Choosing a Japanese SEO vendor has a structural problem: you cannot evaluate the deliverable. The work is in Japanese, the evidence is in Japanese, and the failure mode — content that is grammatically perfect and commercially useless — is invisible from the outside for about six months.
We are one of the vendors you might be comparing, so read this accordingly. We have tried to write the version we would want if the positions were reversed, including the parts where we are the wrong answer.
The three kinds of vendor
| Type | Good when | Watch for |
|---|---|---|
| Global agency with a Japan offering | You need one contract covering many markets, and procurement requires scale | Japanese-language research capability is often thinner than the pitch suggests. Ask who physically does the keyword research, and in what language. |
| Japanese domestic agency | You need depth in the Japanese market and have someone bilingual internally | English-language communication and reporting can be genuinely difficult. Contracts and review cycles may assume Japanese business norms. |
| Small bilingual practice | You want senior attention and native production without an account layer | Limited capacity. No bench if someone is unavailable. Cannot absorb a large multi-market programme. |
We are the third type. If your requirement is one global contract across twelve markets with 24-hour coverage, the first type is the right answer and we would tell you so.
Seven questions worth asking
1. Who does the keyword research, and in which language?
The answer you want describes a person working from Japanese suggest and query data. The answer that should worry you describes translating your English keyword list and checking volumes.
Follow-up: ask how they handle the same term written in kanji, katakana and hiragana. A vendor who has not thought about script variants has not done Japanese keyword research.
2. Who writes the Japanese, and are they the same person who researched it?
Research and writing splitting across a language boundary is where most Japan programmes quietly fail. Ask to see the chain.
3. What happens in YMYL categories?
If your category touches health, money or legal standing, ask who supervises the content and what their qualification is. In Japanese YMYL search, named professional supervision is close to table stakes. A vendor without access to licensed supervisors will simply not be competitive there.
4. How do you measure AI visibility?
AI Overviews launched in Japanese in August 2024 and AI Mode in September 2025. A vendor still reporting only rankings is measuring a shrinking fraction of the outcome. Ask what they track, how often, and to see an example report.
5. What did you record before you started?
Any competent vendor records a baseline — clicks, impressions, positions, AI citation presence — before touching anything. If nobody mentions a baseline, you will not be able to evaluate the engagement later, which is sometimes the point.
6. When would you tell us not to do this?
A vendor who cannot name a scenario in which content SEO is the wrong investment for Japan has not thought about your business. There are several real ones: no Japanese search demand for the category, a revenue requirement inside one quarter, or no capability to answer enquiries that arrive in Japanese.
7. Do you optimize separately for Yahoo! JAPAN?
This one is a diagnostic. Yahoo! JAPAN’s organic results come from Google’s index — there is no separate ranking algorithm. A vendor selling “Yahoo! JAPAN SEO” as a distinct organic service is either behind or hoping you do not know. The correct answer distinguishes organic (same) from Yahoo!’s shopping, local and advertising systems (genuinely separate). Detail in Yahoo! JAPAN SEO.
Red flags
- Guaranteed rankings. Not deliverable, anywhere, by anyone.
- Link packages. Paid link schemes are still sold in Japan. They carry the same risk here as everywhere else.
- Volume article plans. “100 Japanese articles a month” is a scaled content abuse problem with an invoice attached.
- “We’ll translate and optimize your existing content.” This is the specific thing that does not work in Japan, sold as the solution.
- Yahoo!知恵袋 posting as a service. Manufactured Q&A posts. Japanese users spot them, and the downside is reputational.
- Market share figures from 2024 in a 2026 proposal. If a deck says Google 80% / Yahoo! 9%, nothing in it has been re-checked.
How to check quality when you cannot read Japanese
Three things that work without Japanese ability.
- Ask for one sample outline before signing, then have a Japanese speaker you trust — a customer, a distributor, an employee — read it. Not a translator. Someone who buys in your category. Ask them one question: does this sound like it was written for Japan, or converted into Japanese?
- Read the vendor’s own Japanese site. If their Japanese-facing content is thin while their English pitch is polished, you have learned where the capability sits.
- Ask for Search Console evidence, not a case study PDF. Specific numbers, over a stated period, with the starting point included. “56 to 751 monthly clicks over six months” is checkable. “Significant growth” is not.
On price
Native Japanese production costs more than translation, and there is no version of this where it does not. You are paying for Japanese-language research, a writer who can produce publishable Japanese, and in regulated categories a licensed professional to supervise.
The cheapest quotes in this market are usually cheap because one of those three is missing. Ask which one.
Common questions
Should we hire in Japan instead of using an agency?
If Japan is a long-term core market, eventually yes. A single in-house hire rarely covers research, writing, technical and supervision at once, so most companies run a hybrid for the first year or two.
How long before we can judge the engagement?
Two quarters. Judging at 90 days is the single most common reason Japanese programmes get cancelled just before they work.
Does the vendor need to be physically in Japan?
The people doing the Japanese-language work do. The account contact does not.
If you want to talk to us
We are a small Japanese SEO practice. Since 2020 we have worked with 129 clients across 44 industries, all Japan-based, including Japanese subsidiaries of global companies. We produce over 3,000 Japanese articles a year, with more than 250 licensed experts supervising YMYL content, and we monitor AI citations weekly on our own media.
If Japanese content SEO is the wrong investment for your category, we would rather say so than sell you a retainer. More about how we work.
Last reviewed: August 2026.
