Data and methodology

Vizora is operated by MEEQUS (Shibuya Dogenzaka Tokyu Bldg. 2F-C, 1-10-8 Dogenzaka, Shibuya-ku, Tokyo 150-0043, Japan).

Vizora is an information service. It does not recommend buying or selling any security.

Japanese companies: annual securities reports (yuho) filed to EDINET, the FSA's disclosure system. We read the XBRL attached to each report. US companies: 10-K filings on SEC EDGAR.

Every figure keeps a link to the document it came from. Each page lists the document IDs, and the Markdown and JSON versions carry the same references.

We do not use stock prices or third-party data vendors, and we never type in real companies' numbers by hand.

XBRL elements are mapped to a common set of metrics (revenue, operating income, net income, cash flows and so on). Values that cannot be read directly, such as gross profit or free cash flow, are derived from the reported figures, and the page shows which ones are derived.

Per-share figures (EPS, book value per share and dividends per share), payout/equity ratios, issued and treasury shares, and shareholder counts are taken from reported XBRL facts. If payout ratio is missing, it is calculated as dividends per share divided by positive EPS. We do not include stock prices or calculate PER. When reported payout ratio differs from that calculation by more than one percentage point, the reported ratio is kept and a warning is recorded.

Before a year is published we check it: revenue minus cost of sales must equal gross profit, pretax income minus income tax must equal net income, operating income must not exceed gross profit, and operating cash flow minus capex must equal free cash flow. A year that fails is not published automatically.

Segment mix divides each segment’s revenue by total segment revenue, and nonnegative profit by the sum of positive segment profits. Losses are listed separately; difference is profit share minus revenue share in percentage points. Profit shares and margins require every segment’s profit and source confidence of at least 0.8. Older snapshots without source references omit these calculations. Single-segment companies omit the section.

Revenue/profit patterns compare comparable annual periods and include transitions between profit and loss. Growth quality is determined for profitable periods with revenue growth: profit growth exceeding revenue growth by more than 1 percentage point is margin expansion; the reverse is revenue-led; a difference within 1 point is similar growth. Operating leverage is profit growth divided by revenue growth and is omitted when revenue growth is below 0.1%. Period lengths and end-date gaps must be 330–400 days, with matching currency, accounting standard, and consolidation scope. Banks and insurers use ordinary income.

Cash conversion is operating cash flow divided by operating income. The cash/profit gap is operating cash flow minus net income. Cash ratio is cash and equivalents divided by total assets; cash-to-revenue is cash divided by revenue. Net cash is cash minus interest-bearing debt, and net debt to operating cash flow is (debt minus cash) divided by positive operating cash flow. If debt is unavailable, net cash is omitted; a nonpositive operating cash flow omits the net-debt ratio. Shareholder return rates use dividends plus buybacks divided by positive net income or free cash flow; dividend ratios are shown separately. CapEx intensity is capital expenditure divided by revenue. Investment relative to operating cash uses the absolute investing cash flow divided by positive operating cash flow. Ratios with nonpositive denominators are omitted. We do not show cash-quality metrics for banks and insurers because deposits and lending drive their operating cash flow.

The company model graph joins segment revenue to the income statement, net income to operating cash flow, and operating cash flow to capital expenditure, shareholder returns and cash change. The bridge is operating cash flow minus net income; it groups non-cash items and working-capital effects. Segment revenue may include inter-segment sales. The graph is omitted for banks and insurers and when required figures are missing.

Per-employee revenue and profit divide consolidated figures by consolidated employees. Average annual salary, age and tenure are reported separately for the filing company (parent only). We omit implausible employee counts above 3 million. The per-employee company rankings require revenue of at least ¥10bn and 100 employees; industry profit-per-employee medians require at least ¥1bn in revenue, 100 employees and three companies in the industry. Banks and insurers are excluded from operating-income measures.

Notable figures are selected by deterministic templates from published snapshots: segment concentration, large long-term changes, employee/profit divergence, shareholder returns, cash versus profit, and per-¥100 profit. Candidates are ranked by a fixed strength score and at most three are shown. Each candidate keeps its input references; no AI writes these statements. Industry-margin comparisons are omitted when a median is unavailable.

The multi-year comparison uses the earliest available point among up to six annual snapshots and the latest. It shows revenue, operating income and margin, employees, and cash when both endpoint values exist. Percentage changes are omitted if accounting standards, currencies, or consolidation scope differ; accounting-standard transition years are marked. Segment revenue shares connect by stable segment keys; share increases or decreases and unmatched keys appear as new or reorganized flows.

The segment portfolio chart uses revenue CAGR from the earliest available to latest annual disclosure for each current segment, latest operating margin, and latest revenue for bubble size. It is shown when at least two current segments have a reported operating margin. CAGR is shown only with at least two years of positive revenue and matching accounting basis. Segments without enough history remain visible without a CAGR; segments without a reported latest margin are omitted from the plot. When the chart cannot be shown, the page explains which data is missing.

The key-figures strip includes the largest segment by disclosed segment revenue, consolidated employees, parent-company average salary, and operating-margin difference from the primary industry's median when a benchmark exists. Banks and insurers omit the largest-segment and operating-margin comparison. The margin funnel shows only reported inputs that exist; gross margin is omitted without reported cost of revenue.

Balance-sheet components are normalized from the reported statements. Interest-bearing debt is derived only when short-term borrowings, long-term borrowings, and bonds are all present; missing components are never treated as zero. IFRS reports that disclose current and non-current interest-bearing-liability subtotals use their sum; lease liabilities are separately presented and excluded. Parent equity is derived from total equity less non-controlling interests when both exist. Some IFRS reports combine intangible assets and goodwill; the combined amount is kept separately and split only when goodwill is also reported. The assets and funding bars use reported components and show the residual as other; percentages are rounded to 100%. Banks and insurers do not expose these balance-sheet metrics in public company data because their statements use industry-specific formats.

ROA and ROE use net income attributable to parent divided by average opening and closing assets/equity when adjacent comparable annual snapshots exist; otherwise closing balances are used. Parent equity is preferred, with total equity as an approximation if parent equity is unavailable. ROE is omitted for nonpositive profit or equity and decomposed as net margin × asset turnover × financial leverage. Reported ROE is checked against the derived value; differences above one percentage point are warnings.

Asset intensity is property, plant and equipment divided by revenue; fixed-asset ratio is property, plant and equipment divided by total assets; goodwill-to-equity uses separately reported goodwill divided by parent equity. Goodwill is never inferred from a combined intangible-assets-and-goodwill figure. A multi-year goodwill change and cumulative investing cash flow may be shown side by side without asserting a causal relationship.

Geographic revenue is stored with its own XBRL dimension axis and member, separate from operating segments. Overseas revenue is derived as disclosed regional revenue less Japan/domestic revenue, and its share is calculated against company-wide revenue, only when geographic rows are disclosed and their sum reconciles to company revenue within 5%; otherwise overseas share is omitted. Geographic rows and ratios are calculated at Snapshot build time from the source facts.

"Where 100 yen of sales goes" rescales the income statement so that revenue equals 100 yen. It is a common analytical technique (a common-size income statement), shown as coins.

The 100-yen model is omitted when neither reported operating income nor the inputs needed to derive it are available. If a machine-readable operating-income fact exists but is absent from the validated published snapshot, the page says it is pending validation; otherwise it explains that no numeric disclosure is available. Markdown and the JSON API carry the same reason. Bank segment gross profit is not substituted for revenue because the measures are not comparable, and is excluded from industry revenue totals.

Industries are assigned per business segment, not per company, so a conglomerate appears in every industry it operates in. We use about 150 industries in 12 groups.

Each industry has a written definition (what it includes, what it excludes and where excluded businesses belong). An AI model reads the segment names and the business description in the report and assigns up to two industries per segment following those definitions. Staff can override any assignment, and overrides are never replaced automatically.

Industry totals add up the revenue of the segments in that industry. On the size map, a segment that belongs to two industries is split between them so the overall total is not double counted.

The five-axis company profile shows each company's percentile within its primary industry for revenue CAGR (up to six annual periods), operating margin, cash conversion, asset turnover, and net cash divided by total assets. The primary industry is the company's highest-revenue industry assignment. Financial industries are excluded from these axes, and an axis is unavailable when data or at least three eligible peers are missing. A dot plot shows each axis independently; no area or combined score is calculated, and industry rank is not investment advice.

AI writes short summaries and commentary from the figures above and from the text of the report. It never produces numbers: every number in an AI text is checked against the input data, and the text is not saved if any number does not match.

Texts are then scored by a separate judging model for unsupported claims and for boilerplate writing. Only texts that pass are published, and they are labelled as AI-written. Pages without such a text show a summary built mechanically from the numbers.

New annual reports are picked up after they appear on EDINET. Downloaded documents are stored and not fetched again.

If you find an error, please email support@meequs.com (the "Report it" link under Sources on each company page pre-fills the company and document ID). Corrections are recorded against the original fact so the history is kept.