Marketplace questions trip up strong candidates more often than any other interview category. The reason has structure. A marketplace runs on two sides at once, with paired customers, paired economics, paired risk profiles, and paired growth loops. Miss one side, fail the question.
Why marketplaces are different
Most PM frameworks assume a single user. A marketplace works through pairs: buyers and sellers, drivers and riders, hosts and guests, lenders and borrowers. Each side has its own onboarding, retention, economics, and risk profile. A feature that helps buyers can hurt sellers. Pricing that grows volume can shrink the margin.
Interviewers love marketplace questions because the questions expose shallow analysis on the spot. A candidate who proposes a buyer-only feature reveals they have not modeled the seller. The same goes for pushing the take rate up without checking seller exit costs.
This dual-sided structure changes the analysis. The single-user question disappears in favor of a sharper question about side priority and supply tension. As Andrew Chen describes in The Cold Start Problem, every early marketplace lives or dies by getting the first side right before any growth tactics matter.
The MATCH framework
Use MATCH to break down any marketplace question in thirty seconds. The framework maps five forces: matching efficiency, asymmetry between sides, take rate economics, cross-side network effects, and platform health.
Matching efficiency
Liquidity is the first thing a buyer notices on a marketplace. You open Uber, see no cars, leave the app. Try Airbnb, find no listings in town, abandon the platform. Liquidity measures the share of demand that finds supply within a window.
Common metrics include fill rate, request-to-book ratio, time-to-match, and median wait time. Liquidity also depends on geography and time. Uber needs liquidity per neighborhood per minute. Airbnb needs liquidity per market per booking window. Aggregate metrics hide local failures.
In an interview, ask about liquidity before features. A marketplace with thin liquidity needs supply growth before a checkout redesign.
Asymmetry
Every marketplace has a harder acquisition target on one side. Supply or demand starts as the bottleneck. Uber subsidized drivers. Airbnb chased hosts in New York. eBay courted sellers with free listings. DoorDash paid restaurants for early sign-ups.
The constrained side gets the subsidy. Monetization rides on the other side. Misread which side is constrained, and the spend hits the wrong incentive.
Sangeet Paul Choudary recommends starting with the producer side, since that side often holds the network effects.
Take rate
Take rate is the platform's share of gross merchandise value. eBay sits near 13 percent. Etsy near 9 percent. Uber near 25 percent. Airbnb near 14 percent.
Two forces bound the rate. Push too high, suppliers leave or build a direct channel. Stay too low, the platform cannot fund operations or growth. The structure also varies by market side. Uber pulls most of its rake from drivers. Airbnb splits its fee between hosts and guests. Etsy combines listing fees with transaction fees. Upwork takes a sliding percentage from freelancer earnings.
Lenny Rachitsky writes that take rate emerges from two pressures: supplier alternatives and buyer price sensitivity. In interviews, ground take rate decisions in supplier alternatives. A unique seller with no other distribution channel tolerates a higher take. Commodity sellers with many options tolerate a much lower rate.
Cross-side effects
A marketplace gets stronger as both sides add users. More drivers attract more riders. The reverse holds too: more riders pull in more drivers. This is the cross-side network effect.
Same-side effects also shape the dynamics. More drivers on Uber means longer idle times for each driver. Same-side effects sometimes carry a negative sign.
In interviews, draw both arrows on the whiteboard. Show how supply growth helps demand and where it hurts other suppliers. The picture earns credit. A strong cross-side loop is the source of long-term defensibility for any platform business.
Health
Trust failure kills a marketplace. Fraudulent sellers on eBay, fake reviews on Yelp, dangerous hosts on Airbnb, fake gigs on Fiverr: every two-sided business has a trust failure mode. Health metrics include fraud rate, dispute rate, refund rate, and repeat purchase rate.
Healthy marketplaces invest in moderation, identity verification, dispute resolution, and seller vetting before a crisis. Unhealthy ones add it after a scandal.
How to use MATCH in an interview
When the prompt names a marketplace, walk through MATCH at the whiteboard. Identify the constrained side at the start. Estimate the take rate. Sketch the cross-side loop. Call out one health risk.
This sequence shows the interviewer you think in systems. Most candidates jump to features. You will jump to forces.
A worked example: Instacart
Try MATCH on Instacart. Matching efficiency lives at the store level per delivery window, so the right metric is order fill rate per zip per hour. Asymmetry favors the shopper side as the constrained input, since demand for grocery delivery scales faster than the trained shopper supply. Take rate stacks delivery fees, service fees, retailer markup, and ad placements, with retailer alternatives like store-direct pickup applying downward pressure on the rate. Cross-side effects run strong from shoppers to customers and weaker in reverse, with negative same-side effects when too many shoppers chase the same batch. Health risks live in order accuracy and item substitutions, with mismatch rate as the lead indicator.
That walk-through takes ninety seconds in an interview. The interviewer hears a structured analysis with named tradeoffs and a clear monetization view.
Four common traps
Four patterns sink candidates on marketplace questions. First, treating both sides as one user. Second, ignoring the cold start problem. Third, recommending a take rate change without a supplier alternatives analysis. Fourth, skipping health metrics across the board.
Use MATCH and you avoid all four.
Works cited
Chen, Andrew. The Cold Start Problem: How to Start and Scale Network Effects. Harper Business, 2021. https://andrewchen.com/
Choudary, Sangeet Paul. Platform Scale. Platform Thinking Labs, 2015. https://platformthinkinglabs.com/
Rachitsky, Lenny. "Lenny's Newsletter." Substack. https://www.lennysnewsletter.com/