Every freelancer knows the cut hurts. Fewer ask why it's 20%, whether it has to be, and what actually breaks if you remove it. I removed it. Here's the economics.
The number nobody questions
A freelancer invoices $5,000. The platform takes $500 to $1,000. The freelancer waits several days to receive the rest. Before any of that, they may have spent money just to submit the proposal.
We've normalized this. But "20% commission" isn't a law of physics — it's a pricing decision, and like any pricing decision it can be wrong, or at least not the only option. So let's actually take it apart.
Why the commission exists (the steelman)
To be fair to the incumbents, the cut isn't pure greed. A marketplace genuinely pays for:
Trust and dispute resolution — escrow, mediation, fraud prevention.
Payments infrastructure — processing, compliance, chargebacks, payouts across borders.
Demand generation **— the clients are there because the platform spent to acquire them.
**Matching — surfacing the right job to the right person.
These are real costs. The question isn't whether the platform should be paid. It's which side of the transaction should pay, and in what shape.
The problem with percentage-of-labor
Charging a percentage of the freelancer's earnings has three bad properties:
It scales with effort, not with cost-to-serve. Processing a $10,000 invoice doesn't cost the platform 20× what a $500 one does — but the freelancer pays 20× more. The fee is decoupled from the service rendered.
It taxes the party experiencing it as loss. The freelancer already did the work. The cut feels like money taken out of their pocket, because it is. Clients, by contrast, experience fees as part of a price they chose to pay.
It compounds with other frictions. Add pay-to-bid and delayed payouts on top, and you get a system where the freelancer pays to apply, pays to win, and waits to collect.
The 0% experiment
So flip it. Take the fee off freelancer earnings entirely, and structure platform revenue as a flat, transparent fee that doesn't punish people for doing larger or more frequent work. Freelancers keep 100% of what they invoice.
This is the bet behind OpenLance.io, the marketplace built by freelancers for freelancers. The mechanics:
0% freelancer commission. Revenue comes from a flat platform fee, not a percentage of labor.
Escrow on every job. Clients fund each milestone before work begins. This is the trust primitive — both sides can see the money is committed.
Instant payouts. Approval triggers immediate release. No multi-day float.
Flat bidding. One proposal, one cost, any job size — so reputation wins work, not ad budgets.
"But 0% can't be a business model"
The most common objection, and it deserves a real answer, not a dodge.
0% commission is not the same as 0% revenue. The platform still charges — it just charges in a shape that's decoupled from the freelancer's invoice size. A flat fee structure, transaction-level fees that reflect actual cost-to-serve (payments, escrow), and optional paid tiers are all revenue that doesn't require skimming labor.
Will it sustain at scale? That's the genuinely open question, and I'd be lying if I claimed certainty. The honest framing: incumbents chose percentage-of-labor because it's the easiest high-margin model, not the only viable one. Whether a leaner, flatter model can fund the same trust infrastructure is exactly what launching proves or disproves. I'll be publishing what I learn either way.
The part engineers will care about: solving cold-start
You can have the most elegant fee model in the world and still die of the two-sided marketplace problem: no freelancers without clients, no clients without freelancers.
The architecture decision that matters here isn't in the code — it's in the sequencing. I onboarded the client side first, with active, responsive clients posting real work, before opening freelancer recruitment. So freelancers arrive to a marketplace with live jobs, not an empty schema. Supply-side-first or demand-side-first is one of the highest-leverage early decisions a marketplace makes, and for a freelancer platform, having the work ready is what makes the freelancer pitch credible.
What I'm watching
If you're building marketplaces, these are the metrics I'd argue actually matter early, in order:
Client responsiveness / liquidity — are posted jobs getting real engagement? A job nobody answers is worse than no job.
Activation, not just signups — did a freelancer apply to real work, or just create an account and leave?
Repeat behavior — does a completed job lead to a second? That's the only proof the model has product-market fit.
Commission rate is a positioning lever. Liquidity is the survival lever. I optimized positioning to 0% precisely so liquidity has room to be the thing I obsess over.
I don't think 20% is evil. I think it's lazy — the default high-margin choice, not the only sustainable one. OpenLance is my attempt to prove a flatter model can fund the same trust and still leave freelancers whole. It might not work. But the only way to find out is to ship it and publish the numbers.
If you've built marketplaces and know where flat-fee models break first, tell me in the comments — I'd rather learn it from you than from a churn chart.
Top comments (0)