I looked into what it actually takes to monetize this site.
There are tons of Japanese-language guides on the topic, but honestly, surprisingly few of them lead back to a primary source. Of the conditions described as “required,” fewer than half had a written basis in any official document.
Below I’ve split things into what has a basis, what doesn’t, and what nobody writes about but actually matters. This research reflects the situation as of late August 2026.
“AdSense requires a custom domain” → I couldn’t find it in writing
You often read that you won’t get AdSense approval on a free subdomain (like *.netlify.app).
Within what I searched, there is no official Google document that explicitly prohibits it. There are also multiple cases of sites getting approved on a subdomain alone.
That said, one of the conditions Google does require is “verifying that you own the site,” and the owner of the *.netlify.app domain is Netlify. It makes logical sense that you’d be at a disadvantage there. Japanese-language guides also broadly agree that “a custom domain is the safer bet.”
My conclusion: I’m getting one. Not “there’s no basis, so it’s unnecessary,” but rather: if ¥1,000–2,000 a year removes one source of uncertainty, that’s cheap. After getting it there are only two things to do: swap the site URL in the config file and the URL in robots.txt, then rebuild. Netlify automatically 301-redirects the old URL, so the link equity built up until then isn’t lost.
“10–15 articles” → Not an official criterion, but in practice you need them
This wasn’t an official number either. What AdSense asks for is a qualitative condition: “unique, valuable content.”
Still, a human does the review. It’s easy to imagine that a site with only 3 articles is unlikely to be judged “unique and valuable.” In practice I set 10–15 as my target. As I understand it, the point isn’t the number itself but getting the site to a state where it’s obvious “this site is ongoing.”
What nobody emphasizes but actually gets sites rejected: broken internal links
This was the biggest takeaway. For a site with the same setup as mine (a static site + Netlify), the cause of an actual rejection was internal links returning 404.
Everyone talks about content quality, but from the reviewer’s side, a site with broken links is simply “not maintained.” And this is exactly the kind of defect you can check automatically.
I built a check into the build.
// Extract href/src from every HTML file under dist and check that each resolves within dist
for (const m of html.matchAll(/(?:href|src)="([^"]+)"/g)) { ... }"build": "node scripts/gen-ogp.mjs && astro build && node scripts/check-links.mjs"If there’s even one broken link, the build fails. Checking by hand will get forgotten at some point, so it’s pointless unless you go as far as failing the build. In fact, every time I write an article, 1–2 links get caught (mostly slug typos).
Analytics: I chose cookieless over GA4
Without numbers, you can’t iterate on your articles. But GA4 uses cookies, which means adding to the privacy policy and thinking through consent.
I went with Cloudflare Web Analytics.
- Completely free
- No cookies. No fingerprinting either
- You don’t even need to use Cloudflare’s DNS. You just paste one beacon line into
<head> - No consent banner needed
At the scale of a personal blog, there’s hardly ever a situation where you’d use GA4’s full feature set. I judged that not having to show a consent banner is worth more than the extra analysis dimensions I’d get.
Three affiliate networks (ASPs), each with a different role
| ASP | Review | What it’s for |
|---|---|---|
| A8.net | None | High-payout programs like web hosting (rental servers). Start here |
| Moshimo Affiliate | Yes (passes with a few articles) | Amazon/Rakuten partnerships in place from the start + a +12% bonus on sales (its “W reward”) |
| ValueCommerce | Yes | Carries Udemy’s program |
The most important row is the second one. Applying directly to Amazon Associates comes with the condition of “3 qualifying sales within 180 days.” That’s tough for a site that just launched.
Going through Moshimo Affiliate, that condition doesn’t apply. And the commission rate is higher there. If you want to place Amazon links, going through Moshimo rather than applying directly turned out to be the right answer.
Registration requires your real name, address, and bank account, but none of that is published on the site. I confirmed that it’s fine for the site to keep showing just my business name.
AdSense and affiliate revenue are on different scales
These are the figures I could find as realistic market rates.
- AdSense-centric: ¥0.2–0.3 per pageview. ¥20,000–30,000 a month at 100,000 pageviews a month
- Affiliate-centric (via ASPs): about ¥1 per pageview
For the same pageviews, affiliate is more efficient by a whole order of magnitude. But this doesn’t mean “line up high-payout programs and you win.” These are the numbers for introducing things you actually use, in the context of an article.
What I decided not to do
- Ad networks with no review (Ninja AdMax, etc.). RPM is around ¥20, which only comes to ¥30 per 1,000 pageviews. Not only is that not real income, having other companies’ ads on the site makes a bad impression during AdSense review
- Going deep into programming-school programs. The payouts are high (up to ¥55,000), but review articles ranked by payout draw criticism easily and would damage the trust I’ve built through real records and disclosed numbers. I’ve limited myself to recommending only things I’ve actually used
That last call is the most important one. This site’s asset is the real data I disclose, and that’s not the kind of thing you can trade for one high-payout article.
Legal: no Specified Commercial Transactions Act notice needed, with conditions
If you run a site on ads and affiliate links only, you don’t need a notice under the Specified Commercial Transactions Act (Japan’s disclosure law for sellers). You’re an advertising medium, not a seller, so the disclosure obligation doesn’t apply.
However, the obligation kicks in the moment you start selling paid articles, apps directly, or e-books. At that point a business name isn’t enough; you have to publish your real name, address, and phone number. If you’re doing this as an individual, it’s a fork in the road you need to be ready for in advance.
Japan’s stealth-marketing regulation (under the Act against Unjustifiable Premiums and Misleading Representations, in effect since 2023-10-01) is clear. Any article involving compensation must carry a “PR” label, and ASP terms require it too. Violations carry up to 2 years in prison or a fine of up to ¥3 million.
I decided to enforce this with a mechanism. Write one line in an article’s metadata, and a PR label appears at the top of the body.
pr: trueDon’t build something where “forgetting to write it” makes you break the law. The only judgment call is “is there compensation?”; the labeling is left to the machine.
Paid work and sponsors didn’t come from pageviews
I couldn’t find any published threshold like “you start getting writing requests at X pageviews a month.” As far as I searched, no such standard exists.
What I could actually observe was this route: “an article takes off on Zenn/Qiita (Japanese developer blogging platforms) → someone at a company DMs you directly via X.” It’s X followers and recognition in the tech community that matter, more than your own site’s pageviews.
So I changed my approach. I’ll publish the same articles on Zenn too. My own site serves as an SEO asset and portfolio, and I’ll get exposure on the platform side. Zenn also has badges, a tipping feature (¥300–5,000), and book sales.
Where things stand now
| Item | Status |
|---|---|
| Articles | 21 |
| Privacy policy / operator info | Ready |
| Broken internal links | 0 (checked at build) |
| OGP, sitemap, robots.txt | Ready |
| PR labeling mechanism | Implemented |
| Custom domain | Not yet acquired |
| Analytics | Not yet installed |
| ASP registration / AdSense application | Not yet |
Everything left is not “research” but just “roll up my sleeves and sign up.” What I learned from spending half a day on research is that the hard part of monetization isn’t the review criteria, but whether you can keep publishing articles every week after meeting them.
When numbers come in, I’ll publish the actual amounts as they are.
I’ve already published everything on the cost side (Running 3 apps for a month: a ¥22 bill). I plan to publish the revenue side at the same level of detail.