Unit Economics for Two-Sided Marketplaces

Unit economics for marketplaces is harder than for a one-sided business, because every transaction needs two customers, a buyer and a seller, and each side has its own acquisition cost, its own retention and its own incentives. A marketplace does not usually own the goods, so its revenue is a slice of the money that moves through it, and its profit depends on how big that slice is, how much it costs to process, protect and support each transaction, and how many transactions each acquired user generates. This guide defines the key terms, builds a per-transaction model with checked arithmetic, shows how to treat acquisition cost on both sides without counting the same transaction twice, and explains the strategic issues, liquidity, leakage and incentives, that decide whether the numbers hold. It belongs to unit economics and budgeting and forecasting, follows unit economics for SaaS, and leads to common unit economics mistakes and the introduction in what is unit economics.

The vocabulary

GMV is not revenue. A marketplace with $2 million of GMV and a 12 percent take rate has $240,000 of revenue. Compare the cost base to revenue, not to GMV, and be cautious about reports that celebrate GMV growth without showing take rate and contribution.

The per-transaction model

A hypothetical marketplace for handmade goods (illustrative figures).

Check: 2.62 + 0.32 + 0.80 + 1.20 + 0.40 = 5.34; 9.60 - 5.34 = 4.26. Contribution is 4.26 / 9.60 = 44.4% of net revenue and 4.26 / 80 = 5.3% of GMV.

Two observations. First, the contribution margin as a share of revenue sounds healthy, but as a share of GMV it is only about 5 percent, so a small change in cost per transaction moves the result a lot. A 1 percentage point rise in payment and fraud costs on an $80 order is $0.80, which cuts contribution by 19 percent (0.80 / 4.26). Second, the largest items are payments and marketplace-funded incentives, and both are partly under management control.

Break-even

Assume fixed costs (engineering, product, trust and safety team, general overhead) of $150,000 a month.

Break-even transactions = 150,000 / 4.26 = 35,211.3, so 35,212 transactions a month. At $80 per order that is GMV of 35,212 × 80 = $2,816,960 a month. At that scale revenue is 35,212 × 9.60 = $338,035. The marketplace must move about $2.8 million of GMV each month before contribution covers the fixed costs, before any acquisition spend. See break-even analysis.

Two-sided acquisition cost

Each side has its own CAC. Suppose:

• Buyer CAC: $20, with an average of 5 transactions in the first year.
• Seller CAC: $60 (outreach, onboarding, listing support), with an average of 30 transactions in the first year.

A naive view says buyers generate 5 × 4.26 = $21.30 of contribution against $20 of CAC, a small profit, and sellers generate 30 × 4.26 = $127.80 against $60 of CAC, a large profit. But the same transaction is being counted on both sides. A transaction is one event involving one buyer and one seller, and its $4.26 of contribution can only be counted once.

A simple way to handle it is to split each transaction's contribution between the two sides, for example 50/50:

Check: 5 × 4.26 × 0.5 = 10.65; 30 × 4.26 × 0.5 = 63.90.

On this allocation, buyers do not pay back in a year, and sellers only just do. The split is a judgment, and a different split gives different answers, but the lesson holds under any split: the two sides share the same stream of contribution, so you must judge the combined system. The better test is to add up all acquisition spend on both sides in a period and compare it to the contribution earned from those users' transactions in the following twelve months.

Retention and repeat behavior

Marketplace retention comes in two forms:

• Buyer retention: do buyers come back? A buyer who transacts once and leaves has cost $20 to acquire and returned $4.26.
• Seller retention: do sellers stay active, list more and respond well? Seller quality affects buyer experience.

Build cohort tables for both sides. Track not only the share still active, but also transactions per active user, because a user who stays but buys less is a weaker cohort. See cohort analysis and customer retention rate.

Liquidity decides whether the numbers hold

The model above assumes the transactions occur. In a marketplace, they occur only if the right buyer meets the right seller at the right time. If supply is thin, buyers fail to find what they want and leave; if demand is thin, sellers list items that never sell and leave too. Useful liquidity measures include the share of searches that end in a transaction, the share of listings that sell within a set time, and the time to first transaction for new users.

Better liquidity raises transactions per user, which improves payback on both sides, so spending to improve liquidity in a focused segment can have a higher return than spending to grow everywhere. Many marketplaces grow by concentrating on one category or city until liquidity is good, then expanding.

Leakage

Leakage happens when a buyer and seller meet on the marketplace and then complete future transactions off it, to avoid fees. It is a risk for services marketplaces and for categories with repeat relationships. The unit economics suffer, because the marketplace paid to acquire both sides and then lost the later transactions. Responses include:

• Adding value that is only available on the platform: payment protection, insurance, reviews, scheduling, financing.
• Pricing the take rate so that going off-platform does not save much.
• Lowering the fee on repeat transactions between the same parties.
• Monitoring signs of leakage, such as a high first-transaction rate with low repeat rate for the same pair.

Pricing the take rate

The take rate is a trade-off. A higher rate improves contribution but can discourage sellers and encourage leakage. A lower rate attracts supply but may not cover costs. Test using contribution per transaction as the guide, not revenue. In the model above, raising the buyer fee from 2 percent to 3 percent adds $0.80 of revenue per $80 order, a 19 percent increase in contribution (0.80 / 4.26), if volume holds. If volume falls by 10 percent as a result, contribution per month falls by 10 percent but rises per transaction, so total contribution is 0.9 × (4.26 + 0.80) = 0.9 × 5.06 = $4.55 per original transaction, up 6.9 percent from $4.26. The test shows that a modest fee increase can pay for itself even with some volume loss, but the sensitivity to volume is high: at a 20 percent fall in volume the total would be 0.8 × 5.06 = $4.05, below the starting point. See pricing analytics.

Incentives and subsidies

Marketplaces often subsidize one side to build the other. Discounts for buyers, guarantees for sellers, and referral bonuses all count as acquisition or variable cost. Track them explicitly, not buried in marketing. A marketplace whose contribution per transaction is positive only because incentives are recorded outside the model is not yet profitable. Ask whether the business would still work if the incentives ended, and when they will.

Key metrics to report

Common mistakes

• Reporting GMV as revenue.
• Counting each transaction's contribution on both buyer and seller cohorts.
• Ignoring marketplace-funded incentives, refunds and credits.
• Using blended CAC across sides when the sides differ widely.
• Assuming liquidity will arrive without a focused plan.
• Setting take rate without watching leakage.
• Treating fraud and chargebacks as small when they are a large share of a thin margin.

FAQs

What are marketplace unit economics?

They are the revenue, variable cost and contribution of a single transaction, together with the cost of acquiring and retaining buyers and sellers.

What is a good take rate for a marketplace?

It varies widely by category, and no single figure applies. The right rate is one that leaves positive contribution after payments, support and incentives without driving sellers away.

How do you calculate CAC for a two-sided marketplace?

Calculate it separately for buyers and sellers, then judge the combined spending against the shared stream of transaction contribution, taking care not to count each transaction twice.

What is the difference between GMV and revenue?

GMV is the total value of transactions on the platform, while revenue is the portion the marketplace keeps as commission and fees.

Why is liquidity important to marketplace unit economics?

Because contribution only arrives when buyers and sellers complete transactions, so poor liquidity lowers transactions per user and delays payback on acquisition spend.