Thirteen things an English reader can’t find on your site
Ordered by what they cost you, not by what they score. The first four take an afternoon and no budget. Every figure on this page comes from the 128 TSE-listed companies we have reviewed on all 34 questions of the same published rubric, and is recomputed whenever a score changes.
Read this as a checklist, not a verdict
We publish no league table. Each company page lists every question, the score, the document we used, and the date we looked — so that when we are wrong, you can prove it. Our change log is public and records every score we have changed, including the ones changed against us.
One example of what that means in practice. We drafted a rule that would cap any company publishing materially incorrect data in English. Applied to the first 21-company index, it caught three companies — all three because they had published English correction notices. The rule punished exactly the behaviour it was written to reward, so we suspended it. That decision is recorded in our method document, not buried.
Publish an email address an investor can write to
One mailbox, one line on your English IR page. The role does not matter — IR, finance, corporate communications all count. What matters is that it is reachable without opening a PDF.
Most companies do name someone: in the header of the English tanshin. Almost nobody puts that person on the website. Of the 128 companies scored on this question, 97 sit at 3 or below. The 31 that reach the top band each link a dedicated IR contact from the English IR top page, where a reader sees it straight away. Top scores were re-checked in a browser in October, and some companies that had been scored 4 dropped to 2 — see the scoring rules change history.
Cost: one mailboxSay what your English version is actually for
Nearly every English filing carries the same line: in the event of any discrepancy, the Japanese original shall prevail. Fine as a legal backstop. But if that is the only status you ever give the English text, you are telling a foreign investor their version does not count.
Add one sentence to your IR page stating what the English disclosure is, when it is published relative to Japanese, and who maintains it.
Cost: one sentencePut every English document on one page, by date and type
Not a navigation menu — a list. Date, document type, language, link. An analyst building a file on you should not have to guess which of five sections holds the thing they need.
Cost: one pageClick your own English links
The most common single defect we find is an English page title resolving to a Japanese PDF. It is invisible to anyone testing the site in Japanese, and it reads to a foreign investor as though the translation was abandoned halfway.
Related, and worth an hour of a developer’s time: check that an English URL returns an English error page, and that nothing on your English site is labelled as machine-translated when it was in fact written by a person.
Cost: an afternoon of clickingExport the numbers you already publish as a spreadsheet
Five years of revenue, operating profit, segment figures and per-share data, in one Excel or CSV file. You have the data — it is in the tanshin you file every quarter. What you have not done is let anyone download it without retyping.
This is the single most-failed question in the rubric and among the cheapest to fix. Only 31 of 128 companies do it properly. If you want a differentiator that an analyst notices within thirty seconds, this is it.
Cost: a day of a finance analyst’s timeTranslate the AGM resolution results
You already translate the convocation notice, because you need overseas holders to vote. Then you publish 決議ご通知 and the 臨時報告書 carrying the actual vote tallies — in Japanese only. The foreign shareholder you asked to vote never learns the outcome.
It is one document, once a year, and the figures are already tabulated.
Cost: one translation, annuallyKeep the segment note in the English tanshin
Segment revenue and profit, with the definitions stated and held constant between periods. Where English tanshin are abridged, the segment note is usually the first thing cut — and it is the thing a sell-side analyst opens the document for.
Cost: stop abridgingPublish the English tanshin the same day as the Japanese
If you already translate it, the remaining gap is scheduling, not cost. Same-day is achievable — 77 companies already run a zero-day median lag across the five documents we track (the latest two tanshin, the AGM notice, the governance report and the securities report).
Watch the annual securities report specifically. Same-day on quarterly results and a three-month lag on the 有価証券報告書 is a common and easily-missed pattern.
Cost: schedulingReport progress against your own plan, in numbers
Put last plan’s targets and this year’s actuals side by side and explain the variance. Not a narrative of achievements — a reconciliation.
To be explicit about how we score this: beating every target scores just as well as missing one and explaining it. We reward the reconciliation existing, not confession. A company with nothing to apologise for is not penalised, and candour earns no bonus.
Cost: two pages in the integrated reportPublish the results briefing Q&A in English
You hold the briefing, someone takes the minutes, and the Japanese summary goes up. The English reader gets the slides without the questions — which is where the useful information usually is.
Cost: one summary per quarterBoard skills matrix and remuneration linkage, in English
The skills matrix is usually already drawn for the Japanese governance report; it needs translating, not creating. Remuneration is harder: the question is not the numbers but whether the link between pay and performance is explained well enough to be assessed.
Cost: translation, plus one drafting roundAn English annual securities report, in any form
We deliberately score this generously: a full translation, a summary, or an English section of the 有価証券報告書 all count. Most companies still score zero, and we recognise this is a real translation budget rather than a scheduling fix.
If a full translation is out of reach, an English summary of the risk factors and the business description captures most of what a foreign investor is looking for.
Cost: a translation budgetExplain the capital decision, not just announce it
A buyback with no stated hurdle rate, an acquisition with no quantified case, a cross-shareholding sale with no capital-efficiency rationale. The English reader can see that something happened, never why or with what result.
This is the sixth pillar of our next rubric version and is not yet scored against any company. We flag it now so it is not a surprise. The bar is not that the decision was good — we do not judge that. It is that the decision was explained with numbers, and that you later reported how it went.
Coming in v1.2 — not yet scoredIf we have this wrong, tell us
Every score we record names the document we used and the date we looked. If a document exists and we missed it, send us the link. We reply within 10 business days, correct the score if the evidence supports it, and record the change publicly with its date, its label and the previous score. Contact and corrections →
Corrections are free and unconditional. No company pays us, and a score can be corrected on evidence.