How to Monetise a Two-Sided Marketplace

What Is the Right Way to Monetise a Marketplace?

Charge a percentage of each completed transaction, collected inside the payment flow. It aligns everyone: the platform earns only when a match succeeds, suppliers pay only on revenue, and buyers pay nothing visible. Every other model — subscriptions, listing fees, lead fees, featured placement — is either a supplement to the take rate or a substitute for categories where the take rate structurally fails.

The discipline is sequencing. A marketplace that cannot yet reliably match buyers and suppliers has nothing to charge for, and a fee on a weak loop actively teaches both sides to transact elsewhere. Monetisation follows liquidity — the thresholds worth using are the fill-rate and repeat-rate measures from the first-100-users playbook.

What Are the Six Marketplace Revenue Models?

Six models cover essentially every monetised marketplace, and mature platforms usually stack two or three. Each has a structural condition under which it works — the table is the map.

What Do Real Marketplaces Actually Charge?

The published numbers are worth studying because they map fee to role. Etsy's fee policy stacks a $0.20 listing fee, a 6.5 percent transaction fee and roughly 3 percent plus $0.25 payment processing — a goods marketplace whose sellers depend on its demand. Airbnb's help centre documents its move to a single host-only service fee of 15.5 percent of the booking subtotal, replacing the old split of roughly 3 percent host and 14–16 percent guest — at the time of writing, one of the clearest examples of restructuring who visibly pays.

Services platforms sit higher still: Fiverr charges sellers a flat 20 percent plus a 5.5 percent buyer service fee per FreelanceCompare's breakdown, while Upwork's own support documentation describes a variable 0–15 percent freelancer service fee set per contract. The pattern behind the spread matches a16z's observation that managed marketplaces command the top of the take-rate range: the more matching, trust and infrastructure the platform provides, the more of the transaction it can keep.

What Take Rate Should You Set?

Set it off three questions: how much value you add to the match, how fragmented each side is, and what the transaction's margin can bear. High platform value, fragmented sides and healthy margins support double-digit rates; thin-margin or relationship-heavy niches do not, however good the software is.

Then start below your ceiling. A new marketplace earns pricing power with liquidity, and the cost of opening low is small compared with the cost of teaching early users to circumvent you. Raising a rate from 8 to 12 percent on a thriving loop is a communication exercise; recovering users lost to a premature 20 percent is a relaunch.

Does payment flow through the platform?
  NO  -> commission is unenforceable
         -> listing fees, lead fees or subscriptions
  YES -> are transactions rare and very large?
          YES -> flat or capped fees beat percentages
          NO  -> commission. Set the rate:
                  platform adds real matching + trust value?
                  YES -> low-to-mid teens, grow with proof
                  NO  -> single digits until the value exists

When Should You Switch Monetisation On?

After liquidity, before habit. Switch the fee on once the loop clears its thresholds — solid fill rate, visible repeat transactions — and while the platform's contribution is still obvious to both sides. Waiting too long is its own mistake: users who transacted free for a year treat any fee as a betrayal, while users who paid 5 percent from month three barely notice 8 in year two.

Grandfather the pioneers. Founding suppliers who carried the empty months deserve a permanently discounted rate, publicly promised — it costs little at their volume and buys durable loyalty plus a recruiting story for the next cohort. The same founding-cohort logic runs through job board monetisation, and it works because it prices contribution, not just usage.

How Do You Implement Marketplace Fees Technically?

Collect the fee inside the payment flow, never as an invoice after the fact. Platform payment infrastructure supports this natively — Stripe Connect's published model lets the platform add an application fee to each charge while routing the remainder to the supplier, with Express accounts priced at $2 per monthly active account plus 0.25 percent and $0.25 per payout at the time of writing. The fee that collects itself is the only fee that reliably gets paid.

Subscriptions and featured placement reuse the same billing rails as any SaaS feature, which is why they arrive as clean later phases on a well-built MVP. The payment plumbing is standard payment gateway integration work, and inside a two-sided marketplace MVP build the application-fee wiring ships with version one so the take rate is a configuration decision, not a rebuild.

How Do You Stop Users Going Around the Platform?

Disintermediation — meeting on the platform, transacting off it — is the tax every commission model pays, and policing alone never beats it. The durable answer is value inside the paid transaction: payment protection and escrow, verified reviews that only completed transactions earn, dispute resolution, insurance, guaranteed payouts. Users stay on-platform when leaving costs them something real.

Design nudges help at the margin: fees capped per transaction for large deals, rates that fall with volume or loyalty, and messaging that keeps contact details private until commitment. But treat rising leakage as information, not crime — it is the clearest signal your take rate exceeds your added value, and the fix is usually on the value side.

When Should a Marketplace Not Charge Commission?

Three structures break the commission model. Off-platform completion: where deals close in the physical world — property, heavy machinery, recruitment-to-hire — the platform cannot see the transaction, so lead fees, listing fees or subscriptions do the work. Huge, rare transactions: a percentage of a house sale is politically impossible; flat or capped fees fit. And strategic free sides: some platforms deliberately keep one side free forever and monetise the other harder — the standard shape in hiring markets.

There is also the do-not-monetise-yet case: a marketplace still below liquidity thresholds should be spending its energy on matching, not billing. Revenue models are easy to add to a working loop and impossible to extract from a broken one.

What Monetisation Mistakes Kill Marketplaces?

The recurring failures are structural choices rather than execution slips, which makes them visible in advance and cheap to avoid. Each of the six below traces back to the same root: charging, or building billing machinery, ahead of the evidence that the loop deserves it — the pricing equivalent of building platform features before liquidity.

• 1. Charging before liquidity — a fee on an unreliable loop pushes the few working transactions off-platform.
• 2. Copying a big platform's rate without its value — 20 percent needs matching, trust and demand generation behind it.
• 3. Collecting by invoice instead of inside the payment — uncollected commissions are discounts you did not choose.
• 4. Monetising both sides at once from day one — visible fees on buyers and suppliers simultaneously doubles the friction.
• 5. Punishing leakage instead of pricing it — bans and threats drive users away; escrow, protection and reviews pull them back.
• 6. No founding-cohort deal — repricing your earliest suppliers at standard rates burns the loyalty the cold start bought.

PINCLER's Perspective on Marketplace Monetisation

PINCLER is an AI-first custom software development studio, and the monetisation lesson from our catalogue is about sequencing scope: across PINCLER's 79 documented projects — every one fixed-price between $500 and $2,500, median $1,450 and 13 days — the marketplace builds that work ship the application-fee wiring in version one and leave subscriptions, featured placement and volume pricing as later phases, each its own fixed-price build once the loop has earned them.

That order holds because billing features built ahead of demand get rebuilt once real behaviour arrives — the same pattern we see across two-audience products generally. Adding a billing model to an existing platform sits near our integrations median of $1,200 and 10 days, which makes waiting for evidence cheap. The dataset behind those numbers is public at what you can build.

Bottom Line

Monetising a marketplace is commission-first: a take rate collected inside the payment flow, opened at the low end of your category's range and raised as liquidity proves the platform's worth, with subscriptions and placement stacked later and structural exceptions priced by listing or lead instead. Keep users on-platform with value, not rules, and let the founding cohort keep their discount forever. When you are ready to wire it, a free 30-minute call turns the model into a fixed-price build plan.