Repsona: Five Years Solo, ¥12.4M Total Revenue — The Open Ledger of a Task-Management SaaS Its Maker Calls a "Startup Failure"

“As a hobby side project, a great success. As a startup venture, a big failure.” That is how the developer of task-management SaaS “Repsona” summed up five years of operation. The numbers he published on note in January 2024: total revenue of ¥12,403,567 over five years. Averaged out, that is a little over ¥200,000 a month — a razor-thin level for one person to live on.

Repsona is a task- and project-management tool built around simplicity and speed, with development, operations, and marketing all handled by one person. It has 5,873 registered spaces (team workspaces) and 14,641 cumulative users. New spaces are still being created at a pace of roughly one per day.

Most indie revenue disclosures cherry-pick the good months. What makes this one valuable is that it opens five full years of books, failures included, and reports the result of each initiative in real numbers. Let’s go through them.

The five-year timeline

• May 2018: quits his job and starts developing full time
• June 2019: beta launch after eight months of development
• October 2019: official release with paid plans; first month’s revenue is ¥9,900
• January 2020: mobile app released (roughly 100 cumulative installs after five years)
• October 2022: price revision and restrictions on the free plan; revenue rose
• January 2024: publishes five-year total revenue of ¥12,403,567; 5,873 spaces, 14,641 users

Note that these are the numbers of someone who quit his job and went all in, not a side project. Eight months of development, five years of serious operation, and cumulative (not annual) revenue of ¥12.4 million.

Initiatives and results — why ads at ¥500 CPA still didn’t work

The note is candid because it reports the outcome of each attempt numerically.

Google Ads: ¥30,000 in spend brought 60 space registrations, a CPA of ¥500 per signup. Cheap, on its face. But free-to-paid conversion was about 1%, meaning it took roughly ¥50,000 of ad spend to create one paying customer. Recouping that through a low-priced monthly subscription takes years, spending on ads didn’t grow cash.

He also invested in credibility. Obtaining ISO/IEC 27001 (the information-security certification) cost ¥380,000 in audit fees, plus ¥360,000 a year in consulting and ¥150,000 a year in surveillance fees to maintain. It is reassurance for corporate adoption, but against a business doing ¥12.4M over five years, it’s a heavy fixed cost. He also ran PR TIMES press releases and content marketing through blogs and case studies.

The mobile app fared even worse. Released in January 2020, it sat at roughly 100 cumulative installs five years later. Against 5,873 web-side registered spaces, this was the initiative with the worst return on development effort. Mobile support feels “obviously necessary” for a task manager, but on indie resources it can be an area where numbers never justify the expectation.

The October 2022 price revision and free-plan restrictions pushed revenue up, but web traffic and conversion rates fell afterward. Some free users were forced to change plans. The trade-off between revenue and reach showed up plainly. By 2024, web traffic had fallen to around 30 visitors a day.

His own diagnosis: “they simply don’t know it exists”

Is it the features? The pricing? His conclusion sits further upstream. After listing everything he tried, he writes that the reason people don’t discover the product comes down to one thing: “they don’t know about it.”

That verdict carries weight coming from someone who spent five years building. He places the failure not in craftsmanship or price, but in never climbing the awareness ladder.

Task management is a category with countless free options and global giants. There, being known at all is the biggest gate, and product quality only matters after that. You can buy signups at ¥500 each, but only 1% convert after comparison shopping, a realistic exchange rate for a low-awareness product in a crowded category. Compare other one-person SaaS: Inkdrop took a passionate niche with a developer-focused note app, and Healthchecks runs a simple monitoring tool for the long haul, the density of competition differs at the category-selection stage.

Meanwhile, “big failure” is a verdict against the startup (hyper-growth) yardstick. 5,873 teams registered, new spaces are still created daily, and the service has run for five years without stopping. His other verdict, a great success as a hobby project, is no exaggeration either, judged by the numbers.

Conditions for repeatability, and the limits

The transferable asset is the record itself, initiatives verified with real numbers: ad CPA and free-to-paid conversion, the actual cost of certification, the side effects of a price change, the numbers an aspiring indie SaaS builder wants in advance, measured rather than estimated.

The limits are just as clear. One is that these are the numbers for a hyper-competitive category (task management). Pick a niche and the conversion math can change. Another is that credibility investments like ISO certification pay off when corporate contracts are stacking up, carried as fixed cost before awareness exists, only the weight remains. Read it as a sequencing problem. A third: quitting to go full time presumes savings to cover living costs until revenue arrives. Repsona’s first month of revenue was ¥9,900, four months after the beta launch. As also seen in Walica, a bill-splitting app run solo to the 300,000-user scale, in indie development the design of “what pays your rent” shapes outcomes as much as the design of the product.

One more thing. The ¥12.4M five-year total is not proof of failure but a measured data point: this is how far these conditions reached. Change any one condition (full-time solo, small-scale ads, crowded category) and the same effort lands somewhere else. Repsona is still operating, and this ledger is not yet closed. That a first-person record willing to say “big failure” keeps being published is itself a public good for the indie developers who come next.