How Does a Marketplace Make Money?

Marketplaces make money by taxing the transaction loop - most often a commission (take rate) on each booking or order, sometimes joined by subscriptions, listing fees, lead charges or advertising. The software is the meter; the business model is what you charge for matching, trust and demand. Across seven marketplaces we operate, the model that launches cleanest is usually a simple, visible take rate; second streams arrive once liquidity is real enough that sellers will pay for reach or tools.

The five marketplace revenue models

Model	How it charges	Works when	Breaks when
Commission / take rate	% (or flat fee) per transaction	Clear GMV, repeat orders, sellers see value	Ticket too small after fees; leakage high
Subscription	Monthly/annual for sellers (or buyers)	Ongoing tools, CRM-ish value, predictable access	Liquidity weak - people won’t subscribe to an empty room
Listing / insertion fees	Pay to publish or feature	High intent catalogue; scarce attention	Cold start - fees before demand feel extractive
Lead generation	Charge for enquiries or qualified leads	Quote-driven verticals; offline close	Lead quality disputes destroy trust
Advertising	Sponsored placements, promoted listings	Dense traffic and fair auction rules	Thin traffic; ads on emptiness look desperate

Most consumer-looking marketplaces lead with commission. B2B and high-ticket services often lean lead-gen or subscription because the “order” happens off-platform. Hybrid is normal once you have volume - see below.

Payment rails (for example Stripe Connect for marketplaces) collect the money; they do not choose the model. Niche focus improves willingness to pay - what makes a niche marketplace work.

Take rates in the real world: what verticals will bear

What is a good take rate for a marketplace?

Published industry commentary often clusters consumer marketplace commissions roughly in the 10-20% band, with high-ticket B2B and travel-like verticals sometimes lower on percentage but healthier on absolute margin - and some service marketplaces higher when they replace expensive offline brokers. “Good” means sellers still win after the cut and your contribution margin survives payment fees, support and refunds. A 15% take on a £20 basket with heavy support can lose money; an 8% take on a £2,000 B2B job can work.

Operator framing from our portfolio (directional, not a published rate card for each brand):

• Aesthetics and appointment-led verticals care about calendar fill and trust; fee tolerance rises when the platform clearly drives bookings
• Vehicle rental cares about utilisation; percentage of rental value must leave owners ahead of peer-to-peer Facebook risk
• Sponsorship and package marketplaces often behave like catalogue-plus-lead; pure GMV commission may underfit how deals close
• Funeral and high-care verticals punish fee surprise; clarity beats cleverness
• Pub/venue event verticals live on timing accuracy; monetisation fails if fixture data is wrong, regardless of rate

Always model card fees + Connect costs + support minutes per order beside the headline percentage. Unit economics section below is where survival is decided.

Why most marketplaces mix models (and when to add a second one)

One fee keeps onboarding honest. Two fees too early look like rent-seeking on an empty grid.

Add a second stream when:

• Organic demand exists and sellers ask for more visibility (promoted listings)
• You ship seller tools worth a subscription on their own (CRM, analytics, multi-location)
• Offline close is the norm and a lead fee matches how the trade already thinks
• High-volume free listings create noise and a small insertion fee raises quality

Wait when:

• Supply is still hand-recruited
• You cannot explain the first fee in one sentence
• Ads would outrank organic results sellers thought they earned

Our experience: subscription-before-liquidity is how you fund churn. Commission-first, tools-second is how sellers stay.

The unit economics that decide survival

Are online marketplaces profitable?

Yes when contribution margin after payment costs, refunds and variable support stays positive and you can acquire both sides at a cost that payback allows. No when GMV vanity hides negative orders, or when leakage pushes real trade off-platform. Profitability is an ops and economics problem that wears a product costume.

Minimum sketch before you scale ads:

1. Average order value and take rate → revenue per order
2. Minus payment and payout costs
3. Minus variable support (minutes × cost)
4. Minus refund/chargeback allowance
5. Compare to CAC on buyer and seller sides (blended thoughtfully)
6. Check repeat rate - one-off categories need much higher take or ticket

If step 5 only works at fantasy conversion rates, fix the niche or the model before you hire a growth lead. Build cost context: marketplace build cost. Failure modes: marketplace disadvantages.

What our seven marketplaces taught us about which model to launch with

Launch with the simplest fee a seller can repeat back to you. For most of our verticals that meant a transparent transaction cut, collected in-platform, with fee copy at signup - not a PDF buried in help centre. Add subscriptions or promotion only when sellers feel the absence of those tools in their week.

Secondary lessons we reuse on client builds:

• High-care, low-frequency verticals need fee clarity more than fee cleverness
• Calendar and inventory verticals monetise utilisation; percentage of nothing is nothing
• Package and sponsorship shapes may need milestone or lead pricing because GMV is ambiguous until a deal closes
• Never raise take rate to fix a support cost problem you have not measured - fix the queue first

After 50+ marketplace builds over 15 years, and seven we run ourselves, the commercial question is always the same: what will a good seller pay for demand and trust they cannot get cheaper elsewhere?

Commission first, second streams later: that is the fee order that survives empty-grid weeks. Pressure-test your model inside marketplace development.