Marketplace take rate model.
There is no correct take rate, only a structure. Decide who pays, then what each side pays, then look at what you actually realize once every other stream is counted. Free, no sign-up, saved in your browser.
Who this is for
Founders setting a rate for the first time, and founders who suspect the rate they set two years ago is now the wrong one. If you are trying to answer “what should our commission be” with a single number, this will change the question.
How the structure decision works
Three structures, and the choice matters more than the percentage. Charge the supply side and the buyer sees the price the supplier set. Charge the buyer and your supply stays price-competitive with going direct. Charge both and you get predictable margin per transaction, rarely split evenly.
The test that settles it: say the rate out loud as each side would hear it. A supplier hearing “we keep 20% of what you earn” and a buyer hearing “we add 20% to your total” are two different conversations. Whichever sounds worse is the side that will go around you first.
What you get out
Your headline rate against your realized rate, with the gap between them as its own figure. A month-by-month read on which side is gating your transactions, which is what decides whether a fee rise is nearly free or expensive. And a ladder showing what another three, six or nine points would do to GMV and to net revenue, so you can see the trade rather than guess at it.
The reasoning behind all of it is in take rate is a strategy, not a number. The model below is the same one on the traction forecasting page, which frames it around the raise rather than around pricing.
Most marketplace models break for the same reason: founders project demand without modeling supply, or model both without modeling liquidity. This template gives you a cohort-based, two-sided projection across a 24-month horizon, ready for a board meeting or a seed deck.
Fill in the inputs across each section. Numbers update live. The example shown is a fictional marketplace, RigShare, a B2B heavy equipment rental platform, and you can clear it at any time using the topbar. When you are done, save as PDF for your data room.
Your marketplace
What this template covers
Supply cohorts.
Supply growth is the harder side. Most B2B and service marketplaces fail because supply churns silently. They never come back to list new inventory. Model retention honestly here.
Acquisition
Activation and retention
Frequency and CAC
Retention curve preview
The most common modeling mistake is assuming a flat retention rate. Real marketplace retention is a curve: supply churns hardest in months 1 to 3, then stabilizes. If your m1 retention is below 70%, your onboarding is broken. Fix that before forecasting anything.
Demand cohorts.
Demand is loud. Sign-ups feel like progress. But sign-ups without first transactions are vanity. The activation rate is the number that matters, and frequency of use determines whether you build network effects or churn.
Acquisition
Activation and retention
Frequency and CAC
Retention curve preview
For B2B marketplaces, expect 1 to 3 transactions per active buyer per month. For consumer leisure marketplaces (Airbnb-style), expect 0.1 to 0.3. If your number is between those, you have a frequency problem to solve before scaling acquisition.
Liquidity.
Liquidity is what separates a marketplace from a directory. If a buyer searches and gets matched within hours, you have a marketplace. If they get matched after days, or never, you have a listings site that someone is calling a marketplace.
Match metrics
The Cold Start Problem benchmark: match rate above 60% within 24 hours is healthy. Below 40% means demand is leaving for a substitute. The fix is rarely more demand. It is denser supply in the geographies and categories where buyers are actually looking. Solve liquidity before you solve growth.
Revenue model.
There is no correct take rate, only a structure. Decide who pays before you decide what they pay, because the two are not independent: the side you charge is the side that feels the friction, and the side that feels the friction is the side that leaks.
Who pays
Say the rate out loud as each side would hear it. A supplier hearing “we keep 12.0% of what you earn” and a buyer hearing “we add 0.0% to your total” are two different conversations. Whichever one sounds worse is the side that will go around you first.
What each side pays
Everything else you charge for
Most marketplace revenue is not the headline rate. Each of these is computed separately and summed, so you can see what each one is actually worth.
Headline against realized
Which side is the constraint
Supply gates your transactions early and demand gates them from month 6. That means a fee rise costs you far more in the first months than it does later, because every point of retention you lose in that window removes capacity you do not have spare. The same rise after month 6 is close to free, because supply is sitting idle. If you are going to move the rate, moving it after the constraint changes hands is the cheaper trade.
If you raised the supply-side fee
For every point you add to the supply-side fee, assume this many points come off supply retention at months 6 and 12. Leave it at zero if you have no evidence. This is your assumption, not a benchmark. It is stated as yours everywhere it appears, including in the printed model, so nobody reads it as analysis we did.
Unit economics.
The fundamental question for any marketplace: does the take per acquired user pay back the cost of acquiring them, with margin to spare? We model both sides separately because the answer is rarely the same.
Per-transaction economics
Computed unit economics
For a venture-fundable marketplace, target LTV/CAC of 3× or higher within 18 months, with payback under 12 months on at least one side. Pre-seed and seed founders often inflate LTV by assuming retention curves they have not earned yet. Model conservatively, then beat your model.
24-month projection.
The rollup. Active users, transactions, GMV and revenue, projected from your supply and demand cohorts and gated by your match rate. The numbers below are computed live; they will move when you change any input above.
The last row is the one worth reading twice. It names which side is gating transactions that month: the side that runs out first is the side that decides how much volume you get, whatever the other side is doing.
| Metric | M1 | M3 | M6 | M9 | M12 | M15 | M18 | M21 | M24 |
|---|---|---|---|---|---|---|---|---|---|
| Active suppliers | 10 | 36 | 95 | 196 | 379 | 708 | 1,304 | 2,385 | 4,348 |
| Active buyers | 26 | 86 | 206 | 388 | 678 | 1,150 | 1,923 | 3,192 | 5,275 |
| Transactions / month | 24 | 86 | 224 | 422 | 737 | 1,251 | 2,092 | 3,473 | 5,739 |
| GMV / month | $44.4K | $159.1K | $414.4K | $780.7K | $1.4M | $2.3M | $3.9M | $6.4M | $10.6M |
| Net revenue / month | $3.2K | $11.4K | $29.7K | $56K | $97.7K | $165.9K | $277.4K | $460.5K | $761K |
| Constraint | supply | supply | demand | demand | demand | demand | demand | demand | demand |
Cumulative
Sensitivity scenarios.
Three scenarios derived from your base case: pessimistic (70% growth, 80% activation, 85% match), base (your inputs), and optimistic (115% growth, 110% activation, 105% match). The point is not which one is right. It is whether the pessimistic case is still survivable.
| Outcome | Pessimistic | Base | Optimistic |
|---|---|---|---|
| Cumulative GMV | $23.3M | $66.1M | $107.6M |
| Cumulative net take | $1.7M | $4.7M | $7.7M |
| Active buyers (exit month) | 1,743 | 5,275 | 9,026 |
| Active suppliers (exit month) | 1,207 | 4,348 | 8,095 |
| Exit-month MRR | $213.8K | $761K | $1.4M |
| Exit-month ARR (run rate) | $2.6M | $9.1M | $16.4M |
Sophisticated marketplace investors look at the pessimistic case first. They want to know what survives if everything takes 30% longer than you projected. If your pessimistic case still gets you to product-market fit on a reasonable runway, you have a fundable model. If it does not, your base case is too aggressive. Go back and rebuild it.
If you raised the supply-side fee
The scenarios above vary things you influence slowly. This varies the one you can change on Monday. Each rung adds points to the supply-side fee and reruns the whole model.
| Fee change | Rate | GMV | Net revenue | Realized | Supply-bound |
|---|---|---|---|---|---|
| As configured | 12.0% | $66.1M | $4.7M | 12.0% | 5 of 24 |
| +3 pts | 15.0% | $66.1M | $6.7M | 15.0% | 5 of 24 |
| +6 pts | 18.0% | $66.1M | $8.7M | 18.0% | 5 of 24 |
| +9 pts | 21.0% | $66.1M | $10.7M | 21.0% | 5 of 24 |
GMV does not move because you have assumed no retention response to a fee rise. That is the honest default, and it means this ladder is showing you the revenue upside with none of the cost. Set a response above to see the other half of the trade.
Investor summary.
One-page snapshot. Drop this into a board update or include it as a slide in your seed deck.
A model is only as good as its assumptions.
Book thirty minutes with us. We will pressure-test the retention curves, match-rate assumptions and CAC numbers, for free and with no strings, and tell you which inputs we would push back on if we were the lead investor on your round.
Want a copy in your inbox?
Optional. We will email this to the address you enter, and a copy reaches our inbox so we can offer to look at it. You are not added to any list, and everything here works whether or not you use this.