Insights · Auditing a Japanese site
A Japan website SEO audit you can run yourself: ten checks, four states, and where it stops
Before you pay anyone to audit your Japanese pages, run this. It takes an afternoon, needs no Japanese, and catches the faults that stop a page being found at all. It also shows you exactly where a checklist stops being enough.
Published 2 September 2026. Technical definitions from Google Search Central documentation (canonical, hreflang, robots meta), confirmed 2 September 2026. The four-state convention and the order of checks are SuzuLabo's own, as used in the written Snapshot.
Why the first pass should be yours
Most Japanese pages we are asked to look at fail on something a non-Japanese speaker could have found: a canonical tag pointing at the English version, a noindex left over from staging, a form that errors in English. None of that needs a specialist. It needs a list and an afternoon.
Doing the first pass yourself also changes what you buy afterwards. If the mechanical faults are fixed, a paid review can spend its time on the things that do need Japanese: whether the page answers the query a Japanese buyer types, whether the proof reads as proof, whether the form asks for the right things. That is a better use of a fixed fee than being told your sitemap is missing.
Four states, so the result is a list of decisions
We give every finding one of four states, and we suggest you do the same, because a list of observations is not an audit until each line tells you what to do.
The ten checks
Run these on the three to five Japanese URLs that matter: the Japanese home page, the main service or product page, and the contact route. Each check needs a browser and, for a few, the page source.
| Check | How to do it yourself | What a bad answer looks like |
|---|---|---|
| Does the URL return 200? | Open it in a private window, or run curl -I on it. | Redirect chains, 404s on the Japanese path, or a 200 that shows the English page. |
| Is there text without JavaScript? | View the page source (not the inspector) and search for a sentence from the page. | The body is empty or a loading placeholder. Crawlers that do not execute scripts see nothing. |
| Is indexing allowed? | Search the source for noindex; check robots.txt for the path. | A staging-era noindex left on the live Japanese section. |
| Is there a self-referencing canonical? | Search the source for rel="canonical". | The Japanese page's canonical points at the English page. Google is being told the Japanese page is a duplicate. |
| Is the language declared? | Look at the <html lang> attribute. | lang="en" on a Japanese page, or no attribute at all. |
| Do the language versions point at each other? | Search for hreflang in the source of both pages. | Only one direction, or codes such as jp (not a language code) instead of ja. |
| Is it in the sitemap? | Open /sitemap.xml and search for the URL. | The Japanese pages were added after the sitemap was generated and never made it in. |
| Does the title say what the page is, in Japanese? | Read the <title> in the source. | An English title on a Japanese page, or the company name alone. |
| Is the company identifiable? | Look for a Japanese company name, address and a contact route on the page or one click away. | Nothing but a logo and a form. Japanese B2B readers look for this before they write. |
| Does the form work end to end? | Submit a test inquiry and check the confirmation, the auto-reply and where the message lands. | Validation errors in English, no confirmation, or a reply that reaches nobody. |
Two notes on method. Use a private window, so your own cookies do not change what you see. And record the date next to each answer: a check without a date cannot be compared with the next one.
Where the checklist stops
Everything above is about whether the page can be found and used. None of it tells you whether it will be, because that depends on things a checklist cannot see without Japanese:
- Whether the page targets a query anyone types. A Japanese page built from a translated keyword list can pass every check above and target nothing. The reasons are in Japanese keyword research: four assumptions that quietly break it.
- Whether the Japanese reads as written for Japan. Register, terminology, the order information is given in. A native reader sees it in seconds; a checklist never does.
- What ranks for the query now. Opening the Japanese results page and reading what kind of page Google is rewarding is the step that turns a list of faults into a plan.
- Whether the proof works. Case studies, client logos and certifications carry different weight in Japan by industry. Which of yours travel is a judgment, not a check.
That second half is what the written Snapshot covers, and the reason it is priced as a fixed piece of work rather than an hourly audit: the first half is already done by the time you send it, or can be.
What the first pass looked like on our own site
We ran the same ten checks on our own English pages on 2 September 2026, twelve URLs, as part of a routine review. Every URL returned 200, every page had a self-referencing canonical and a declared language, the two pages that are deliberately not indexed carried noindex, all indexable URLs were in the sitemap, and the form had been tested end to end in August. Nothing was Blocked. Several items were Needs work and had nothing to do with the checklist: description tags that ran past 160 characters, and a company location that was in the structured data but not in the visible text, which an AI assistant had filled in wrongly.
That is the pattern to expect. The mechanical checks clear quickly. What remains is content and judgment, and that is where the second half of the work sits.
FAQ
Frequently asked questions
Do I need to read Japanese to run these checks?
No. All ten checks work from the page source, a browser and a test submission. Reading Japanese becomes necessary at the point the article marks: query targeting, register, what ranks, and proof.
Which pages should I check first?
The Japanese home page, the one page that describes what you sell, and the contact route. Three URLs cover most of the risk; five is plenty for a first pass.
What is the single most common Blocked item?
A canonical tag on the Japanese page that points to the English page. It is invisible to visitors and tells Google the Japanese page is a duplicate. It usually comes from a template that was never localized.
Is a "no JavaScript" check still relevant?
Yes. Google renders JavaScript; not every crawler does. If the page source has no text, you are relying on rendering you cannot verify from outside. It is a Blocked item when the page matters.
How often should this be repeated?
After every change to the Japanese section, and otherwise quarterly. Keep the dated answers; the comparison is the value.
What does the paid Snapshot add to this list?
The second half: Japanese query targeting, a read of the results page for your queries, the register and proof judgment, and a prioritized list of about fifteen actions with a 30-day order. The illustrative sample shows the format.
If you have run the ten checks and want the second half done
Send the URLs and what you found. We will reply in writing with whether the pages are ready for the second half, what we would look at first, and a quote if you want it done. That reply is free.