Bazaar
The bustling marketplace where buyers and sellers fuel each other.
The Pattern
What the Bazaar pattern is
On one side of a Bazaar are sellers offering something and on the other are buyers looking for it, and the Bazaar keeps a percentage of every sale between them (the take rate). Each side is there because the other one is, so the activity itself is the product. Empty stalls drive shoppers out of a real bazaar faster than empty aisles ever do. Keeping the number of buyers and the number of sellers in balance is the operator's job.
Think: eBay, Airbnb, Etsy, Uber, StockX, OpenTable, Coinbase.
The Engine
How it makes money
A Bazaar is built on matching, which is how it gets a seller and a buyer to say yes to each other. In practice that's the trust, search and transaction layer, and it has to work well enough that both sides want to deal through you instead of going around you. Underneath matching sits liquidity (enough active buyers and sellers in one place that matches happen quickly), which is the second keystone. If you starve either side of the market, or lose the trust layer, the marketplace starts to fall in on itself.
A Bazaar brings buyers and sellers into one place, where it's easy for them to find each other and trust each other enough to do a deal. The platform doesn't produce anything itself, and all of its value comes from taking the friction out of that process. Every successful match makes the next one easier and cheaper, and that compounding is what the business runs on. In a mature Bazaar, buyers don't think to look anywhere else, and sellers can't afford to leave because that's where the buyers are.
The Discipline
Operating principle
Liquidity is what protects a Bazaar from competitors. As more buyers and sellers deal with each other on the platform, the platform gets harder to replace. But that liquidity belongs to one geography and one category and it's never global, which means winning in one city or category doesn't carry over to the next.
Trust and matching speed build on themselves the same way. Every transaction that's quick and fair, with nobody getting burned, makes the next ten more likely to happen on the platform instead of around it. The reverse is just as true, because once buyers stop trusting it or either side starts to thin out, the marketplace comes apart far faster than it took to build.
Healthy
The healthy pattern is to keep subsidizing the harder side until the marketplace grows on its own, and from then on to defend the take rate hard.
Unhealthy
The tension never goes away, because if you subsidize one side for too long you bleed cash, but if you try to balance the two too early, neither side shows up.
We'd never launch a marketplace nationally on day one, and we'd saturate a single zip code, city or vertical until that first market is locked in.
What Works & What Breaks
Healthy vs. breaking
What healthy looks like
What breaks it
The Profile
Profile
Best Fit Founder
Bazaars need an ecosystem builder, a systems thinker who tolerates a lot of chaos and is obsessed with unit economics. They're as much a product person as an operator, and they think a lot like a behavioral economist.
Go-To-Market
You're acquiring both sides of the market together, and you do it one geography or category at a time. The standard move is to over-subsidize the harder side, usually supply, until liquidity tips, and then let the cheaper side flow in. A Bazaar should never launch nationally on day one.
Ideal Focus
The focus is driving liquidity, which means reducing friction on every match while protecting the take rate. We'd measure it with match rate, time-to-match, fill rate and repeat usage per cohort, and pay much less attention to logo count or feature count. Every product decision should answer one question: does this make the marketplace more liquid for the next match? If it doesn't, it goes to the back of the list. The number one thing that stalls a marketplace's momentum is the founder's attention drifting toward features and away from liquidity.
Revenue Pattern
Revenue grows non-linearly once liquidity arrives, so it picks up speed instead of rising at a steady rate. It's transactional and driven by volume, and it rises and falls closely with the seasons in your category. Outside of subscription tiers, there's very little revenue underneath it that's recurring and under contract.
Unhealthy Signs
Risk, Speed & Burnout
Sales Cycle
Revenue starts the day liquidity does, because the take rate earns money on every transaction. Getting to liquidity is the slow part, and it can take 18-36 months in each new geography or category. Profitability typically takes 5-10 years.
Hiring Sequence (first 5 after founder)
- Head of Supply: owns whichever side is harder to acquire, which is usually supply
- Head of Demand / Growth: brings in buyers and turns them into customers
- Trust & Safety lead: runs disputes and fraud prevention, which protects liquidity
- Data / Marketplace Analytics lead: owns the matching and pricing data, plus the cohort signals
- Category / Geo Launcher: takes the playbook into the next city or category
A Bazaar runs on search and matching infrastructure, ratings and reviews, escrow and payments, dispute resolution, fraud detection and messaging between the two sides. On top of that sits a lot of data engineering for the matching algorithms.
Running a Bazaar is a constant balancing act. You're handling problems on both sides at once while you run growth experiments, and operationally it's exhausting.
Leading KPIs
- GMV - the topline volume the take rate is applied to, and the truest measure of marketplace activity.
- Take rate % and how stable it is over time - the price you charge for liquidity. A rising take rate often hides slowing volume and speeds up disintermediation.
- Liquidity ratio (active demand / active supply, per geo or category) - balance is healthy, and a sustained imbalance is a leading indicator of churn on the starved side.
- Time-to-match - this is liquidity the way the buyer feels it. The lower it is, the higher repeat usage runs.
- Buyer retention % - the economics only work when buyers come back without paid reactivation.
- Seller retention % - supply churn costs more than demand churn, because supply is harder to replace.
- CAC, buyers and sellers separately - a blended CAC hides that one side is usually 5-10x more expensive than the other.
- Repeat purchase rate - repeat behavior by cohort is the cleanest signal of real marketplace health, because growth spending can't mask it.
- Off-platform leakage rate (estimated) - if users find each other on the platform and then deal off it, your take rate only exists on paper.
- Supply-side concentration (top 10% as % of GMV) - when a few sellers control most of the volume, they have outsize negotiating power and they're an outsize risk if they leave.
Feature Focus
Network effects let it scale.
Churn on one side puts the platform at risk.
CFO Watchpoints
What to watch
Defense
Defensibility
Concentration is generally low on the demand side once you're at scale. Supply is where the risk sits, because a few power sellers, hosts or drivers can hold the platform hostage.
The winner takes most of the market within a category and geography. Once liquidity tips, the leader's advantage keeps growing, and competitors have to subsidize indefinitely just to stay relevant. Watch the specialists that match better in a single category, because they're usually more dangerous than a generalist coming at you head on.
A Bazaar is protected by direct network effects and brand trust, plus the data from its transaction history and the habits that lock in users on both sides. When it works, we'd call it the strongest combined moat in business, but it's worth almost nothing the moment liquidity slips.
For any one user on either side, switching is cheap. Across the whole market it's expensive, because the value comes from the other side being there. What actually kills marketplaces is a coordinated migration, when a big group of users leaves together.
Financial Impact
The money
Impact
Cash flow is negative early, because you're subsidizing both sides, and growth turns non-linear once liquidity kicks in. The economics are asset-light and get much better with scale. At maturity a Bazaar can reach extraordinary margins, 60%+ contribution margin (what's left of revenue after the costs that come with each sale), but only after years of subsidy.
Target P&L Allocation
| Category | Target |
|---|---|
| Client Acquisition / Ad Spend | 30% |
| Total Labor | 15% |
| Fixed Overheads & G&A | 30% |
| Legal & Professional | 5% |
| Operating Income (EBITDA) | 10-20% |
Negative early (subsidizing both sides); non-linear growth after liquidity
Source: Compass archetype one-pager, target allocation as a percent of revenue
Support agents are on salary plus a QA (quality assurance) and CSAT (customer satisfaction score) bonus. The bonus is tied to how well disputes get resolved, response time and service ratings, rather than to ticket volume.
Account executives who onboard sellers are on a 40/60 base/commission split. Their accelerators (higher commission rates past a target) are tied to seller activation, meaning the first listing goes live and the first transaction is completed, rather than to signed contracts alone.
Refund and chargeback risk is moderate to high. Buyers file chargebacks against the platform even when the seller is at fault, and trust and safety claims can spike without warning. We'd reserve 1-3% of GMV for it.
You collect from the buyer when the sale happens and pay the seller later, so a Bazaar's working capital need is often negative (which is favorable). That timing is one of the model's biggest advantages, and we'd protect it hard.
Revenue Quality Score
6 / 10. Gross margin is high and usage recurs, but the revenue is transactional rather than contracted. It moves with category cycles, and the take rate is always under pressure.
Trajectory
Where this goes
The ceiling comes from saturating the category and the geography, and from burning out the supply side. Once every available host, driver or seller in a city is matched, organic growth slows dramatically. From there you either make more from each user (ads, payments, lending, subscription tiers) or expand into adjacent categories, and those are usually slower and more expensive than expected.
Pivoting is hard, because the core asset is liquidity in one specific category and that doesn't transfer. Expanding next door is easier (Airbnb adding Experiences) than moving the matching mechanism somewhere new.
At scale the exit is an IPO (initial public offering), like Airbnb, Uber and Etsy, or a strategic acquisition by a platform that wants the demand. Mid-size marketplaces struggle in M&A (mergers and acquisitions), because acquirers are rarely equipped to run them day to day.
Common Founder Failure Modes