Aggregation
Агрегация
Accumulation Holds Depth 4 · Mine
an environment built for broad participation
Moats it grows into
- #2 Marketplace liquidity
A place where a listing clears fast and a search returns a match on the first try is where both sides come back; a rival with an empty market can promise neither.
- #1 Network effects
Each seller makes the place more useful to every buyer and each buyer to every seller; the participants' presence is the wall a rival has to rebuild.
Summary
A strategy of building the place where many parties come to find each other, and owning the place rather than any of the parties. Value comes from the count on each side: more sellers draw buyers, more buyers draw sellers. It holds because the density of participants is what a rival cannot copy, and every new participant makes the place harder to leave.
How it works
The mechanism is density on two sides that need each other. A one-sided crowd is usership; aggregation adds a second side, and each side shows up for the other. A reef is coral laying down structure, and fish, algae and crustaceans moving in because the others are already there; the structure is worth more than any resident. A mall signs anchor tenants to draw traffic, traffic draws the small tenants, and the landlord charges rent on all of it. Craigslist owns no inventory, only the page where a city's sellers and buyers expect to meet.
The edge is liquidity. A buyer who finds what they want on the first visit never tries the second place; a seller whose listing clears lists there again. A rival faces both sides at once, and each side waits for the other to arrive. The aggregator's work is to lower the cost of showing up and to keep the match rate high; the participants do the rest.
Why it holds — or only takes
Holds, because the strategy digs two of the deepest network moats directly. Marketplace liquidity is its direct output: a place where a listing clears fast and a search returns a match is the place both sides return to, and a rival with an empty market can promise neither. Network effects stack on top — each participant makes the place more useful to every other, on both sides, so the crowd defends itself. Both resist AI: a model can build the page, not the crowd on it. Neither is solo-reachable at scale, and liquidity borrowed from a bigger platform through its API is that platform's moat, not yours.
Do you have it
- Participants come to you to find each other, and you hold no inventory of your own.
- Time to first match is short enough that a newcomer returns without being reminded.
- Both sides list or search with you first, and elsewhere only after failing here.
- A rival subsidises one side heavily and still cannot get the other side to show up.
- Participants trade with each other in ways you never designed.
How to dig it
- Solve one side by hand first — seed the supply yourself or be the first buyer — until the other side arrives.
- Constrain the market to one category and one city until matches clear within a day, then widen.
- Make showing up free on whichever side is harder to attract.
- Instrument the match rate and the time to first match, and treat any drop as an outage.
- Give participants identity and reputation inside the place, so their standing cannot be carried out of it.
How it erodes
- Participants meet through you once and take the relationship off the platform.
- A vertical rival takes the most valuable category with a better match, and the general market thins from the top.
- Supply multi-homes: sellers list everywhere, and your side of the market stops being exclusive.
- The take rate climbs until a lower-take rival pulls the price-sensitive side away.
- Agents search every market at once, and the place stops mattering; only the match does.
Solo, without capital
A solo builder can aggregate one small market — a niche where both sides fit in a spreadsheet — and make the matches by hand until the page does it. Marketplace liquidity at scale is out of solo reach, and network effects only partly in it; both take years and subsidy money to reach a mass market, and a funded rival can buy one side outright. Neither moat rents cleanly: supply borrowed from a bigger platform is that platform's moat, and it can be withdrawn.
What the matrix says of its moats
- #2 Marketplace liquidity Solo ✗ no Capital ● high AI ↑ AI-resistant Rent ~ partly
- #1 Network effects Solo ~ partly Capital ● high AI ↑ AI-resistant Rent ✗ not for sale
In nature, in business
- Nature
- reefs
- Business
- cities, malls, Amazon, Craigslist
Cases
- Airbnb — Recruited the first hosts by hand and photographed their flats, so guests found real supply before hosts had a reason to list.
- eBay — Let anyone list and anyone bid, and became the one place a rare item was worth listing because the bidders were there.
- Craigslist — Runs one page per city with free listings in most categories and no inventory, and is still where a city expects to meet.
Combines with
- Usership — The one-sided crowd is the seed; add a second side that needs it, and the crowd becomes a market.
- Curation — Selecting who may participate keeps the match quality high while the count is still small.
- Skimming — A small cut of every match funds the place without taxing either side enough to leave.
- Chokepoint — Once both sides expect to meet here, the place becomes the passage every transaction passes through.
Fights with
- Highest bidder — Buying exclusivity on one side turns the open place into a distributor, and the other side stops arriving on its own.
- Decentralization — A place with no owner cannot capture the liquidity it creates; the moat is in the single point everyone passes.