The Marketplace Fragmentation Test
A marketplace only works if the market is fragmented and the per-transaction dollars are small enough to keep it in the middle
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 3
- Confidence
- 92%
Fragmentation — how many businesses exist relative to transaction volume — determines whether a marketplace can work. If a few concentrated players do most of the volume, they have their own sales teams, need the marketplace less, pay lower commission, and disintermediate. High per-transaction dollar value also makes going around the marketplace attractive.
Origin
A principle Dan Hockenmaier has applied repeatedly to evaluate B2B marketplaces, contrasting ride-sharing with high-value materials sourcing.
Core principles
- 01Fragmentation = number of businesses relative to transaction volume; check what share the top ~5% of suppliers do
- 02The more concentrated a side, the more leverage it has, the less it needs you, and the lower the commission it will pay
- 03High absolute dollars per transaction makes disintermediation worth the effort
- 04B2B often underperforms as a marketplace category precisely because it is less fragmented
How to run it
- 1
Measure concentration
Calculate what percentage of total volume the top ~5% of suppliers do; the higher that share, the less fragmented the market.
Watch out If ten companies do 80% of volume, they are big enough to have their own sales teams and operations, so they need the marketplace less and pay less.
- 2
Check dollars per transaction
Estimate how many total dollars are attached to each transaction; above a certain amount, disintermediation becomes very attractive.
Pro tip On a ride, commission is only two or three dollars, so it is rarely worth a driver's effort to call the passenger and go around Uber.
Watch out When per-order commission runs to tens of thousands of dollars, a supplier would rather pick up the phone and transact directly to save the fee — the marketplace breaks.
- 3
Judge whether you add enough value to charge
Confirm you bring enough value that you can charge a meaningful commission and still keep both sides on-platform.
Watch out If you can't charge anything meaningful, the business simply will not work — no matter how large the market.
In the wild
Manufacturers of beauty products source aerosol cans and inputs from a small number of big suppliers, with transactions of tens of thousands to millions of dollars each. The commission on such an order is too high relative to the value added, so the supplier would rather call the buyer directly and save the fee.
→ Low fragmentation plus high per-transaction dollars makes a marketplace unviable in that space.
Common mistakes
Launching a marketplace in a concentrated market
When a few players do most of the volume they have their own sales and operations, need the marketplace less, pay lower commission, and disintermediate — the marketplace can't sustain a business.
Ignoring absolute dollars per transaction
High-value transactions give both sides a large incentive to transact directly and save the commission, so even a fragmented-looking market can fail once per-order dollars get large.
Is it for you?
Best for
Founders and investors evaluating whether a proposed marketplace, especially in B2B, can actually work.
Not ideal for
Already-validated, highly fragmented consumer marketplaces where the question is moot.
From the transcript
“B2B also comes with something else which is much lower fragmentation in many cases and you need fragmentation for a good Marketplace”
“the more concentrated either side of your Market is the more leverage they have the less likely they are to need you”
“fragmentation is basically just a measure of how many total businesses are there in the space relative to the transaction volume”
“how many total dollars are attached to each transaction in the marketplace when it goes above a certain amount it becomes much more attractive to…”
From the episode
Developing a growth model + marketplace growth strategy
Dan Hockenmaier (Faire, Thumbtack, Reforge)