How did net revenue evolve after take rate and incentives?
№ 030 · Net revenue waterfall
Definition
MetricNet marketplace revenue after take rate, promotions, incentives and refunds
Unitcurrency, decomposed; net take rate in % of GMV
From GMV down to net revenue: minus payouts to suppliers, minus subsidies, plus other revenue. Reveals how thin the actual marketplace cut is. Build by subtracting payouts and incentives from GMV. Example: $51.8M GMV, -$46.6M to suppliers, -$1.2M subsidies, +$0.6M other = $4.6M net.
- GMV
- Gross merchandise value. Total value of goods sold through the platform before returns, discounts and take rate.
Benchmarks
| Bottom 30% | Median | Top 30% |
|---|---|---|
| net revenue flat or down while GMV grows, meaning incentives, refunds or category mix are consuming the take rate | net revenue growth roughly tracking GMV growth; incentives flat as % of GMV year over year | net revenue growing faster than GMV, with incentives and promos stable or falling as % of GMV; public reference: Grab holding incentives near 10% of on-demand GMV while revenue percentage of GMV holds at 16% in mobility |
consumer and on-demand marketplaces, 2024-2026; the only large public disclosures of incentives as % of GMV are SEA on-demand platforms, so absolute levels are model-specific · Incentive accounting differs: some platforms net incentives against revenue (Grab nets partner incentives), others book them in sales and marketing. State the treatment before comparing. Food delivery blended net take of 13-14% sits well below published 15-30% commission tiers because of subsidies and category mix.
Category split omitted: Net-of-incentive revenue splits by category are disclosed only by individual public companies, not by two independent benchmark datasets.
When it looks bad
The gross-revenue bar grows but the incentive and refund deduction bars grow faster, so the net bar is flat and the gap between the first and last column widens every period.
Gross take $3.0M on $20M GMV (15%), minus $0.9M promos and $0.3M refunds leaves $1.8M net (9%); next quarter GMV is $23M but net is still $1.8M because promos hit $1.4M.
What to do about it
- Split incentives into partner-side and consumer-side lines, mirroring Grab's disclosure, and hold each to a % of GMV budget; the Q4 2025 Grab print shows the discipline is feasible at scale, holding total incentives within 30-40bps YoY.
- Kill always-on sitewide vouchers and move to targeted incentives triggered by lapse risk or first-transaction friction; measure each incentive line by incremental net revenue per subsidy dollar, not GMV lift.
- Introduce merchant-funded promotions and advertising as a counter-bar in the waterfall; Grab's deliveries margin expansion of 35-38bps in 2025 was driven primarily by advertising contribution.
- Report refunds and disputes as their own deduction bar with a named owner, so quality problems stop hiding inside a blended contra-revenue line.
Sources
- Grab Holdings on-demand incentives 10.4% of on-demand GMV in Q4 2025 vs 10.1% a year earlier; partner incentives netted against revenue by definition s205.q4cdn.com ↗
- Grab Holdings incentives flat QoQ at 10.1% of on-demand GMV; mobility revenue held at 16% of mobility GMV grab.com ↗
- PulseRevOps published delivery commission tiers of 15-30% collapse to a blended 13-14% net take after subsidies, promotions and grocery mix pulserevops.com ↗
- Mostly Metrics (CJ Gustafson) public marketplace take rates diverge from headline fee schedules once pass-through and promos are netted; compare net, not list, rates mostlymetrics.com ↗
- Tidemark Capital take rates should be adjusted to remove pass-through revenue before benchmarking; net take is the comparable figure tidemarkcap.com ↗
- Everything Marketplaces (Marketplace Capital) investors convert GMV to net revenue before valuing; present GMV, take rate and net revenue together everythingmarketplaces.com ↗