The Three-Signal Marketplace Screen
Fragmentation, uniform needs, high matchmaking barrier — no marketplace works without all three.
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 4
- Confidence
- 92%
Before deciding a business should be a marketplace at all, Lauzier screens the category on three attributes. High fragmentation on both sides means aggregation creates value. Relatively uniform needs means supply can be commoditised into a matchable unit. A high existing barrier to matchmaking and vetting means there is real work for an intermediary to do. Weak on any one and the marketplace model fights the market.
Origin
Benjamin Lauzier's screening heuristic from ~15 years building marketplaces (Lyft, Thumbtack) and advising marketplace founders. He uses Thumbtack as the cautionary case on the uniform-needs criterion — service marketplaces score badly on it and are correspondingly hard.
Core principles
- 01If a handful of big players control a side, buyers can find them directly and the aggregator adds nothing.
- 02Fuzzy, idiosyncratic supply needs make matching nearly impossible — an electrician wants only certain jobs, only on days they are free, and will cancel for something better.
- 03The higher the effort people currently spend finding and vetting each other, the bigger the opportunity for an intermediary.
- 04Services marketplaces are feasible but structurally disadvantaged on the uniform-needs axis.
How to run it
- 1
Test fragmentation on both sides
Look for a long tail of buyers and sellers with no handful of dominant players controlling the market. Aggregation only creates value where the counterparty is genuinely hard to enumerate. Nobody needs a marketplace to find one of five airlines.
Pro tip Check fragmentation on both sides — one fragmented side and one concentrated side usually means the concentrated side will disintermediate you.
- 2
Test uniformity of needs
Ask whether supply can be commoditised into a comparable unit. eBay sellers want to sell clear, distinct inventory. Thumbtack electricians want particular job types, only when free, and may cancel when something better appears — a fuzzy definition of supply that makes matching brutal.
Pro tip Write down the unit of supply in one sentence. If you cannot, the needs are not uniform enough.
Watch out Non-uniform needs do not make a marketplace impossible, but they are not a compelling attribute for building one — go in with eyes open.
- 3
Test the matchmaking barrier
Measure how hard it is for the two sides to find and vet each other today, and how much effort that costs them. The higher the barrier, the bigger the opportunity, because you can implement processes that simplify the exchange. You cannot flag down a stranger's car or ask to sleep in their home.
Pro tip Vetting effort counts as much as discovery effort — trust infrastructure is often the real product.
- 4
Reject or proceed
Only proceed with a marketplace model when all three signals are present. Failing one is a signal to consider a different model — direct supply, SaaS, or a managed service.
Pro tip Run this before writing 'Airbnb for X' on a deck, not after.
In the wild
Unlike eBay, where sellers list clear distinct inventory, Thumbtack's electricians only want specific job types, only on days they are available, and may cancel a job when something better appears. Two electricians perceive the same unit of demand completely differently.
→ The marketplace was feasible but structurally difficult, which Lauzier cites as the reason service marketplaces are so hard to run.
Drivers and riders are both massively fragmented; the need is uniform (get me from A to B); and the pre-existing matchmaking barrier is enormous — you cannot flag down a random private car and trust its driver.
→ All three signals present, which is why the model supported multi-billion-dollar marketplaces.
Common mistakes
Forcing a marketplace onto a concentrated market
Where a handful of large players control a side, buyers can just go direct. The aggregator adds no value and the big players have every incentive to disintermediate.
Ignoring the unit-economics question entirely
Even a category that passes all three signals can fail on business model. On-demand car washes and similar 'Uber for X' ideas died because nobody would pay what it costs to send a person to do the job.
Is it for you?
Best for
Founders and investors evaluating whether an idea should be a marketplace at all, before any product is built.
Not ideal for
Existing marketplaces at scale — the model question is already settled and liquidity management matters more.
From the transcript
“high fragmentation I think you want this long tale of buyers and sellers without a handful of big players controll in the market”
“I think you also want a relatively uniform uh set of needs uh that means that it can be like your supply can be commoditized…”
“high enough Bearer uh in the matchmaking or the creation um I think how hard it is for like people to find each other today…”
From the episode
How marketplaces win: Liquidity, growth levers, quality, and more
Benjamin Lauzier (Lyft, Thumbtack, Reforge)