Biznax uses these research standards to guide the preparation of supplier profiles, buying guides, comparisons, glossary entries, templates, and tools. Research should help readers understand what the available evidence supports, how it was interpreted, and which questions remain unresolved.
Define the question and scope. Before collecting information, identify the purchasing question, relevant industry, product category, geography, and time period. A profile of one legal entity must be distinguished from coverage of its parent group, brands, distributors, or individual facilities. Research should not imply broader coverage than was undertaken.
Choose sources by their purpose. Official business registers are useful for registered identity. Company websites, catalogs, and technical documents establish what a supplier publishes about its offerings. Certification bodies and relevant scheme databases may help examine certificate claims. Technical standards, regulators, and authoritative professional resources support explanations within their respective scopes.
A source’s authority depends on the claim being examined. A company catalog may support a description of its advertised product range without independently establishing production capacity, performance, or compliance.
Trace information to its origin. Directories, search results, press coverage, and social posts can identify research leads. Important details should be followed to their original source where possible. Several websites repeating the same description do not constitute several independent confirmations.
Search snippets, inaccessible documents, and AI summaries must not be presented as material that has been examined in full. If an important source cannot be accessed or its meaning cannot be established, qualify or omit the claim.
Document the evidence. Research notes should identify the source, page or document title, URL, publication or revision date where available, and access date. Record the relevant statement, its context, and the field or conclusion it supports. Distinguish information supplied by a company from additional corroboration.
For changeable details, preserve enough context to explain when the information applied. Prices require currency, quantity, units, and conditions. Lead times require a defined starting point and scope. Capacity figures should identify the product, facility, and period concerned where these are known.
Examine certification claims carefully. Record the named holder, issuing body, certificate number, standard or scheme, covered activities and sites, and validity information where available. A logo or undated certificate image alone is insufficient. Checking a management-system certificate must not be presented as testing or approving every product sold by the holder.
Resolve differences transparently. Conflicting names, dates, addresses, specifications, or figures require examination of source quality, recency, and scope. Prefer an explanation supported by evidence over an assumption that conveniently fills a field. If a conflict remains unresolved, state the uncertainty.
Absence of an accessible record does not itself establish misconduct. Equally, registration or an identity-check label does not establish product quality or commercial reliability. Identity decisions follow the separate Identity Check Methodology.
Make comparisons reproducible. Identify selection criteria, measurement units, assumptions, and missing data. Distinguish supplier-reported specifications from independently obtained results. Claims of product testing, interviews, or facility visits require records of work actually performed. Hypothetical examples must be clearly recognizable.
Calculations and tools should be checked against known examples, including boundary cases and invalid inputs. Review unit conversions, formulas, rounding, and the treatment of unknown values. Outputs must communicate their assumptions and relevant limitations.
Apply human review. AI may assist with discovery, organization, or drafting under the AI Policy, but its output is not evidence. Before publication, review the support for important claims, the accuracy of attribution, and whether conclusions exceed the research performed.
Research should be revisited after credible error reports or significant changes. Material corrections follow the Corrections Policy. Readers can submit evidence through the Contact Us page to help resolve inaccuracies or gaps.