Network Effects · Two-Sided Markets
Network Effect & Two-Sided Market
What is a network effect?
- The cycle where value grows with every new user
- Two-sided markets must first solve the chicken-and-egg problem
- Breaking through the tipping point with a one-sided subsidy
Truly understanding network effects means changing how you split your budget. For an ordinary product, you match what it costs to win one customer against what that customer pays. In a two-sided market, you recover the acquisition cost on one side from revenue on the other. If the KRW 100 million (≈ US$74,800) you spent signing up 500 suppliers for free comes back as fees from the buyers who joined to reach them, it is the right investment even when the supplier side runs at a loss. So this concept is a tool for choosing which side to run at a loss. Split your budget evenly between the two sides, and neither will cross the tipping point.
In practice, you look at three things separately. The first is scope. In markets like delivery or secondhand trading, where value arises only within a local area, 300 suppliers in one neighborhood matter more than 500,000 sign-ups nationwide. The second is multi-homing cost. If it costs a driver or seller nothing to keep a competing app open at the same time, a rising market share creates no lock-in. The third is congestion. When supply grows faster than demand, orders per driver fall and a negative effect kicks in within the same side, so you have to manage growth on both sides as a ratio. Economies of scale means unit costs fall as you grow. A network effect means users create value for one another, so the investment logic is different.
The most common mistake in the field is reading the number of sign-ups as a network effect. If only 4,000 of 100,000 members have traded even once a month, a new sign-up adds virtually no value for existing users. Even riskier is mistaking a subsidy-driven number for the tipping point. If more than 30% of suppliers left in the month the fee waiver ended, they were never attached to the network. They were a group buying the discount. Judge the tipping point by density and repeat-transaction rate at the regional and category level, not by nationwide cumulative totals. When it collapses, it does so much faster than it was built.
John KimInstructor John Kim, author of "B2B Marketing Design" and CEO of Nanum Consulting
Sungmin JooExpert in assessment-based coaching leadership, entrepreneurship education, and design thinkingThe idea was popularized by Metcalfe’s Law, named for Ethernet inventor Robert Metcalfe: a network’s value is proportional to the square of its users. Two-sided market theory grew out of platform economics research by Jean Tirole and others.
Consider a hypothetical case. Company A brokers interior construction jobs in a regional mid-sized city, with 47 employees and KRW 900 million (≈ US$673,400) in monthly transaction volume. It had signed up 3,200 contractors, but only 12% of quote requests got a reply within 24 hours, and the customer repeat-request rate had dropped to 8%. If it couldn’t turn this around within 6 months, the next funding round would be in doubt.
- When the platform planning lead dug into three months of transaction data, only 410 of the 3,200 drivers had submitted even one quote in the past 30 days. “Our suppliers are 410, not 3,200” was the first sentence of the meeting.
- The CEO stopped the nationwide expansion and focused the budget on the two cities that generate 62% of transactions. The goal was 15 active drivers per neighborhood, and ads in every other region were cut entirely.
- The supply manager visited the top 80 drivers in person and waived their fees for 6 months in exchange for a promise to reply to quote requests within 24 hours. The deal was: “We’ll bring in the jobs; you just answer them.”
- On the demand side, they cut ad spend by 40% and instead showed “average response time: 3 hours” on the detail pages. The quote-request conversion rate rose from 4.1% to 7.3%.
- Six months later, active drivers in the two cities had grown from 410 to 1,150, the 24-hour response rate from 12% to 61%, and monthly transaction volume from KRW 900 million (≈ US$673,400) to KRW 1.4 billion (≈ US$1,047,400). After the fee waiver ended, churn was only 9%.