Annolyse
BriefingsCompaniesInsightsPrinciplesCompareChatWatchlist

Explore

  • Briefings
  • Companies
  • Insights
  • Compare

Resources

  • Search
  • Methodology

© 2026 Annolyse.

Methodology

How Annolyse reads NZX results

Annolyse converts public company filings into structured metrics, checks key fields before publishing, and then writes editorial analysis from those available figures.

Operating method

Read

Annolyse reads available company materials and records the disclosed figures in a consistent format where the source supports it.

Check

Publication is blocked when core fields are missing, units are implausible, formulas do not reconcile, or cross-statement inconsistencies appear.

Calculate

Derived metrics use one shared definition for each measure, including required inputs, missing-input handling, display format, and version.

Explain

Briefings and insights use canonical numbers, precomputed historical observations, and sector/source context to explain what changed, whether cash backed earnings, how balance-sheet risk moved, and what remains unresolved.

Review

If confidence or source-quality checks stay weak, the briefing stays out of public publication until it has been reviewed or corrected.

Calculation and validation notes

Formula discipline

  • Canonical metric definitions live in a central registry. Each registered metric declares its formula, required inputs, missing-input behaviour, display format, version, and downstream consumers.
  • Ratios are calculated from normalised fields only after the underlying values pass completeness and scale checks.
  • Growth rates use the prior comparable period where available, rather than mixing half-year and full-year values as if they were like-for-like.
  • Annualisation is used for context ratios such as half-year leverage, not as an earnings forecast.
  • Every public metric row may include qualityFlags. The field is empty when no known caveat applies and populated when the value is denominator-distorted, basis-discontinuous, transaction-distorted, non-comparable, insufficient-history, restated/reclassified, or an implausible outlier.

Comparison logic

  • Company pages show every published period for that issuer, ordered by result chronology.
  • Insight pages use each company's latest published result unless the page states a narrower window.
  • Sector comparison pages use the curated sector taxonomy and suppress metrics that are not meaningful for the issuer or source data.
  • Cross-company rankings and key metric tables consume the same qualityFlags contract: caveat values are labelled, suppress values render as n/m, and hard-suppressed values are not ranked or used as normal evidence.

Validation behaviour

  • Publication can be blocked when core fields are missing, units are implausible, or ratios depend on distorted denominators.
  • Historical analytical context is based on observations such as normal range, edge of range, above normal, below normal, or insufficient history.
  • Large metric movements receive caveats rather than being automatically treated as errors.
  • Known source-quality limitations are documented and revisited as Annolyse coverage improves.

Metric glossary

Profit and growth

Revenue

Reported operating revenue for the period, using the issuer's stated units and normalised into NZD millions when the filing provides a stable unit basis.

Revenue growth

Current reported revenue compared with the prior comparable period. Mixed half-year/full-year absolute values are not treated as like-for-like.

EBITDA-equivalent

The issuer's primary disclosed operating earnings measure. Examples: utilities often disclose EBITDAF, while telecommunications companies may disclose EBITDAI. Annolyse uses the issuer-labelled measure where it is the primary operating metric.

EBITDA margin

EBITDA-equivalent divided by revenue. Suppressed where revenue is missing, zero, or not a meaningful denominator.

PBT growth

Profit before tax movement against the prior equivalent period. Large swings receive a caveat where they may reflect one-off items or denominator effects.

Profit before tax

Reported profit before income tax. Used as the cleaner pre-tax bridge when NPAT is distorted by tax, discontinued operations, or unusual post-tax items.

NPAT

Net profit after tax attributable to ordinary shareholders, where the filing clearly reports this as a separate line item attributable to ordinary shareholders.

NPAT growth

Net profit after tax movement against the prior equivalent period. Large or distorted changes are treated as caveats, not recommendations.

Effective tax rate

Tax expense divided by profit before tax. Suppressed when PBT is negative or too small to produce a stable percentage.

Cash and balance sheet

Operating cash flow

Net cash generated from operating activities for the reporting period.

Cash conversion

Operating cash flow divided by EBITDA-equivalent. Values above 100% usually indicate working-capital release rather than a structural improvement in earnings quality.

Free cash flow

Operating cash flow less capital expenditure, shown before and after lease payments where the filing provides enough detail.

FCF pre-lease

Operating cash flow less capital expenditure before lease principal payments. This helps separate operating cash generation from lease-financing structure.

FCF post-lease

Free cash flow after lease payments where the filing provides lease-payment detail. Used as the stricter cash-flow measure for debt and dividend capacity.

Capex

Actual capital expenditure for the period, drawn from investing-cash-flow lines such as payments for property, plant, equipment, intangibles, or development assets.

Capex intensity

Capital expenditure scaled against revenue. It is useful for comparing reinvestment burden, but can be distorted by lumpy projects or asset sales.

Working-capital movement

The cash impact from changes in receivables, inventory, payables, and related operating balances where disclosed.

Debtor days

Receivables scaled against revenue to estimate how long cash is taking to collect. Used only where receivables and revenue are comparable.

Inventory days

Inventory scaled against cost of sales or revenue when the available data supports it. Suppressed where the business model makes inventory days irrelevant.

Net debt

Borrowings less cash and cash equivalents. Negative values indicate net cash.

Gross borrowings

Interest-bearing borrowings before cash offset. Trade payables and ordinary operating liabilities are excluded.

Net debt / EBITDA

Net debt divided by EBITDA-equivalent. Suppressed when EBITDA is negative because the ratio is not meaningful.

Net cash direction

Companies holding more cash than borrowings are labelled as net-cash. A decline in their cash position is not described as leverage deterioration; that framing applies to companies with net debt.

ROE

NPAT divided by current total equity where the filing discloses or supports both values clearly enough to calculate the ratio.

Total assets

Reported total assets from the balance sheet. Used as a completeness check and as context for property, infrastructure, and capital-heavy issuers. Required for full-year results. Half-year (interim) filings frequently omit a restated balance sheet under NZ IFRS — the most recent full-year balance sheet is carried forward unchanged and a fresh balance sheet is published with the next full-year release. When that happens the field is shown as n/a for the half-year period rather than blocking publication, and balance-sheet reconciliation is suppressed for that period rather than flagged as inconsistent. Period-over-period comparisons treat the half-year value as not applicable rather than zero.

Total equity

Reported equity attributable to owners where available. Used in ROE, price-to-book, and balance-sheet context.

Dividends and valuation

DPS

Dividend per share declared for the period, expressed in cents per share where the filing provides a per-share value.

Nil dividend

When no dividend is declared, Annolyse shows that directly rather than forcing a payout or coverage ratio.

Payout ratio vs NPAT

Declared dividend per share divided by basic EPS where both inputs are available on the same period basis. Extreme or non-meaningful ratios are suppressed rather than forced.

Annual payout ratio vs EPS

For full-year results, final dividend plus the interim dividend divided by full-year basic EPS where those inputs are available. This is separate from the currently declared dividend payout ratio.

Payout ratio vs FCF

Dividends divided by free cash flow. Suppressed when free cash flow is missing or not meaningful.

Dividend yield

Latest dividend per share annualised against the latest available market price where market data is available.

Market capitalisation

Latest available share price multiplied by shares on issue where both inputs are available.

Trailing P/E

Latest market capitalisation divided by trailing twelve-month NPAT. Full-year rows use the latest FY NPAT; half-year rows use latest HY plus the prior second half where the prior FY and HY are available. Negative, near-zero, or unsupported earnings are shown as not meaningful.

Forward P/E

Calculated only when a forward EPS input is available. If no forward EPS source is available from filings or market data, the metric is intentionally omitted.

Price to book

Latest market capitalisation divided by total equity from the latest published result where both values are available and comparable.

Price to FCF

Latest market capitalisation divided by trailing twelve-month free cash flow pre-lease. Negative, distorted, or unsupported cash flow is shown as not meaningful.

NTA per share

Net tangible assets per share where disclosed. It is most useful for property, investment, and asset-heavy companies.

Conventions

Notation and labels

  • n/m means not meaningful. It is used where a denominator is negative, zero, too small, or analytically misleading.
  • n/a means not available from the available filing data.
  • No dividend is used when no dividend was declared. Nil dividends are not labelled as covered or stable.
  • Suppressed values are intentionally hidden when inputs are incomplete, stale, negative, or too distorted to compare responsibly.
  • Metric quality flags are additive caveats carried through the public API, briefing tables, insights, rankings, and chat. Caveat-level flags add context; suppress-level flags hide values from normal comparison; hard-suppress flags prevent the value from being used as analytical evidence.
  • An analytical flag is a caveat or explanatory note that calls attention to an unusual value. It is not an error label and not an investment recommendation.
  • Needs review means the figures are not clean enough for public publication without further checking.
  • A calculation pass means displayed numbers reconcile to the underlying source fields. Advisory warnings are review signals, not automatic evidence that a displayed number is wrong.
  • Representative sample checks compare a cross-sector group against current source materials. Source gaps, scope drift, material drift, and missing statement coverage are tracked separately from arithmetic failures.

Units and signs

  • Money values are presented in NZD millions unless a page explicitly states another unit.
  • Per-share dividends are presented in cents per share.
  • Declared dividend payout and annual dividend payout are labelled separately when both can be calculated.
  • Cash outflows, finance costs, and losses preserve their economic sign where the source data supports it. Income tax is treated as a signed tax effect: tax expense or charge is negative, while a tax benefit or credit is positive.
  • Net cash is treated as a distinct state, not as negative leverage deterioration.

When Annolyse says not meaningful

  • Negative or near-zero denominators can make growth, payout, P/E, and leverage ratios misleading.
  • One-off statutory distortions can make NPAT-based ratios less useful than cash-flow or pre-tax measures.
  • Financial-services, property, and investment issuers may need sector-specific measures before cross-company ranking is fair.

Coverage and timing

  • Company and insight pages reflect the latest published Annolyse coverage set.
  • Newly published briefings may appear on individual company pages before showing up in sitewide listings.
  • Market-data-dependent views update when Annolyse has a valid recent quote for the company.

Periods and calendars

  • Prior equivalent period means half-year compared to half-year, or full-year compared to full-year. Mixed-period comparisons are avoided unless explicitly labelled.
  • Insight pages use the latest published result for each company unless the page explicitly states a different comparison window.
  • Expected dates are inferred from each company's historical reporting cadence, not from issuer guidance.
  • Half-year (interim) results under NZ IFRS commonly omit a restated balance sheet — the most recent full-year balance sheet is carried forward unchanged and the next refreshed balance sheet is published with the next full-year release. Annolyse treats balance-sheet fields (currently total assets) as not applicable for half-year periods when the source filing does not disclose them, shows n/a for the half-year value, and suppresses balance-sheet reconciliation invariants for that period rather than flagging them as inconsistent. Income-statement and cash-flow fields remain required on every period regardless of shape.

Publication rules

Publication safeguards

  • Briefings are published only when the required source and calculation checks pass.
  • When source materials are unusual or confidence is weak, publication is held back for review.
  • If a required check fails, the briefing is blocked rather than published with an unsupported number.

Publication timing

  • During active monitoring windows, Annolyse generally publishes a briefing within an hour of the NZX announcement being posted.
  • Timing can be slower when the source materials are unusual, a required check blocks publication, or market data is temporarily unavailable.
  • Newly published briefings can appear on company and briefing pages before every index or listing page has refreshed.

Source availability

  • If a field is not disclosed clearly enough in the available market and filing data, Annolyse prefers omission over producing a potentially misleading number.
  • Financial reports and results announcement forms are the primary financial sources. Results presentations are normally supporting context for segment detail, outlook, tax explanations, project pipelines, and management framing; they do not override statutory statement values when a proper report or statutory results announcement is available. If no primary report attachment exists and a presentation contains explicit statutory statement appendix pages, that presentation may be selected as the source material for that record.
  • Commentary documents are linked only when there is clear result linkage, a result-adjacent window, issuer-specific context, AGM or investor-day timing, or a source-backed material corporate transaction announced before a result. Ambiguous commentary is excluded rather than guessed into a briefing.
  • Movement screens rank the largest metric changes in the published coverage set and should be read as outlier detection, not investment advice.
  • Currency translation for dual-listed issuers preserves the reported currency unless a page explicitly states a conversion basis.

Limitations

  • No buy, hold, or sell recommendations.
  • No broker consensus, price targets, or management-call transcript analysis.
  • No forward-looking financial forecasts beyond what is explicitly disclosed in filings, presentations, or deterministically associated commentary.
  • Valuation views depend on available market data inputs.
  • Scanned PDFs with poor OCR quality cannot always be converted reliably.
  • Non-standard formats are excluded from publication when the available materials cannot support reliable figures.
  • Financial-services and property-trust issuers have balance-sheet structures that do not always fit the standard metric set.
  • Half-year packets for some issuers have thin treasury disclosures; full-year data is used as fallback where available.
  • Currency translation for dual-listed issuers preserves the reported currency rather than forcing NZD conversion.
  • No management-call tone scoring or social-media sentiment scoring.
  • Sector overlays affect interpretation and framing only. They do not change arithmetic, replace source values, or loosen publication checks.
  • No guarantee that every NZX issuer has enough structured data for every insight page.