MENTA, Shingo Irie's 30th Indie Project: From ¥1.4M Monthly Revenue to a Share Transfer to Lancers — the Full Story

There are several indie-developer exit stories in Japan, but MENTA is the only one where primary sources cover everything: day-one revenue, failed initiatives, the process of building a team, and the organizational changes after the sale. Drawing on three public sources from different vantage points (Shingo Irie’s own note posts, a Lifehacker interview, and the engineering blog of the acquirer Lancers) this article breaks down what was inside “the 30th product that finally hit.”

Timeline of the Business

What the Business Was

MENTA is an online mentoring service where “anyone can become someone’s teacher, or someone’s student.” It is a matching platform where users pay a monthly fee to consult directly with working engineers and other professionals; mentor registrations grew to around 4,000. The core of the design is that instead of one-off spot consultations, the model of Coconala or TimeTicket, MENTA sold an ongoing relationship with unlimited questions. When Irie himself learned programming, he was saved by a mentor who taught him at close range. MENTA is that formative experience turned into a product. He has said that “the more a service is tied to your own formative experience, the more vividly you can imagine the user, so you don’t miss the mark.”

The platform earns a commission on gross transactions. Working backwards from the published figures as of January 2020 (over ¥9M gross, ¥1.4M revenue), the take rate comes out to roughly 15%. The rate itself was never disclosed, so that figure is our editorial back-calculation.

The 12 Years Before Number 30

Irie was born in 1982 in the Goto Islands of Nagasaki. He built his first service in 2008 and went independent as a freelancer in 2011. In his early independent years he released 10 apps, and every one of them flopped. However, “people who saw the apps sent inquiries for contract development work”. The failures functioned directly as sales assets.

In April 2018, having calculated that his savings from contract work would “last even with zero income for a year in the worst case,” he stopped taking contract work and went all-in on his own services. MENTA was his 30th product overall. The launch preparation was meticulous: when he recruited 50 test users on Twitter, about 200 applied, and he had secured 200 mentors before opening. Stocking the supply side first, the lifeline of any matching service, is what converted 8,000 day-one visitors into ¥90,000 of actual transactions.

From Generalist to Specialist — What PMF Actually Felt Like

At launch, MENTA was a mentoring service with no restriction on subject. But after ¥700,000 in gross transactions in month one, month two also came in at ¥700,000. An early stall. The turning point was the January 2019 specialization in programming. This was not an armchair strategy pivot, though. In reality, the vast majority of registered mentors were already in programming. On Twitter, users themselves had started describing a learning pattern of “Progate for basics → MENTA for practice”. And a reputation had formed that MENTA cost “one-thirtieth of a programming school.” What Irie did was ratify those facts and rewrite the top page for programming beginners. After that, he recalls, “gross transactions grew as a natural consequence”. That was PMF. Our reading matches his summary, but the skill we would credit is not the idea of specializing so much as the order of operations: he let the market give the answer first, and only then abandoned a general-purpose service already moving ¥700,000 a month.

The other growth engine was the “Posts” feature launched in May 2019. Mentors publish articles to make their skills visible, and learners can post “mentor wanted” requests. It is a CGM structure where user-generated content brings in search traffic, and customer acquisition without ad spend began to compound.

The Initiatives That Whiffed

Behind the up-and-to-the-right curve, Irie has also disclosed the initiatives that failed, in concrete terms. A campaign offering “¥100,000 prize money to the mentors who taught the most students” fizzled because the entry conditions were vague and drove no action. An online drinking party, held out of anxiety during a stagnant period, ended in awkward silence. A scholarship campaign using Polca attracted exactly one backer. Two lessons came out of this. People act only on dead-simple initiatives where the required action is unambiguous, like “tweet this and get a coupon”. The harder one, visible only in hindsight: digging deeper into existing users should have come before scrambling for new ones.

The Limits of One Person, Team Building, and the Sale

In late 2019, feeling the ceiling on his development speed, Irie assembled a team of contractors. His recruiting method was distinctive: nearly all members were MENTA mentors themselves. He delegated development, design, customer support, and server management, and moved himself into strategy and planning. At that point, he had a backlog of 30 initiatives waiting to be implemented.

In October 2020, MENTA joined the Lancers group via share transfer (price undisclosed). The stated deciding factor: “the flow of people learning on MENTA and then working on Lancers felt natural.” It was a decision to solve a resource shortage that had outgrown indie development by partnering with a company that had traffic-referral power in the skill-share market. After joining the group, the organization shifted from contractor-centric to 10 full-time employees, the infrastructure migrated from a single Sakura Cloud server to AWS (ECS/Fargate), and releases were automated with CircleCI. Registered users reached 30,000. Irie himself stepped back from hands-on development into a planning and decision-making role, and in December of the same year he also sold his cloud estimate/invoicing service CLOUD PAPER. That makes 2 of his 30 products reaching M&A.

Reading Between the Numbers

29 duds before the 30th hit. A 3% hit rate is normal. Irie built 30 services over 12 years, and MENTA was the one that grew into a real business. That batting average almost exactly matches Pieter Levels, who openly says “4 hits out of 70 products” (about 5%). Irie himself preaches volume: “The number of attempts is the key. Picasso left 13,500 works. Hokusai left 34,000.” Indie success is a game of attempts, not marksmanship, someone who retreats after one failure is structurally incapable of producing a hit. One line runs through this case: 29 duds delivered the 8,000 day-one visitors for number 30. Twelve years of shipping and narrating the misses in public built the audience that gave the winner its launch velocity.

PMF came from the decision to discard, not the ability to build. MENTA’s numbers moved not when features were added, but the moment he abandoned scope and narrowed the target. Narrowing while sitting on ¥700,000/month of gross transactions is a frightening call, but after narrowing, gross transactions grew more than 10x. And the grounds for specialization were not a hunch but observed data: the composition of the mentor pool and the vocabulary users had naturally begun using. The cause of the plateau was not missing features but ambiguity about “who is this service for”, structurally identical to Bannerbear, which took off after realizing “what should be niched down is the target market, not the problem”.

The buyer was not “the company that pays the most” but “the company that has distribution.” Lancers holds a customer base and brand in skill sharing. In a market with roughly 400,000 programming learners a year and a claimed 90% dropout rate, MENTA’s growth bottleneck was awareness, so a share transfer at the moment when it needed referrals more than money was rational. Placed alongside Zenn’s decision to choose Classmethod, a buyer that would not rush monetization, the axis for choosing an acquirer becomes clear: pick the partner who fills your service’s single greatest deficiency.

The dominant exit pattern in Japan is “sell, but keep building.” After joining the group Irie continued to lead MENTA’s planning and decisions, and has since kept up indie development and a YouTube channel (over 6,000 subscribers). Selling is not retirement but a purchase of money and time for the next round of attempts, and the “earn → sell → build again” loop is the real face of indie-developer FIRE.

How Reproducible Is This?

• What transfers: three patterns apply to matching services in general — “stock the supply side (200 mentors) before opening,” “when growth stalls, specialize to match what users are actually doing,” and “run only initiatives where the required action is unambiguous”
• What doesn’t: the 8,000 day-one visitors existed only because of the audience (followers) Irie had built through years of public output; an unknown builder cannot reproduce that launch. The precondition of savings that could survive “a year of zero income in the worst case” should not be overlooked either. And because the sale price is undisclosed, the economic return on two and a half years of invested time is ultimately unknowable

Related Cases

• Zenn: business transfer 4.5 months after launch — a domestic indie exit from the same era
• Bannerbear: $10K MRR after 7 failures — the overseas twin of “narrow the target and grow”
• PhotoAI: 4 hits out of 70 — indie development as a game of attempts