What should our take-rate be? It is the third question every marketplace founder asks me, right after “how do I get suppliers” and “should I use Sharetribe.” The answer is the same every time, and it disappoints them. There is no right number. There is only a right strategy, and the number falls out of the strategy. When founders chase a number first, they end up with an arbitrary fee that hurts whichever side they accidentally undervalued.
I have built and operated marketplaces with my own money. I have set take-rates that worked, set ones that did not, and watched competitors set ones that broke their economics inside a year. The pattern in every failure was the same: the take-rate was set against a benchmark, not against a strategic choice. This article is the framework for choosing the strategic choice.
The Wrong Question
“What is industry standard?” is the wrong starting question, because there is no industry standard. Etsy charges 6.5 percent. eBay charges 13. Airbnb charges around 14. Uber charges around 25. Upwork charges 10 to 20 depending on tier. These are not random. Each of those numbers reflects a specific decision about who creates value, who captures it, and what is being defended.
Setting your take-rate by averaging those benchmarks is like setting your salary by averaging the pay across every industry. It tells you nothing about your situation. The right question is not “what is normal?” The right question is “what does my marketplace need to do, and what does that imply about how I should price it?”
Who Creates The Value
Every marketplace fee structure should answer one question: which side of the marketplace creates more of the value, and which side feels the friction of paying for it. The answers are not symmetrical, and they often surprise founders.
On Airbnb, the host creates most of the value (the property, the photos, the experience). The guest captures it. Airbnb charges the guest more than it charges the host, because guests are willing to pay for trust, search, and customer support, while hosts would route around any platform that taxes them too heavily.
On Uber, the driver creates the value (the car, the time, the labour). The rider captures convenience. Uber takes a large chunk from the driver because drivers have low alternatives, and rider-side prices are constrained by the cab industry as a benchmark. This is why driver pay has been a continuous source of friction for the platform.
On Etsy, both sides create some of the value (sellers create the goods, buyers create the audience that makes the platform attractive to sellers). Etsy charges sellers relatively little, because seller migration to alternatives is the biggest threat to its survival. The take-rate is structured to retain supply.
“Your take-rate is the most public piece of strategy you ship. It tells your suppliers and your buyers exactly how you have decided to share the value. Get it wrong and one side leaves.”Darren Cody, Marketplace Studio
Supply-Side Fee, Buyer-Side Fee, Or Both
The first decision is who pays. There are three options, each with different consequences.
Supply-side fee only. The marketplace takes its cut before paying out the supplier. The buyer sees the listed price and pays exactly that. Best when buyer price-sensitivity is high and switching cost on the supply side is also high. Etsy works this way. So does Substack.
Buyer-side fee only. The supplier receives their full listed price; the buyer pays a service fee on top. Best when the marketplace adds obvious value to the buyer (trust, payment protection, search) and you need to keep supply price-competitive with off-platform alternatives. StubHub historically worked this way. So do most ticketing marketplaces.
Both sides pay. A small fee on each. Best when both sides receive obvious value and you need predictable margins on every transaction. Airbnb works this way (around 3 percent host, 14 percent guest). The trade-off: every time you raise either fee, the side with the lower switching cost notices first.
The Hidden Take-Rates Founders Forget
The headline take-rate is rarely the whole picture. The marketplaces that work long-term tend to have multiple, smaller take-rates layered on top of a transparent base. Founders set the headline number and forget there are other levers. Here are the levers that actually exist.
Payment-processing margin. If you pass through the payment processor cost, you take a small loss on every transaction. If you mark it up by even 0.5 percent, you have created a quiet additional revenue stream that does not touch the headline rate. Most large marketplaces do this.
Premium placement.Boosted listings, sponsored slots, search rank purchases. eBay's sponsored listings program is now a meaningful share of its revenue, separate from take-rate. Etsy's ads program is similar. This is take-rate that suppliers opt into, which feels different to them than take-rate that gets deducted automatically.
Subscription tiers. A monthly fee for premium features (analytics, multi-listing, custom storefront) charged to suppliers. The attractiveness: it is a fixed cost suppliers can plan around, which feels less hostile than a percentage of every transaction.
Insurance and trust products. Damage protection, identity verification, dispute mediation. These are paid services that solve real problems, and they typically have very high gross margins. They also tend to tie users deeper into the platform, which compounds.
Float. The interest you earn on transaction balances held between buyer payment and supplier payout. At scale, this is meaningful.
Real Benchmarks, With The Reasoning
Benchmarks are useful as comparisons, not as targets. Here are the rates of marketplaces I would bet most readers know, with the rationale behind each.
| Marketplace | Headline rate | Who pays | Why this rate works |
|---|---|---|---|
| Etsy | 6.5% | Seller | Low fee retains diverse, low-margin supply |
| eBay | 13%+ | Seller | Variable by category; supply has high switching cost |
| Airbnb | ~14% (3% host + ~11% guest) | Both | Trust and search are buyer-valued |
| Uber | ~25% | Driver | Drivers have low alternatives in many cities |
| Upwork | 10–20% | Supplier (talent) | Tiered: lower for repeat clients |
| DoorDash | 15–30% (restaurant) | Supplier | Demand aggregation; supply is fragmented |
| StubHub | ~25% combined | Both | High trust premium on both sides |
| Fiverr | 20% supplier + 5.5% buyer | Both | Highly atomized supply |
Patterns worth noticing: high take-rates exist where the marketplace solves a real aggregation problem and the supplier has limited alternatives. Low take-rates exist where supply churn would destroy the platform faster than the higher fee would generate revenue. The choice is structural, not aesthetic.
Raising Take-Rate Without Losing Supply
Most marketplaces start with a low take-rate to attract supply, then raise it as the marketplace becomes important to the supplier's revenue. This is not deception. It is a value-share evolution. The mistake is raising without giving anything back.
When we have helped clients increase their take-rate successfully, three things have been true at the same time. The marketplace was already producing a meaningful percentage of the supplier's revenue (typically more than thirty percent, often more than fifty). The increase was paired with a tangible new benefit (better tools, marketing reach, payment protection). And the timing was telegraphed: suppliers got ninety days of notice and a clear explanation of what was changing and why.
When clients have done this poorly, the rate increase landed cold, untied to any value improvement, and supplier sentiment turned within a quarter. We have seen ten-percent rate increases trigger fifteen-percent supply churn when handled badly. The losses compounded for two reasons: the suppliers who churned were often the higher-volume ones, and the public conversation among remaining suppliers shifted, which slowed new supply acquisition for the next year.
The Freemium Illusion
A common founder instinct is to launch with a zero or near-zero take-rate to attract supply. The instinct is half-right. Some kind of supply concession during the first-twenty stage is sensible (we covered this in the cold-start post). But a permanently low or zero take-rate is rarely a strategy. It is a deferred decision.
The problem with freemium-forever take-rates: they train your suppliers to expect something for nothing, which makes any future fee feel like a betrayal. They also starve the marketplace of the operating capital it needs to invest in trust, search, and supply acquisition, which is what makes the marketplace valuable to suppliers in the first place. A marketplace with no take-rate is a marketplace with no budget to earn supplier loyalty.
A more honest pattern: launch with a low rate, communicate clearly that it is an introductory rate, and tie the future increase to specific milestones the marketplace will hit (volume, features, supplier earnings thresholds). Suppliers respect a clear plan. They resent feeling tricked.
How To Test A Take-Rate Change
Take-rate changes are one of the few places in marketplace strategy where A/B testing is genuinely informative. The standard pattern: pick a single category or geography, change the rate, watch three numbers for ninety days. Listing rate (does new supply slow?), churn rate (do existing suppliers leave?), and effective revenue per transaction (does the change actually improve unit economics?).
The trap is testing too small a change too briefly. A one-percent rate change on a two-month window will not produce a statistically significant signal in any but the highest-volume marketplaces. Test changes of at least three percent over at least ninety days, and segment the analysis by supplier tenure: long-tenured suppliers react differently than new ones.
The other trap is testing the rate without testing the messaging. The same fee change, announced two different ways, can produce dramatically different supplier sentiment. The fee structure is half of the experiment. The framing is the other half.
Take-rate is a strategic statement. It tells suppliers what you think they are worth and tells buyers what you think you are worth. Set it deliberately. Defend it when the numbers support it. Change it when the strategy changes. Never let the number be an accident.