Hiroshi Mikitani Built Rakuten on Merchant Trust, Then Bought a Baseball Team

The Banker Who Left in 1997

Hiroshi Mikitani was an Industrial Bank of Japan analyst in 1995 when the Kobe earthquake flattened his hometown and, by his account, exposed how slow institutions were to respond. He left the bank in 1997 and founded Rakuten with a handful of employees. The first product was not a marketplace but a shopping mall for small merchants: Rakuten Ichiba, launched in 1997, let independent shops list goods and pay commission rather than buy inventory.

That model put Rakuten in direct competition with larger retailers without taking on their warehousing costs. By the early 2000s, Rakuten Ichiba had tens of thousands of merchants and was processing billions of yen in transactions. Mikitani’s public framing was consistently merchant empowerment, not price disruption.

The mechanism is worth spelling out, because it explains why the model scaled the way it did. A conventional retailer buys stock, stores it, and sells it at a markup; its balance sheet carries the inventory risk and its margins depend on how fast goods move. Rakuten inverted that. Merchants kept their own inventory, set their own prices, and handled their own shipping, while Rakuten collected monthly fees and a commission on sales. The company’s capital went into software, servers, and merchant recruitment rather than warehouses and trucks. That meant Rakuten could add a new merchant without building a new distribution center, and it could grow transaction volume faster than its fixed costs. The trade-off was control: Rakuten could not guarantee delivery times or product quality the way a vertically integrated retailer could, so it substituted rules, rankings, and training for ownership.

The Kobe episode is the origin story Mikitani himself returns to, and it fits the pattern of a man who concluded that established institutions move too slowly. Whether the earthquake alone explains the decision or merely crystallized a restlessness already present is a matter of interpretation; what is documented is the sequence — analyst at a major bank, departure in 1997, founding of Rakuten the same year. The timing also placed him at the front edge of Japan’s internet commerce wave, when consumer broadband was spreading and the country’s retail landscape was still organized around department stores and chains.

It is worth pausing on what a bank analyst in mid-1990s Japan would have seen, because it frames the bet. The country’s financial system was still working through the aftermath of the asset bubble, and its big institutions were oriented toward protecting existing relationships rather than financing new ones. A young analyst inside that system could observe, at close range, how decisions moved through committees and how long it took for capital to reach a new idea. The Kobe earthquake made the same point in physical terms: institutions that should have moved quickly did not. Mikitani’s conclusion — that relationships could be organized more directly, without the intermediary layer — is the thread that runs through everything Rakuten later did. The founding year matters for a second reason as well: 1997 was early enough that the infrastructure was still thin, which meant a company had to build capability rather than rent it, and late enough that the consumer internet was clearly arriving.

A Marketplace Built on Merchant Loyalty

Rakuten’s early advantage was its “shop of the year” ranking system and its insistence that merchants manage their own storefronts. The company charged monthly fees plus commissions, and it pushed merchants to adopt email newsletters and customer-service standards. This was unusual in Japan, where retail was dominated by department stores and convenience-store chains with centralized control.

The ranking system deserves attention because it was a governance tool disguised as a contest. By publicly recognizing high-performing shops, Rakuten set an aspirational bar without owning the stores, and merchants competed for standing that translated into visibility on the platform. The newsletter requirement had a similar logic: it pushed merchants to build direct relationships with their own customers, which made those merchants more dependent on the platform as the channel for those relationships. Rakuten was, in effect, teaching small shops to do marketing they had never done, and in the process making itself indispensable to them.

Rakuten listed on the JASDAQ exchange in 2000 and later moved to the Tokyo Stock Exchange. Its market value rose through the 2000s as it expanded into travel, banking, and credit cards. The company’s annual securities report for the fiscal year ended December 2005 shows revenue of roughly ¥100 billion, a figure that grew steadily as Rakuten added financial services.

The listing mattered beyond the capital it raised. It gave Rakuten a public currency — shares — that it could later use to buy companies abroad, and it imposed disclosure discipline on a young firm. The move to the Tokyo Stock Exchange’s main market signaled that the company had graduated from the startup tier. The financial-services expansion was not incidental: travel bookings, banking, and credit cards all sat on top of the same customer base the marketplace had assembled, and each new service gave existing users another reason to stay inside Rakuten’s ecosystem. Revenue of roughly ¥100 billion for fiscal 2005 is a marker of how far the commission model had carried the company in under a decade.

There is a subtler point in the merchant-training model that is easy to miss. A platform that only collects fees is a landlord; a platform that teaches its tenants how to sell is a partner, and partners are harder to leave. The newsletters, the service standards, the rankings — each one transferred a skill to the merchant while deepening the merchant’s reliance on Rakuten as the place where that skill paid off. That is the engine behind the loyalty the company came to depend on, and it explains why the merchant base proved durable even as larger, better-capitalized competitors entered Japanese e-commerce. Rakuten could not outspend a logistics giant on warehouses, but it could out-invest in the merchants themselves, and that investment compounded in a way that physical assets do not.

Buying a Baseball Team as a Brand Bet

In 2004, Rakuten bought the failing Kintetsu Buffaloes baseball franchise and moved it to Sendai, creating the Tohoku Rakuten Golden Eagles. The purchase was widely read as a marketing move. Mikitani said at the time that the team would strengthen Rakuten’s brand in northern Japan and give the company a physical presence beyond the screen.

The Eagles won the Japan Series in 2013, nine years after the purchase. The team’s success coincided with Rakuten’s expansion into credit cards and banking, which became a larger share of profit than the marketplace itself. By 2015, Rakuten’s financial-services segment was contributing a substantial portion of operating income, according to the company’s annual reports.

The baseball purchase is the clearest example of the pattern the record later reveals: Rakuten buying a relationship rather than an asset base. A regional team generates loyalty that no ad campaign can manufacture, and it embeds the company in a local community — Sendai and the wider Tohoku region — where Rakuten had no physical footprint. Skeptics at the time questioned whether a web company could run a baseball club, and the early years were not triumphant on the field. The 2013 championship changed the narrative, but the more important return may have been the brand association itself, which outlasted any single season. The financial-services growth in the same period is the quieter story: the marketplace supplied the customers, and banking and cards supplied the margins.

What makes the baseball move instructive is the asymmetry it accepted. A team is a civic institution before it is a business, and civic institutions carry obligations — to fans, to a city, to a league — that a marketplace does not. Rakuten took on those obligations in exchange for a kind of standing that advertising cannot buy. The nine-year gap between purchase and championship is itself the point: brand bets of this kind pay off slowly and unevenly, and the company had to absorb several seasons of doubt before the payoff arrived. The same period, meanwhile, shows the financial-services arm becoming the profit center, which is a reminder that the visible, emotional bet and the quiet, arithmetic one were running on parallel tracks.

The Global Push and Its Limits

Rakuten expanded abroad in the 2010s, buying PriceMinister in France, Ikeda in Brazil, and OverDrive, a US digital-library service, in 2015. It also invested in Lyft and other startups. The global strategy was not uniformly successful: Rakuten wrote down parts of its overseas operations and eventually sold some assets.

The overseas campaign followed the domestic playbook — acquire a platform with an existing user or merchant base, then layer Rakuten’s services on top. PriceMinister gave Rakuten a French marketplace; Ikeda gave it a Brazilian one; OverDrive gave it a library lending channel in the United States. The Lyft investment was a different kind of bet, a stake in a ride-hailing platform rather than an acquisition of one. The write-downs and divestitures that followed are the counter-evidence to the idea that the model traveled easily. Marketplace dynamics are local: merchant habits, payment preferences, and consumer trust differ by country, and a commission model that worked in Japan did not automatically reproduce itself abroad. Selling assets later was an admission that some of those relationships were not worth owning at the price paid.

In 2019, Rakuten launched its own mobile network in Japan, a capital-intensive bet that weighed on profits. The company’s founder remained chairman and CEO, and Rakuten’s market value in 2024 was far below its 2015 peak, reflecting investor skepticism about the mobile buildout.

The mobile network was the most expensive version of the same instinct. Building a carrier requires spectrum, towers, and years of spending before subscribers arrive, and the drag on profits was immediate while the payoff was speculative. Investors marked the shares down accordingly. For readers, the lesson is not that the bet was wrong but that it was structurally different from the earlier ones: the marketplace and the baseball team could be grown incrementally, while a network demands capital up front regardless of demand. The gap between the 2015 peak and the 2024 valuation is the market’s price for that difference.

It is worth separating the two kinds of failure in this period, because they teach different things. The overseas write-downs were a failure of translation: the playbook was sound in Japan but the local conditions — merchant culture, payment habits, trust in an unfamiliar brand — did not match the assumptions. The mobile buildout was a failure of category: it asked a company built on other people’s inventory and other people’s storefronts to carry its own heavy fixed assets. One was a bet that the relationship model would travel; the other was a bet that the relationship model could absorb a business it was never designed for. The market’s reaction to the second was harsher, and reasonably so, because it was the larger departure from the founding logic.

What the Record Shows

Mikitani’s career is a case study in building a marketplace around merchant relationships rather than logistics. Rakuten Ichiba never matched Amazon’s fulfillment scale, but it created a durable base of small Japanese merchants who depended on the platform. The baseball team and the financial-services arm were brand and profit extensions of that base.

The 1997 decision to leave banking for a commission-based mall was the founding bet. The 2004 baseball purchase was the brand bet. The 2019 mobile network was the capital bet. Each one followed the same logic: own the relationship, not the inventory.

That logic has a cost as well as a benefit, and the record shows both. Owning relationships rather than inventory let Rakuten grow without the capital burden of a retailer, and it gave small merchants a channel they could not have built alone. But relationships are harder to consolidate across borders than warehouses are, which is why the global push produced write-downs, and relationships do not shield a company from the heavy fixed costs of a network buildout. The durable base of Japanese merchants remains the through-line: it funded the financial services, it gave the brand a customer list, and it is the reason the company’s history reads as a series of extensions from a single founding idea rather than a series of unrelated ventures. For anyone watching a platform company today, the Rakuten record suggests a useful question: is the company accumulating relationships it can monetize in many ways, or is it accumulating assets that must pay for themselves? Mikitani’s answer, across three decades, was consistently the former — and the market’s verdict on the most capital-heavy version of that answer was decidedly mixed.

The record also invites a second question, about durability. Relationships are an asset that appreciates with use and depreciates with neglect; a merchant who feels the platform has stopped investing in them is a merchant who starts looking elsewhere. Rakuten’s early training programs and rankings were investments in that asset, and the later financial services were withdrawals from it — legitimate withdrawals, since each new service gave the merchant’s customers another reason to stay, but withdrawals nonetheless. The distinction matters for anyone reading a platform company’s accounts: the balance sheet will show the servers and the spectrum licenses, but it will not show the accumulated trust of tens of thousands of small shops, which is the thing the company actually built and the thing that funded everything else.