A senior-care info site earning ¥17,200/month in profit sold for ¥1.4 million — inside a valuation of 81 months of profit

On February 17, 2023, Rakko Inc., which operates a website marketplace, published an analysis of 1,050 completed deals from its own “Rakko M&A” service. Buried in it is one deal whose numbers, at first glance, don’t add up: a senior-care and nursing-home information site with ¥17,200/month in profit sold for ¥1.4 million.

The deal data, against what the same survey says is typical

The same survey shows the broader market operates at a very different order of magnitude:

The multiple distribution is also broken out in detail: 12–24 months accounts for about 40%, 6–under-12 months for 28.8%, 24+ months for 21.9%, and under 6 months for 8.5%. “Within 2 years of profit” covers about 80%, and within 3 years covers 93%. The nursing-home site’s 81 months, 6 years and 9 months, sits well outside the remaining 7%.

Multiply ¥17,200/month by the average 17.8x multiple and you get ¥306,160. Going by the market average, this site should have closed around ¥300,000 — it sold for 4.6x that. What was the extra roughly ¥1.09 million actually paying for? That’s the substance of this case.

The decisive factor: the valuation axis shifted from “profit” to “inventory”

There’s no dramatic turning point in this deal. The operator didn’t do anything to grow profit, monthly profit stayed at roughly ¥17,200 throughout. What drove the price up was not operational effort, but the buyer’s evaluation axis moving away from profit entirely.

The clue lies in the ratio of monthly PV to page count. About 30,000 PV against 30,000+ pages, roughly 1 PV per page per month, meaning most pages are barely being read at all. That ratio is unusual for an article-format blog. It points to a directory-style structure that catalogs facilities one by one. If value is attached to pages that aren’t being read, the buyer isn’t buying “current traffic.”

Dividing the sale price by page count gives about ¥47 per page. Compare that to the same survey’s figure of ¥2,370.8 per article for sites in the under-¥10,000-profit tier, over 50 times lower (the segments differ, so this isn’t a strict apples-to-apples comparison, but the order-of-magnitude gap is clear). What the buyer is paying for is not the quality of any one page but the comprehensiveness that only exists once 30,000 pages are bundled together.

Why does comprehensiveness carry value? Search demand for senior-care facilities is made up of a mass of narrow, location-specific queries, “[region name] + nursing home.” Search volume per query is small, but cover every region in the country and the aggregate search surface becomes enormous. And for a competitor to build the same coverage would require both the production cost of 30,000 pages and the time it takes for Google to trust them. To the buyer, ¥1.4 million was judged against the cost of building 30,000 pages from scratch, and inheriting a fully-formed directory database at ¥47/page was, by that math, cheap.

What pushed back against the YMYL discount

One thing to flag: the same survey explicitly lists “YMYL categories (medical, financial, legal)” as a negative-evaluation factor. A site touching senior housing and care sits on the periphery of YMYL. Even so, it commanded an 81x multiple.

Two factors from the survey’s positive-evaluation list can plausibly explain what pushed back against that: being “niche and highly specialized,” and low personal dependency. Rakko lists “high personal dependency (can’t be outsourced)” as a negative factor, but a directory-style facility-info site doesn’t rely on an individual author’s personal experience or voice. Swap out the operator, and the published data and structure keep functioning unchanged. What an M&A market is ultimately buying is “will this keep running the same way after the handoff”, and on that single dimension, this site outweighed its YMYL discount.

Other deals in the same survey follow the same logic:

The food/diet site with the most PV (110,000) has a multiple of 30, lower than the trivia site’s 33.3, despite the trivia site having only 8,600 PV. PV volume barely determines the multiple. What deals that beat market average share in common is a set of time-only-buys-this factors: years of operating history, a YouTube channel, 30,000 pages, and a year-round revenue base. The trivia site’s evaluation explicitly notes “traffic declines only gradually even after years without an update.” What the sales market is pricing here is neglect-resistance.

Where this structure breaks down

The same survey also lists negative-evaluation factors: black-hat SEO, YMYL, high personal dependency, curation/AI-generated content, and trend/news-driven content. The last two sit uncomfortably close to this deal. A 30,000-page site could easily be judged an “automated-generation flood” depending on how it was built. This deal likely earned its price because the volume represented genuine coverage of real-world data, but the source material doesn’t disclose the production method. The distinction is important to hold onto: volume itself isn’t value. Volume is only value when it maps one-to-one with search demand.

The buyer’s risk is also plain. Paying ¥1.4 million against ¥17,200/month in profit means a simple payback period of 6 years 9 months. If demand for senior-care searches doesn’t hold, or if search-engine evaluation shifts, this valuation stops working. As Rakko notes, “rising demand from an aging society” was cited as an evaluation factor, meaning this ¥1.4 million price includes a bet on the category’s future demand.

A great deal also isn’t disclosed: this deal’s operating history, revenue breakdown, days to close, and buyer profile are all unpublished. What’s known is limited to price, profit, PV, page count, and Rakko’s evaluation comments.

How far can this be copied?

The reproducible part is the mindset. Even at ¥10,000-something a month in profit, a structure that comprehensively covers a long-tail of demand, combined with neglect-resistance, can command value as an asset, a different exit than chasing monthly revenue. Removing personal dependency, building something that keeps running after the operator changes, expands your options even if you’re not currently planning to sell.

What isn’t reproducible is equally clear. 30,000 pages can’t be built quickly, time itself is the thing being evaluated. The aging-society tailwind is a byproduct of category selection, not something built through effort. And 81x is an outlier within the 1,050 deals. The figure to anchor expectations on is the 12–24-month range at the center of the distribution.

One more thing: the same survey also reports that 90% of listings receive their first negotiation inquiry within 24 hours of listing. Whether a valuation is reasonable is something the market judges within a day.

Related reading

• ScrapingBee, and how its co-founders sold their stake — an example where a profit-based multiple applied cleanly to a sale.
• MENTA’s acquisition by Lancers — the evaluation axis when a platform acquires a solo-built service.