The Two Faces of the Matching Model: How Rakuten and Recruit Connect Supply and Demand, and Why It Ends in Winner-Takes-All
Some business models neither make products nor hold inventory, yet earn steadily over the long run. The matching model is one of them. It is traditional and reliable, and especially well suited to IT companies. But once it gains momentum, it tends to push a market toward dominance by a handful of platforms. The key to understanding it is not the label "platform" but two questions: where value comes from, and where the money is collected.
1. Value comes from bringing supply and demand together
The matching model does something simple: it connects supply and demand, provides a place to transact, and earns money by brokering. USS, which runs used-car auction halls, is the classic form. It does not own every car; it lets sellers and buyers complete transactions in one venue. Rakuten Ichiba and Recruit's various booking and job sites move the same structure onto the internet.
Two axes clarify the design. The horizontal axis is the source of value: the company concentrates on "the power to bring parties together," or matching capability. Value does not come from what the company produces, but from helping supply and demand find each other. The vertical axis is how profit is captured: the company places itself in a "three-party market," becomes the platform, lets third parties match with one another, and charges for those matching opportunities.
Put the two together, and the outline of the model emerges: the model earns not by owning resources but by arranging for resources to meet.
2. Rakuten: merchants first, users second
Rakuten Ichiba is Japan's largest e-commerce platform and the most direct example of the matching model. Its core question is singular: can it keep attracting merchants that shoppers find appealing? The more merchants it has, and the more attractive they are, the more users come to buy. The more users arrive, the more merchants want in. Once this loop starts turning, the platform keeps growing.
Rakuten's approach is the "mall model." Think of a real estate company that runs a shopping mall: it rents space to tenants, collects fixed rent, and takes a percentage of sales. Rakuten builds Rakuten Ichiba on the internet; merchants handle product listings, order management, and shipping; Rakuten collects listing fees and sales-based commissions. In other words, what Rakuten really has to maintain is servers, merchants willing to open stores, and customers.
This model is often compared with Amazon. Amazon runs a distribution-retail model: it buys, holds inventory, and sells, like Amazon in the United States or Askul. Beyond the servers an e-commerce site needs, it must build large warehouses and carry heavy inventory, which demands substantial upfront investment and cash. The mall model is far lighter up front, but its weakness is equally clear: without merchants, customers do not come. To grow, the first task is not advertising but recruiting merchants one by one.
Rakuten Ichiba later grew into a platform with roughly 57,000 stores and about ¥6 trillion in gross merchandise value, and went on to promote the "Rakuten ecosystem": Rakuten Mobile, Rakuten Travel, Rakuten Card, Rakuten Pay, and other businesses side by side. Travel and payments are, in essence, extensions of the matching model, connecting travelers with accommodation and users with payment occasions.
3. Recruit: replicating one model across industries
If Rakuten went deep in one enormous market, Recruit Holdings went wide. Indeed covers jobs, Hot Pepper covers dining, Hot Pepper Beauty covers beauty, Jalan covers travel, Zexy covers weddings, and SUUMO covers housing. The industries differ; the underlying structure is the same.
Take Hot Pepper Beauty. Salons post information on the platform, users book through the system, and the platform takes a fee whenever a booking is made. The platform does not own the salons or run their operations; it simply helps "people who want to look better" meet "people who can provide the service" more efficiently. Recruit's approach shows that the model's applicability does not depend on whether an industry looks like the internet. It depends on whether the industry has a large, fragmented supply and demand, and whether matching can create real value.
4. Winner-takes-all: after the internet erased distance
The matching model has a strong tendency in IT: winner-takes-all. The reason is not technology but space.
A physical shopping mall, even one billed as Japan's largest, can serve only people within a certain distance in the Tokyo metropolitan area. An online mall can be reached from anywhere in the world. Delivery and connectivity costs remain, but compared with the physical world, reach has expanded enormously. The result is that users concentrate on a few platforms, and a few companies dominate the market.
Japan's e-commerce market is a clear example. Rakuten Ichiba, Amazon, and LINE Yahoo account for nearly all of the main market, and their combined gross merchandise value comes to roughly ¥14 trillion. That figure shows both the commercial efficiency of the matching model and how hard it is for a newcomer to break in.
5. What happens when platforms get too strong
Platforms grow quickly and push competitors out. That is not merely the outcome of competition; it also changes the balance of power in transactions. A sufficiently strong platform may pass unfavorable terms on to suppliers and consumers in order to increase its own profit. The influence of Meta Platforms mentioned in the talk, and recent government concerns about platforms such as Google's parent Alphabet and Japan's LINE Yahoo, which have grown powerful through advertising and matching models, point to the same structural problem: once a platform becomes the necessary route for supply and demand to meet, who sets the rules and whether the terms are fair stop being the platform's own business.
Two things need to be separated here. A platform creates value through matching; that is the source of the model's efficiency. A platform gains stronger bargaining power by controlling the matching gateway; that is a byproduct of the same mechanism. The former deserves credit; the latter deserves scrutiny. The talk does not go into specific regulatory designs, but the concern it raises is clear: the stronger the platform, the more its terms and rule-making should be discussed in the open.
Conclusion
Back to the two axes: value comes from matching, and profit comes from third parties completing transactions on the platform. These two sentences are both the manual for the matching model and its risk warning. Rakuten showed how large the model can grow through the merchant-user loop; Recruit showed how many industries it can be replicated into; Japan's e-commerce landscape showed how easily it concentrates. For anyone building a platform, the real design work is not the website but which side of supply and demand to start from and what terms keep both sides willing to stay. For anyone using a platform, understanding this structure helps judge whether the market in front of them is efficient or one where bargaining power has become clearly unbalanced.

