Third-party sellers on Amazon generated an estimated $575 billion in gross merchandise volume in 2025 (Capital One Shopping Research, 2025), more than the entire U.S. e-commerce market was worth a decade ago. Those sellers are not Amazon employees. They are independent businesses plugging into a platform Amazon built to connect buyers and sellers. That platform is a two-sided marketplace, and understanding how it works is one of the highest-leverage ideas in modern growth.
Quick Summary
- A two-sided marketplace connects two distinct user groups who each need the other to get value.
- Supply and demand must grow together, you cannot have one without the other.
- The cold-start problem is the defining challenge: how do you attract side A when side B is not there yet?
- Successful marketplaces often subsidize one side early, then monetize later.
- Network effects compound over time, making established marketplaces extremely hard to displace.
What It Actually Is
Think of a farmers market. The market organizer does not grow the food or eat it. Instead, the organizer creates a venue where farmers (supply) and shoppers (demand) can find each other. The organizer's job is to make sure there are enough farmers to attract shoppers, and enough shoppers to keep farmers coming back.
A two-sided marketplace is the digital version of that. The platform sits in the middle, sets the rules, and takes a cut of each transaction. Classic examples:
- Airbnb connects hosts (supply) with travelers (demand).
- Uber connects drivers (supply) with riders (demand).
- Etsy connects sellers (supply) with buyers (demand).
- Amazon Marketplace connects third-party sellers (supply) with shoppers (demand).
The platform itself rarely owns the inventory or performs the service. Its value comes entirely from the density and quality of both sides.
Why It Matters
Two-sided marketplaces have become some of the most valuable businesses ever built because of a structural advantage: cross-side network effects. When more drivers join Uber, the app becomes more useful for riders. When more riders use Uber, it becomes more attractive for drivers. Each side's growth pulls the other side along.
PwC had projected the sharing economy, largely built on two-sided marketplaces, would reach $335 billion by 2025 (PwC, Global Sharing Economy Report), a milestone the sector has now passed. Amazon's marketplace services revenue alone hit $184.1 billion in full-year 2025 (Amazon Annual Report, 2025). As of the most recent reporting, more than 60% of Amazon units sold came from independent third-party sellers (eDesk, 2025), not from Amazon's own retail arm.
For a growth marketer, understanding marketplace dynamics matters even if you never build a marketplace from scratch. Many modern B2B SaaS products, content platforms, and app stores are effectively two-sided. The same cold-start tactics and monetization principles apply.
How It Works
The Four Stages
Which Side Should You Seed First?
| Scenario | Seed This Side First | Reason |
|---|---|---|
| Supply is scarce or hard to recruit | Supply | Demand will not show up if there is nothing to buy |
| Demand is the bottleneck | Demand | Suppliers will not join an empty market |
| Both sides are hard | Supply in a niche | Easier to dominate a small market than spread thin |
| One side is also the product | Supply | Hosts ARE the product on Airbnb |
The most common answer is to seed supply first in a narrow geography or category. You need something for demand to discover before you spend money acquiring demand.
Real-World Examples
Airbnb's Supply-First Launch (2008): The founders personally photographed early listings in New York City. They offered to improve host photos for free, which directly increased booking rates. By building supply quality first in a single city, they gave travelers a reason to choose Airbnb over hotels. The tactic was manual, unscalable by design, and it worked.
Uber's City-by-City Expansion: Uber did not try to launch everywhere at once. It launched in San Francisco, subsidized both drivers (guaranteed hourly minimums) and riders (free first rides), and built density before expanding. By 2024, Uber operated in over 70 countries with 7 million active drivers globally (Uber Investor Relations, 2024).
Etsy's Craft Fair Strategy (2005-2007): Etsy's early team physically attended craft fairs to recruit sellers. They signed up hundreds of artisans before the site even launched to the public. When buyers arrived, there was already a curated, high-quality supply catalog waiting. Etsy reached $13.2 billion in GMS in 2023 (Etsy Annual Report, 2023).
The Cold-Start Problem
The cold-start problem is the defining challenge of any marketplace: you need both sides to have value, but neither side will join without the other.
There are four proven approaches:
- Fake it until you make it: Craigslist listings were manually seeded in the early days. Reddit's founders created fake accounts to populate discussions.
- Single-side value: Build a product that is useful to one side even without the other. OpenTable gave restaurants free reservation software before diners used the platform.
- Subsidize one side: Offer the supply side free access, guaranteed income, or promotional pricing early on.
- Aggregate existing supply: Zillow pulled real estate listings from MLS databases before any agents signed up directly.
The Subsidy Trap: Subsidizing one side to attract the other works, but it can mask a fundamentally broken unit economics model. If drivers only show up because of guaranteed minimums, and riders only show up because of discounted rides, you may never find the price point where the marketplace is self-sustaining. Always model what happens when subsidies end.
Monetization Models
| Model | How It Works | Example |
|---|---|---|
| Transaction fee | Platform takes a % of each transaction | Airbnb (3% host + 14% guest) |
| Subscription | One or both sides pay a recurring fee | Amazon Seller (monthly + referral fee) |
| Listing fee | Sellers pay to list items | Etsy ($0.20 per listing) |
| Freemium | Basic access free, premium features paid | Upwork (free to list, paid connects) |
| Data/Ads | Platform sells data or ad placements | Amazon Sponsored Products |
Most mature marketplaces use a hybrid. The key question: which side captures more of the value, and which side is more price-sensitive? The more price-sensitive side usually gets the lower price (or a subsidy early on).
Common Mistakes
Mistake 1: Trying to be global on day one. Geographic density matters. A ride-hailing app with 1,000 drivers spread across 50 cities is worse than one with 1,000 drivers in a single metro. Start local.
Mistake 2: Optimizing for quantity over quality on the supply side. More hosts, more drivers, more sellers is not always better. One bad Airbnb experience can churn a guest permanently. Quality controls (ratings, vetting, guarantees) are a growth lever, not just an ops issue.
Mistake 3: Monetizing too early. Taking a cut before either side sees real value will kill growth. Most successful marketplaces waited until they had strong liquidity before turning on fees.
Mistake 4: Ignoring disintermediation. If buyers and sellers can find each other on your platform and then transact off-platform to avoid your fee, your moat is weak. Contracts, trust systems, and platform-only features (payments, dispute resolution, insurance) keep transactions on-platform.
Mistake 5: Treating both sides as identical. Supply and demand have completely different needs, pain points, and messaging. Your acquisition channels, onboarding, and retention tactics should be built separately for each side.
The Single-Player Mode Test: Before you have both sides, can your product deliver value to just one side? If yes, you have a wedge. Yelp was a review site (useful to readers) before it was a marketplace for restaurants. Notion was a personal productivity tool before it became a team collaboration platform. A strong single-player mode lowers the cost of acquiring your first side.
Key Takeaways
- Two-sided marketplaces create value by connecting two distinct groups, not by owning supply or performing services.
- Cross-side network effects are the core moat: each side's growth makes the other side more valuable.
- The cold-start problem requires manual seeding, subsidies, or a single-side product to solve.
- Start with geographic or category density before expanding.
- Monetize after liquidity, not before.
- Quality controls on the supply side are a growth lever disguised as an ops problem.
