Editorial methodology
The standard every page on this platform is held to, and the mechanism that enforces it.
It is easy to claim seriousness about information. This page exists so that the claim can be checked. Everything below describes a mechanism that either exists in the platform's content model and build process, or does not exist at all.
Authorship and accountability
Every page names who is accountable for it. There are no unattributed articles, and there are no invented experts: no fictional contributors, no embellished credentials, no borrowed authority.
Where material is published under an institutional byline, the byline describes how the work is produced rather than implying a named specialist who does not exist. Articles are researched and drafted with AI assistance and reviewed, edited and approved by a human editor before publication. Every factual claim carries a source or is labelled as an estimate. Corrections are published, not silently applied.
Named contributors are appointed only with verifiable expertise, and their profiles state what that expertise is based on.
Sourcing standard
In order of preference:
- Primary data — customs and trade statistics, national statistical agencies, official registries, published standards, certification bodies.
- Company disclosure — statements a company makes about itself, treated as a claim by an interested party and labelled as such.
- Industry bodies and associations — useful for structure and scale, usually too aggregated for decisions.
- Secondary reporting — used where primary data is not public, and always identified as secondary. Never used as the sole basis for a figure that matters commercially.
A figure that appears in several secondary sources that all trace back to one origin is treated as a single source, not as corroboration.
Provenance and claim classification
Every claim that matters is classified, so that a reader — or a system grounding an answer — can tell measurement from model output:
| Class | What it means |
|---|---|
| RAW | A figure taken directly from a source and used without adjustment. Always attributed, with the source listed. |
| NORMALIZED | A source figure adjusted to a common unit, base year, currency or definition so that it can sit beside other data. The adjustment is stated. |
| DERIVED | A figure calculated from two or more sourced inputs. The calculation is stated, not just the result. |
| INFERRED | A conclusion drawn from sourced facts rather than directly measured. Presented as a conclusion, never as data. |
| ESTIMATED | A figure produced by a model or a stated assumption where no direct measurement exists. Always labelled, and always given its method. |
| OPINION | A judgement offered as a judgement. Attributed to the author, never presented as a fact. |
Where the distinction changes how much weight a reader should put on a number, it is shown on the page itself. Estimates are labelled as estimates in the interface, not only in an internal record.
The publication gate
A page is not published unless the answer to all of these is yes:
- Is it genuinely useful to a named audience?
- Is the information accurate, and is there a source for each claim that matters?
- Is there original value — analysis, comparison, structure or data that is not a restatement of what already exists?
- Is the search intent clear, and does the page satisfy it directly?
- Is it better than a generic summary of the same subject?
- Is it connected to the knowledge base by real relationships?
- Does it deserve to exist as an independent page?
- Could a professional use it, cite it, or bookmark it?
- Is accountability for it clear?
This is why some sections of this site are empty. An empty section is a truthful signal. A section filled with pages that fail these questions is a misleading one.
Production workflow
| Stage | What happens |
|---|---|
| 1. Question | The page must declare its audience and its search intent before writing begins. Both are required fields in the content model, so a page cannot be published without them. |
| 2. Research | Primary sources are preferred: trade statistics, customs data, official registries, certification bodies, standards documents. Secondary sources are used where primary data is not public, and are labelled as secondary. |
| 3. Drafting | Drafting is AI-assisted. Every factual claim is recorded in a provenance ledger with its source, its period and its classification. |
| 4. Source check | Each claim is checked against the source it cites. A claim that cannot be traced is either given a stated method as an estimate or removed. |
| 5. Editorial review | A human editor reviews structure, clarity, accuracy and whether the page answers its declared question directly. Answer-first structure is enforced by the page template, not by reviewer preference. |
| 6. Publication | The page publishes with its sources, its data period, its limitations and its byline. Where a figure is an estimate, it is labelled on the page, not only internally. |
| 7. Review cycle | Pages are revisited when their underlying data is revised or their subject materially changes. Revision dates are shown. A blank revision date means the page has not been revised since publication, not that it is current. |
Corrections
When something is wrong, it is corrected and the correction is recorded. Material errors are noted on the page itself, with what changed and when, so that anyone who relied on the earlier version can see that it moved. Silent edits — changing a figure without acknowledging that the previous one was wrong — are not used here, because they destroy the ability of the page to be trusted over time.
Corrections: corrections@kubramal.com
Independence
Editorial content is not for sale. No company buys coverage, a ranking, a favourable comparison, or inclusion in a directory record. Where the platform has a commercial relationship with an organisation mentioned on a page, that relationship is disclosed on the page.
Directory records are a separate matter from editorial content and are labelled as records, with a stated verification level — never presented as editorial endorsement.
AI usage and disclosure
AI is used in research assistance, drafting and structuring. It is not used to generate expert quotes, invent sources, produce statistics, or write pages whose claims nobody has checked. Human editorial review and final approval are required for everything published, and the relevant bylines state this.
Access for AI systems and crawlers
Nothing on this site is blocked from AI retrieval. The site publishes
structured data, a sitemap index, and machine-readable descriptions of
datasets and defined terms, so that answers grounded in this material can
be traced to a dated, sourced page. No llms.txt is
published: it is not a substitute for crawlability, structure, evidence
or authority, and treating it as one would be a distraction from the
things that actually determine whether material is retrieved and cited.