Introduction: What a Marketplace Is and Why It's Hard
This chapter sets up the mental model: what a marketplace actually is, the three actors involved, and why marketplaces fail in ways linear businesses don't.
What a Marketplace Is
A marketplace is a platform that connects people who want to transact and takes a cut of what passes between them. It does not own the supply, does not hold inventory, and usually does not do the work. It runs the rules.
eBay connects people with used cameras to people who want cameras. Airbnb connects people with spare rooms to travellers. Upwork connects freelancers to clients. Uber connects drivers to riders. In every case, the platform is not the supplier and not the consumer. It is the infrastructure, the match-maker, and the referee.
Contrast with other shapes of business:
Linear commerce Amazon buys a book, stores it, sells it to you
SaaS Notion writes software, sells access to many users
Ad network Meta runs a service, sells attention to advertisers
Marketplace Etsy owns no pottery; it owns the place where pottery is sold
A marketplace's product is the transaction itself, and the trust infrastructure that makes the transaction possible.
The Three Actors
Every marketplace has three actors, and you will keep encountering them:
Platform runs the rules, takes a cut, owns the data and the brand
Supply side sellers, hosts, drivers, freelancers, artists
Demand side buyers, guests, riders, clients, fans
The platform's job is to make the supply side and demand side transact in a way that profits everyone, including itself. When any of the three stops benefiting, the marketplace breaks.
Some marketplaces have more than two sides. DoorDash has three: customers, restaurants, and drivers. YouTube has at least three: creators, viewers, and advertisers. The complexity scales with the side count, but the core idea does not change.
Why Marketplaces Are Harder Than They Look
A linear business grows its customer base. A marketplace has to grow two customer bases, in parallel, each of which is worth less if the other isn't there.
You've opened a cafe with no customers. That is a sad cafe, but fixable: put an ad in the window, discount a few coffees, wait. A marketplace with no supply is not a marketplace at all. Buyers arrive, find nothing, leave, don't come back. The same thing happens in reverse: a marketplace with no buyers is a room full of sellers wondering why they showed up.
Three compounding problems:
- The chicken-and-egg problem. You can't get demand without supply, and you can't get supply without demand. Day one is a coordination trap
- Liquidity threshold. Below a certain density of activity, the marketplace is unusable, and above it, it is self-sustaining. Getting from one to the other takes years and cash
- Two audiences, two messages, two product surfaces. You are always running two products at once, each with its own churn, its own economics, and its own grievances
The consolation: once these problems are solved, a marketplace becomes a compounding business with deep defensibility. The value belongs to the pool of buyers and sellers, which is almost impossible to copy.
What You Get if It Works
- Defensibility. A working marketplace is hard to unseat. Supply and demand don't want to multi-home if they don't have to
- Capital efficiency. You do not own the cars, the apartments, or the inventory. The margin profile, once volume is there, is good
- Compounding value. Every new user on one side makes the platform more valuable to the other side
What It Costs
- Years of subsidised growth. Most successful marketplaces spent a long time paying users to show up. Uber's early subsidies were famous; DoorDash, Instacart, and Airbnb all had variants
- Asymmetric failure modes. A product can be half-built and useful. A marketplace can be half-built and useless
- Two churn curves. Supply churn and demand churn compound, and they are often linked
Why You Might Care
Marketplaces show up everywhere:
- Labour (Upwork, Fiverr, TaskRabbit)
- Travel and lodging (Airbnb, Booking)
- Transportation (Uber, Lyft, Bolt)
- Food (DoorDash, Uber Eats, Deliveroo)
- Goods (eBay, Etsy, Poshmark, StockX)
- Content (YouTube, SoundCloud, Substack in parts)
- Capital (Kickstarter, LendingClub)
- Dating (Tinder, Bumble, Hinge)
If you are building one, the chapters ahead are the concepts you'll wish you'd internalised earlier. If you are using or investing in one, the same concepts let you read its numbers without flinching.
What This Tutorial Covers
- Chapters 02 to 06 are the core concepts: two-sided markets, liquidity, matching, pricing, and trust
- Chapters 07 to 09 go deeper: network effects, unit economics, and the operational reality of imbalanced markets
- Chapters 10 and 11 cover scaling, plus five case studies
- Chapter 12 distils the patterns and anti-patterns
The thread that runs through all of it: marketplaces are not products, they are coordination problems. Solve the coordination and everything else becomes possible. Skip it and nothing else matters.
Common Pitfalls from Day One
"We are like a marketplace but we handle the supply ourselves." Then you are not a marketplace; you are a vertically integrated retailer with extra steps. Not a bad business, just a different one with different rules
"We have 10,000 users." Useless unless they are transacting. Vanity metrics mean less in a marketplace than anywhere else
"We'll focus on supply first, then demand." Or the reverse. The question is not which side first; it is what value you deliver to that side before the other side exists. Chapter 02 covers the four strategies
"We will just list things and people will come." This is the classifieds fallacy. Listing is not a transaction. A listing that never sells is a disappointment, not a product
Next Steps
Continue to 02-two-sided-markets.md to understand the two-sided model that underlies every marketplace.