Markets and business models
GMV and take rate
GMV is dollars transacted. Net revenue is GMV times the take. Never compare a GMV number to SaaS ARR.
A marketplace connects two or more distinct groups and takes a cut of the value exchanged. Airbnb connects guests and hosts. Stripe Atlas is not a marketplace — it is SaaS. The economic object you keep is net revenue, not the dollars that pass through.
Concept
GMV (gross merchandise value) is the total dollars of goods or services transacted. Net revenue is what the company keeps: roughly GMV × take rate minus refunds and credits. A $1B GMV marketplace at a 10% take rate is a ~$100M revenue business.
Net revenue ≈ GMV × take rateWalking through take-rate bands
Take rates vary by category (illustrative ranges, not laws):
| Category | Rough take-rate band | Why |
|---|---|---|
| High-touch services / some B2B | ~15–30%+ | High willingness to pay for match quality |
| Ride-hail / delivery | ~15–30% to consumer side dynamics | Ops heavy; incentives muddy the "take" |
| Travel lodging | ~10–20% combined host/guest fees | Habit + trust; competition on rake |
| Digital goods / some software markets | ~15–30% | App stores historically ~30% |
| Low-friction commoditized goods | ~5–15% | Easy to disintermediate; price sensitive |
What matters is take relative to value added. If you only provide a thin listing page, a 25% rake invites bypass. If you provide trust, payments, insurance, discovery and ops, a higher rake can stick.
Category GMV of $500M/year, take rate 12% → net revenue = $60M. If incentives/promos cost $15M, "effective take" is closer to ($60M − $15M)/$500M = 9%. Boards eventually look through promotional GMV to contribution after incentives.
Marketplace gross margin should be compared on net revenue, not GMV. An 80% margin on take rate with a 15% take rate is a different animal from 80% margin SaaS on the full subscription.
If you take 15% of a $100 GMV order, net revenue is $15. Payment processing might be ~$0.30 + 2.9% of $100 ≈ $3.20 if you process the whole payment — brutal if booked against $15.
A household-level sketch
Home-services marketplace:
AOV = $280
Jobs per year per active household = 3
Take rate = 18%
Net revenue per active household / year = 3 × $280 × 0.18 = $151.20If CAC is $90 and year-1 retention of transacting households is 50%, expected year-1 net revenue ≈ $151 × 0.75 (mid-year join) ≈ $113. Minus incentives $20 → ≈ $93 against $90 CAC — barely payback in year 1. Year 2+ retained households (no CAC) are where the business prints money.
Common mistakes
- Pitching GMV as if it were ARR.
- Ignoring incentives so the take looks 12% when effective take is 9%.
- Comparing marketplace GM on GMV to SaaS GM on subscription revenue.
Check yourself
- $1B GMV at a 10% take. What is net revenue? About $100M.
- $60M take minus $15M incentives on $500M GMV. Effective take? 9%.
Key takeaways
- Marketplace companies live on net revenue (
GMV × take), not GMV vanity. - Take must match the value added, or participants bypass.
- Effective take subtracts incentives.
- Compare margins on net revenue, not GMV.
Further learning
Why take is useless without matches: marketplace liquidity.
Related
Marketplace economics
Take rates, liquidity, chicken-and-egg dynamics and why some marketplaces become giants.
Business models
The main ways software companies make money, and the economics each implies for margin, sales motion and venture scale.
Gross margin and COGS, including AI
What belongs in COGS vs opex, why classic SaaS runs ~75–80% gross margin, why AI products often sit at 50–60%, and the levers that move the number.