By sector
Marketplace diligence
How to assess a two-sided business: liquidity, take rate, disintermediation and whether the network effect is real or asserted.
Realistic time Six to twelve hours
Marketplaces are harder to diligence than software because the important properties are structural rather than metric. A marketplace either becomes liquid — buyers reliably find what they want, sellers reliably find demand — or it remains a directory with a payment button. The distinction determines everything and it is not visible in a revenue chart.
The vocabulary invites overstatement. Gross merchandise value measures transactions flowing through the platform, not revenue, and a company reporting $40m of GMV may have $2m of actual revenue. Both numbers are legitimate; only one of them is the business.
The other structural question is whether the marketplace can hold on to the transaction once it has introduced the parties. In categories where buyer and seller expect a long relationship, they have every incentive to transact directly after the first introduction, and the platform's revenue depends on preventing something that both sides want.
01
Liquidity — the only thing that matters
Liquidity is the probability that a listing transacts and that a search finds something acceptable. It is measured differently in every category — fill rate, match rate, search-to-transaction conversion, time to first booking — and the specific measure matters less than whether the company has one and tracks it.
The critical insight is that liquidity is local, not global. A marketplace can look healthy in aggregate while being illiquid in every individual city, category or time slot that a real user actually searches within. Ask for the metric broken down by whatever the relevant unit is, and expect the aggregate to look considerably better than the parts.
Repeat behaviour is the confirming signal. Buyers who return without being paid to return, and sellers who keep listing, indicate that the marketplace is delivering something. One-time transactions driven by discounting indicate that it is buying activity.
Check
- What is the liquidity metric, and is it tracked by city, category or segment?
- What proportion of listings transact within a defined period?
- What proportion of searches end in a transaction?
- Repeat rate for buyers and for sellers, separately.
- How many markets are genuinely liquid versus launched?
02
Take rate and unit economics
Take rate is the share of transaction value the platform keeps. It varies enormously by category and the level is less informative than its trajectory and its defensibility. A rate that has been cut to sustain growth is a different signal from one that has been raised without losing supply.
Establish net revenue rather than GMV, and establish what the platform pays for: payment processing, insurance, guarantees, customer support, refunds and incentives. In several well-known categories the contribution margin after all of these is thin or negative, and the aggregate GMV figure conceals it entirely.
Incentive spend deserves separate attention. Discounts and credits used to stimulate both sides are a real cost of the current growth rate, and a marketplace whose activity falls sharply when incentives pause has not yet demonstrated demand.
Check
- GMV and net revenue, clearly separated.
- Take rate by category, and its trend.
- Contribution margin per transaction after all variable costs.
- Incentive and discount spend, and what happens when it pauses.
- Refund, chargeback and dispute rates.
03
Disintermediation and defensibility
In any category where the same buyer and seller will transact repeatedly, both have an incentive to move off-platform after the introduction. Cleaning, tutoring, freelance work and trades are the classic examples. The platform must therefore provide ongoing value beyond matching — payments, guarantees, scheduling, dispute resolution, insurance — or it is charging for something it delivered once.
Ask directly how much leakage the company estimates and how it measures it. Founders who have thought about this have a number and a strategy; founders who have not will describe the problem as unusual in their category, which it rarely is.
Network effects should be interrogated rather than accepted. The question is whether each additional participant makes the platform more valuable to the others, and in what unit. A national network effect claimed for a business that is really a collection of independent city markets is one of the most common overstatements in the category.
Check
- What stops buyer and seller transacting directly after the first match?
- Estimated leakage rate, and how it is measured.
- What ongoing services justify the take rate?
- Is the network effect national, local, or by category?
- What would a well-funded competitor need to replicate this?
04
Supply and demand balance
Every marketplace is constrained on one side and the constrained side is the real business. Establishing which one it is — and whether the company knows — is a fast way to understand how well the founders understand what they are building.
The cost and durability of acquiring the constrained side is then the central economic question. Supply that churns quickly must be continuously reacquired, and that cost is often understated because it sits in operations rather than in marketing.
Check
- Which side is constrained, and does the company agree?
- Cost of acquiring the constrained side, and its churn rate.
- How concentrated is supply? What if the top 10% left?
- How long does a new market take to reach liquidity, and at what cost?
Stop and think
Red flags
- GMV presented as the headline number with net revenue hard to find.
- Liquidity reported only in aggregate, never by city or category.
- Activity that falls sharply whenever incentive spend pauses.
- No answer to the disintermediation question in a repeat-transaction category.
- A national network effect claimed for a business that is a set of local markets.
- Supply concentration where a small number of sellers drive most transactions.
- Negative contribution margin per transaction described as an investment in growth with no path stated.
Take these into the room
Questions to ask the founders
- Which side is constrained, and how do you know?
- What is your liquidity metric, and what is it in your best and worst markets?
- What happens to volume when you turn incentives off?
- How much leakage off-platform do you estimate, and how do you measure it?
- What is contribution margin per transaction after everything variable?
- How long did your most recent market take to reach liquidity?
- What would it cost a competitor to replicate your supply?
From the decoder