough evidence to scale the whole organization.
In the wild
Before an open marketplace existed, local Facebook groups already let parents buy and sell strollers, cribs, and other goods. The behavior proved supply and local trust, but closed groups could not provide enough demand for uncommon items such as cars. A small team focused on expanding liquidity while adding fraud and authenticity protections.
→ The company formalized behavior the community had already validated and eventually scaled Marketplace to more than one billion monthly users, according to the episode.
Common mistakes
Staffing the pitch
Putting a hundred people on an unproven idea increases cost and scrutiny before the team knows what users need.
Ignoring organic behavior
Teams can miss that users have already built a rough version of the product and revealed the real constraint.
Treating trust as a later feature
A marketplace cannot scale safely if fraud, authenticity, and human trust are postponed until after distribution.
Is it for you?
Best for
Large-company teams incubating a new product whose core behavior can be tested with a small group.
Not ideal for
Infrastructure migrations or compliance programs whose value cannot be proven through a small market-facing experiment.
From the episode
How to own your career growth and become a powerful product leader
Deb Liu, Ancestry (ex-Facebook, PayPal)