a founder pitches a marketplace at a series a meeting in 2026. the headline slide reads "$50m in annualized gmv, up 4x year over year." the partner nods politely and asks one follow-up. "and what's the take rate?" the founder pauses, says "about 3%," and adds "we have room to expand it." the partner writes one number on her pad — $1.5m — and the conversation is effectively over.
gmv is the dollars flowing through the platform. it's not revenue. it's the potential revenue you could earn if you could take a cut, and the size of the cut you actually take decides whether you're a small saas company or a real business. founders lead with gmv because it's the biggest number on the slide. investors discount gmv to zero in fifteen seconds because they've watched twenty marketplaces with huge gmv numbers fail to ever build a defensible take rate.
the question that decides a marketplace round in 2026 is not how much volume is flowing. it's how much value the platform actually captures, and whether the capture rate is going up or down.
what take rate looks like across the references
every founder pitching a marketplace should be able to name their take rate, their direction over the last four quarters, and the comparable benchmark for their category. without those three numbers, the partner is doing arithmetic in their head and arriving at a discount.
the spread is enormous. stripe at 3% is a different business from uber at 25% even at identical gmv — uber captures eight times the revenue per dollar of platform volume. that's why a $50m gmv platform at stripe's take rate is a small business and the same gmv at uber's take rate is a real one. category gravity matters, but the band within a category also varies. higher take rate means more value captured per transaction, and also more risk that buyers or sellers route around the platform once they meet. the question every partner is running in their head: is this take rate sustainable, or is it the pre-disintermediation peak.
the comparison that decides the round
picture two marketplaces, both at $20m in gmv, pitching the same partner the same week.
marketplace a runs a 4% take rate today. revenue is $800k. last year it was 3.5%. the year before it was 3%. the trajectory is up because the platform has launched a payments product, a fraud-protection layer, and a buyer-protection guarantee sellers will pay for. at the same gmv next year with take at 5%, revenue is $1m. management is investing in features that justify higher capture.
marketplace b runs a 12% take rate today. revenue is $2.4m. last year it was 14%. the year before it was 16%. sellers started routing around the platform after they met the buyers, and competitive marketplaces in the same category run 8%. at $20m gmv next year with take trending toward 8%, revenue is $1.6m. management is fighting compression.
at the same gmv and three times the absolute take rate today, marketplace b is worth less than marketplace a in two years. the direction matters more than the level. the partner who has lived through one disintermediation event will pick marketplace a inside the first meeting.
the direction of the take rate is what tells you whether the marketplace is becoming more valuable to its participants or whether they're shopping for an exit.
the three things investors actually check
what serious marketplace diligence runs through:
take rate trajectory by cohort. for a given cohort of sellers (or buyers), what was the take rate in their first quarter on the platform, and what is it now? if it's gone up, the platform is adding services they pay for. if it's gone down, they're learning how to do less business with you. cohorted take rate exposes disintermediation faster than any other metric.
organic versus paid gmv. what percentage of the gmv flowing through the platform was acquired by paid marketing in the same quarter? if it's over 60%, the marketplace is a paid-acquisition channel pretending to be a marketplace. real marketplaces compound on organic — referrals, search, repeat usage. if you cut paid spend by 50% next quarter, what happens to gmv? the answer in the founder's head is the answer the partner cares about.
buyer-seller retention by cohort. the buyer who came in q1 of 2024 — are they still transacting in q1 of 2026? the seller who joined a year ago — what percentage of their volume still runs through you versus side-channels? marketplace retention is harder to measure than saas retention because there's no subscription, but the cohort math is the same shape.
what to do with this if you're the founder
stop leading with gmv on the deck. lead with revenue, take rate, and the four-quarter take-rate trend. add the cohort retention number for the largest seller cohort. if those numbers are strong, gmv is a confirmation slide. if they're weak, no amount of gmv saves the pitch.
the founders raising marketplace rounds in 2026 who are clearing process are the ones who walk in with the take-rate trend chart on slide three and the disintermediation question pre-answered. "here's our take rate by cohort, here's how we hold it, here's why it's going up next year." the partner has nothing to discount because the founder already did the discount themselves and arrived at the honest number.
how zift handles this
zift computes take rate by cohort, by month, against the stripe and bank data it already ingests. on monday morning the briefing names the cohort whose take rate moved, the dollar amount, and which side of the marketplace drove it. the board-pack version of the same chart sits in the export without anyone having to assemble it.
if you're a finance lead at a marketplace running multi-region or multi-category consolidation, zift handles that too.
gmv is what you tell people at a dinner party. take rate is what the term sheet is priced against.