Say you are building a marketplace and you decide to charge no commission. Anyone who has built one will ask the same question first: how does it make money?
That question deserves a real answer. Free describes where the price sits. If you cannot say where the price went instead, you have made a promise you may later break.
This is our answer for Open Lance, including the parts that are harder than taking a percentage.
Most free marketplaces are free only at the start
Most zero-fee marketplaces are percentage businesses running a promotion. Everybody involved usually understands the pattern.
The platform launches with no commission because it has to solve the cold-start problem. It needs buyers and sellers before either side has a reason to stay. It acquires both sides, waits until switching becomes painful, then introduces a small fee. After that, it raises the fee.
In that pattern, the fee is the eventual harvest of a position built while the platform was subsidised. It is less a price for a service than a charge made possible by control of the market.
Users are right to expect this. That makes life harder for anyone who means something different. The suspicion has been earned by the industry, and words will not remove it. The only real answer is to publish where the money actually comes from, then avoid doing the thing everyone expects.
The revenue comes from four lines
On Open Lance, revenue comes from four places. None of them is a slice of a transaction.
Verification
Verification covers identity, right to work, professional insurance and qualifications. These checks cost money to run. They also shorten a freelancer’s sales cycle measurably, which makes them worth paying for.
Verification is priced close to cost plus a margin. It is paid by whichever side wants the assurance.
This is the cleanest revenue line because the value is easy to understand. A verified profile wins work faster. Nobody has to be persuaded about how that works.
Placement
Placement means paid visibility, labelled as paid visibility. This is the revenue line with the highest risk of corrupting the product. A search result you can buy is a search result nobody trusts.
The discipline is to keep paid slots few, obvious and separate from organic ranking, instead of blending them into it. When a buyer cannot tell which result is paid and which result is earned, the whole marketplace becomes worth less. The revenue from placement does not come close to covering that loss.
Tooling
Tooling covers contracts, invoicing, time tracking, expense handling and tax summaries. This is unglamorous software, but it replaces the three separate subscriptions that freelancers often assemble for themselves.
Tooling is sold as a subscription that can be cancelled without losing marketplace access. That constraint matters. Once tooling becomes mandatory, it becomes a commission with extra steps.
Organisations
Companies that engage many freelancers need approvals, consolidated billing, spend reporting and compliance records. This is the largest single line and the most conventional one.
It is business software, priced per organisation. It competes against spreadsheets and procurement systems, instead of competing against other marketplaces.
The economics change when revenue is tied to services
The arithmetic is different, and it took us a while to internalise it.
A percentage marketplace earns more from a larger transaction at essentially no extra cost. Revenue scales with gross value, which is why every such platform pushes toward bigger jobs and longer engagements.
A services marketplace earns per user, per period, roughly independent of what flows through the marketplace. That changes which work is worth hosting. Small, fast jobs no longer become unprofitable simply because the platform cannot take much from them.
The long tail that a percentage model prices out becomes viable. That matters because the tail is where a great deal of real work lives.
It also changes the enforcement problem. The platform stops caring whether a relationship stays on-platform for the transaction. If the freelancer keeps their tooling subscription and the client keeps their organisation account, both continue paying for what they use. Neither side has a reason to route around anything. The enforcement problem disappears because the incentive to escape disappears.
The hard parts are engineering and discipline
There are three hard parts, and none of them is optional.
- Escrow has to be genuinely cheap. Escrow means the money held until the agreed conditions for release are met. At fifteen percent, you can afford sloppy engineering, manual reconciliation and a support team unpicking edge cases. At zero, you cannot. The escrow has to be a properly modelled state machine, meaning a clearly defined set of states and allowed movements between them. It also needs automated release conditions and payment rails chosen for cost. That is real engineering effort before any revenue exists.
- Disputes have to be designed out. Cheap adjudication is not enough; disputes have to be rare. That puts the weight onto the moment of agreement: explicit scope, defined acceptance criteria and staged milestones, so any disagreement is small and caught early. Every dispute prevented is one nobody pays to resolve.
- The revenue has to be genuinely optional. A user must be able to decline every paid line and still transact fully. That constraint costs money, and it is the only thing preventing gradual drift back to a toll.
Taking a percentage would have made more money sooner
The opportunity cost should be stated plainly, because a decision like this matters only if it cost something.
At a conventional ten percent, our transaction volume would have produced materially more revenue than the four lines currently do. It would also have taken a fraction of the engineering.
The percentage model is more lucrative and dramatically simpler. It needs one integration, one number and no product surface to maintain.
We chose the harder route for a commercial reason. Percentage marketplaces compete on network size. We would lose that competition against incumbents with a decade of head start and enormous acquisition budgets.
We think the better position is to compete on the thing those incumbents structurally cannot copy without destroying their own revenue. That makes the business harder, but it also gives us a route they do not have.
That is the honest version. Percentage fees are not wicked. We could not have won that game, and we might win this one.
Growth is slower without a subsidy
The percentage model has one real advantage that we gave up. We should be precise about it.
A commission platform can spend enormously on acquisition because it knows the lifetime value of a user. That value is a percentage of everything the user will ever earn through the platform. It is a large, predictable number, and it justifies paying a lot to acquire someone.
Our lifetime value is a subscription and some optional services. That is a smaller number, so we cannot outspend anyone.
Practically, growth has to come from the product being obviously better for a specific person, instead of from buying attention. For us, that person is the freelancer who has looked at their annual platform fees and done the arithmetic.
Those freelancers are not a mass market. They are a motivated minority, and they bring their existing clients with them. That is the only acquisition channel we can actually afford.
It is slower. It is also more durable. A user who arrived through a paid channel leaves when a competitor outbids you. A user who arrived because the economics are better only leaves if the economics change.
The common objections are fair
We hear three objections often.
- "You will just add a fee later." This is the most common objection and the most reasonable one, given the history. The only answer is time plus published economics, and we accept that the answer takes years to become convincing. Anyone joining now is taking a position on our future conduct, and they should price that in.
- "Free means low quality." The theory is that a fee filters out unserious participants. There is something to it. Our answer is that filtering should be done by verification and reputation, instead of by a toll on every transaction. A percentage fee filters out cheap jobs, not bad workers, which is the wrong filter.
- "Someone has to pay for disputes." Correct, and this is why designing them out matters more than adjudicating them cheaply. A platform with a dispute rate of half a percent can absorb the cost from general revenue. One at five percent cannot, and would have to charge. The engineering that keeps the rate low is the thing that makes the model viable, which is why it got most of the attention.
A buyer or freelancer can test the model quickly
If you are choosing where to work or hire, four questions cut through the marketing quickly.
- What is the total you will pay over a year at your expected volume? Look at the number, not the headline percentage.
- What happens to that if you and a client work together for three years? Does the platform keep charging for an introduction it made once?
- Can you take your reputation with you? If you cannot, the price will rise eventually, because it can.
- Where does the platform say its money comes from? If the answer is only a transaction percentage, every future product decision will be aimed at increasing transaction value instead of serving you.
The main risk is pressure to add a fee
The model has one obvious failure mode, and we should name it.
If the optional revenue never reaches sufficient scale, the pressure to introduce a small transaction fee becomes enormous. Every argument for that fee will sound reasonable at the time. It is the path every predecessor took.
What we have done about it is structural, rather than a promise. The revenue lines are built as products with their own roadmaps and their own accountability.
That changes the question at a bad board meeting. The question becomes why verification is underperforming, rather than why we do not just charge two percent.
Whether that holds under real pressure is not something we can prove from here. Anyone evaluating the platform should weigh it accordingly.
The same test applies to other intermediaries
The reasoning generalises to any intermediary charging proportionally for services that cost flat.
The test we would apply is simple. If you removed the percentage tomorrow, could the remaining services stand on their own as things people would choose to buy?
If the answer is yes, the percentage was a position rent and it is vulnerable to anyone willing to unbundle it. If the answer is no, the percentage was genuinely funding the service and it is defensible.
Most intermediaries have never had to answer that question because nobody made them. The ones that eventually get asked will find the answer uncomfortable. The ones that answer it themselves first will keep their users without needing a contract to do it.
The engineering behind our version is set out in the Open Lance case study, and the same reasoning about what a fee is really for shows up in how we price our own engagements.
