Choose marketplace pricing by three checks: who receives measurable value, how the transaction actually closes, and whether contribution margin survives payment costs, support, refunds and disputes. Get those three right and the fee shape follows. Get them wrong and a tidy percentage still loses money on every order.

This guide is about choosing, calculating and testing a pricing strategy. For a survey of revenue models in the round, start with how marketplaces make money. If you are still proving demand, pair this with how to validate a marketplace idea before you hard-code complex fees into product.

Fee models you can choose
GMV, take rate, revenue per order and contribution margin
Illustrative worked example
Decision framework
Competition
Seller margins and your marginal costs
Vendor differentiation
Network effects and off-platform leakage
Transaction size and volume
Quality, quantity and support burden
Who should pay
Test willingness to pay before you ship fees
Pricing transparency at signup and checkout
Conclusion
Next step
References

Fee models you can choose

Most marketplace pricing strategies use one lead model, then add a second stream only when liquidity and seller value are real. The common options:

Seller commission / take rate. You take a percentage (or a flat amount) of each completed order or booking. Simple to explain when the platform clearly closes the trade.
Buyer or service fee. The buyer pays a fee on top of, or inside, the checkout total. Useful when sellers are margin-sensitive and buyers still see clear matching or trust value.
Subscriptions. Sellers (sometimes buyers) pay monthly or annually for access, tools or tiered features. Works when ongoing product value exists even between orders.
Listing or insertion fees. A fee to publish or renew a listing. Raises catalogue quality when demand exists; feels extractive during a cold start.
Lead fees. You charge for enquiries or qualified leads when the deal often closes offline. Common in quote-driven services.
Value-added services / promoted placement. Optional extras such as promoted listings, fulfilment aids or premium support. Keep the base fee understandable, then let high-intent sellers buy reach.
Hybrid models. Combinations such as a modest take rate plus promoted placement, or a low subscription plus a per-order fee. Hybrid is normal later; stacking fees too early looks like rent on an empty room.

Payment rails collect money; they do not choose your commercial model. Stripe Connect, for example, supports application fees on Connect charges and related patterns such as destination charges with application fees. That plumbing should match the fee you already decided, not invent it. See Stripe Connect for marketplaces for implementation context.

GMV, take rate, revenue per order and contribution margin

Founders often debate the headline percentage and skip the maths that decides survival. Use these four terms in plain English:

Gross merchandise value (GMV). The total value of goods or services traded through the marketplace over a period, before your cut. Useful for scale; dangerous as a vanity metric on its own.
Take rate. Your fee as a share of GMV (or of the completed order). A 12% take rate on £100 of GMV is £12 of marketplace revenue before costs.
Revenue per order. What you actually earn on a typical completed order after combining percentage fees, flat fees and any buyer-side charges. This is the number support and payment costs have to fit under.
Contribution margin. Revenue per order minus the variable costs of that order: payment and payout fees, expected refunds and chargebacks, and variable support time. If contribution margin is negative, volume makes the hole bigger.

A pricing strategy is only finished when contribution margin is credible on realistic assumptions, not when the take rate looks tidy in a slide deck.

Illustrative worked example

Illustrative only. This is a teaching calculation, not a benchmark, forecast or promise for any vertical.

Assume a niche product marketplace with:

Average order value: £80
Seller commission (take rate): 12%
No separate buyer fee
Payment and payout costs: 2.0% of order value + £0.20
Expected refunds and disputes: 1.5% of order value (averaged across orders)
Variable support: 8 minutes at £0.45 per minute (£3.60)

Arithmetic:

Marketplace revenue per order = 12% × £80 = £9.60
Payment and payout costs = (2.0% × £80) + £0.20 = £1.60 + £0.20 = £1.80
Refund and dispute allowance = 1.5% × £80 = £1.20
Variable support = £3.60
Contribution margin = £9.60 − £1.80 − £1.20 − £3.60 = £3.00

At £3.00 contribution per order, growth only helps if acquisition cost and repeat rate leave room after that £3.00. Raise support minutes, refunds or payment costs and the same 12% can turn negative without anyone changing the published fee. That is why you model contribution margin before you celebrate GMV.

Decision framework

Use the questions below to choose who to charge, what they receive, when they pay, and how hard it will be to change later. Work through them in order; do not jump straight to a favourite percentage.

Who receives measurable value? Matching, demand, trust, payments, dispute handling, tools or all of the above.
Who can pay without killing the loop? Sellers, buyers, or both in a split that still converts.
When should they pay? Per transaction, per lead, per listing, or on a subscription cadence.
How easily can trade leak off-platform? If parties can finish outside your rails, pure take rate is fragile.
What seller margin remains after your fee? Thin vendor margins force lower take rates or buyer-side fees.
What are transaction frequency and value? Low ticket with high support needs a different shape from high ticket with rare orders.
What support burden does each order create? Minutes per order belong in the model, not in hope.
How will you test price changes? Cohort tests, introductory windows with a clear end date, and fee comprehension checks before a permanent cutover.

The sections that follow expand the classic pressures that move those answers.

Competition

If vendors have no credible alternative channel, fee tolerance rises. Niche focus helps here: a narrow market is often easier to serve well and harder to replace. Where a stronger marketplace already owns demand, you usually need clearer value or a simpler, lower-friction fee to win a trial. Cheaper alone rarely beats denser demand; cheaper plus a sharper job-to-be-done can.

Price against the value you create relative to the next best option, not against a rumour of someone else's rate card.

Seller margins and your marginal costs

Vendor margin sets a ceiling on seller commission. If sellers already operate on thin contribution, a heavy take rate pushes them off-platform or out of the category. Where seller margins are healthier, a higher take rate can still leave them ahead of acquiring demand alone.

Your own marginal costs matter just as much. Payment fees, payouts, messaging, moderation and dispute handling all scale with volume. Category-specific costs (fragile goods, regulated services, high-touch bookings) may justify a sliding scale by product type rather than one flat percentage everywhere.

Vendor differentiation

Not every seller creates the same value on the platform. High-volume or high-quality vendors may deserve better economics, priority support or free access to tools that casual sellers pay for. That is still a pricing decision: you are allocating margin to keep the sellers who make the marketplace worth visiting.

Optional paid upgrades (promoted placement, analytics packs, multi-location tools) let you keep the base fee simple while giving ambitious sellers a way to buy more outcomes. Keep upgrades optional and clearly valuable; do not bury mandatory costs inside “premium” language.

Network effects and off-platform leakage

Stronger networks (more relevant supply, more repeat demand, better trust signals) usually support healthier fees because both sides lose more by leaving. Weak networks punish aggressive monetisation: people meet once, then finish the next deal elsewhere.

Design pricing with leakage in mind. If quotes, contact details or fulfilment naturally move offline, consider lead fees, subscriptions for tools, or payments incentives that make staying on-platform rational. A pure take rate on a leaky loop trains users to bypass you.

Transaction size and volume

Twenty percent of £20 and twenty percent of £2,000 are different businesses. Low-value, high-frequency orders need fees that survive payment costs and still feel fair on a small ticket. High-value, low-frequency orders can often carry a lower percentage with healthier absolute revenue per order, provided disputes are controlled.

Build a simple forecast from expected order value, orders per active seller, take rate and contribution margin. If the only way the model works is fantasy conversion, change the niche, the fee shape or the support model before you scale acquisition. Build-cost context lives in how much it costs to build an online marketplace.

Quality, quantity and support burden

You serve sellers and buyers. More verified quality (credential checks, clearer listings, fair dispute handling) can justify higher fees because both sides pay for reduced risk. A pure volume land-grab usually needs lower early fees and higher support capacity.

Support minutes are part of pricing. Categories with complex fulfilment, scheduling or complaints need either higher revenue per order or tighter product rules. Product design choices (clear status flows, fewer ambiguous listings, better messaging) can cut support cost as effectively as raising the take rate. That is where product design and pricing strategy meet.

Who should pay

Charge the side that receives the clearest, most measurable value and can pay without collapsing conversion. Many consumer marketplaces lean on seller commission because sellers recognise paid demand. Some service marketplaces split fees when buyers also value trust and convenience. Lead-led verticals often charge the professional who receives the enquiry.

Perception matters. Even when sellers fund the fee, buyers notice if the final price feels padded without explanation. Even when buyers fund a service fee, sellers notice if demand quality does not justify staying. Make the value story match the invoice.

Test willingness to pay and fee comprehension before you hard-code complexity

Do not ship a multi-tier hybrid fee schedule because a spreadsheet can express it. Founders should test two things early:

Willingness to pay. Can target sellers or buyers explain why the fee is worth it in their own words? Run interviews, concierge transactions or a manual billing pilot before engineering every edge case.
Fee comprehension. Can a new user predict what they will pay after one pass through signup and checkout? If people cannot estimate the fee, they will distrust the marketplace even when the rate is fair.

Prefer a simple first fee you can explain in one sentence. Add subscriptions, promoted placement or lead products only when users ask for the outcome those fees buy. When you change prices later, test on a cohort, state the change clearly, and avoid silent compounding of small fees.

Pricing transparency at signup and checkout

Show the fee shape where people decide: seller signup, listing publish flows and buyer checkout. Hidden mandatory charges taught late in the journey destroy trust and create support load. UK consumer law also expects clear pricing practices for traders; the Competition and Markets Authority publishes price transparency guidance (CMA209) on presenting total prices and avoiding drip pricing, alongside broader unfair commercial practices guidance (CMA207). This article is practical product guidance, not legal or tax advice. Get counsel for your exact model and jurisdiction.

Practical rules of thumb for product:

State who pays what before the irreversible commit
Show examples with real order values, not only percentages
Keep optional upgrades clearly optional
Align the fee copy in marketing, onboarding emails and checkout totals

Conclusion

Marketplace pricing is a strategy problem first and a percentage second. Pick the model that matches who gets value and how trade closes, then prove contribution margin under honest payment, support and refund assumptions. Keep the first fee simple enough to test. Make the cost obvious at signup and checkout. Change it deliberately when evidence says you should, not because a competitor rumour spooked you.

If you need the wider model map, return to how marketplaces make money. If you are scoping the product itself, see marketplace development.

Next step

If you want help turning a fee model into a buildable marketplace product, contact Code23. We have built under the Code23 brand since 2005.

Marketplace Pricing Strategy: Choose, Calculate & Test Fees

Contents