Governance · evidence framework
Evidence standards: how the desk labels sources, certainty, and limitations
A reliable page explains not only what it says but what kind of evidence it is using. This standard defines how the site distinguishes checkout events, primary research, reviews, registries, and regulator documents.
Evidence categories used on the site
A cart event is a recorded storefront observation. A primary paper is an original research report. A systematic review synthesizes multiple studies using stated methods. A registry record describes a clinical study and its status. A regulator document provides jurisdiction-specific authorization, safety, or public-notice context. A laboratory document describes a specific submitted sample or method.
These labels are not interchangeable. A cart event cannot establish product safety. A registry entry cannot establish a final study outcome. A laboratory report cannot establish every batch or shipment. A research paper cannot establish current merchant inventory or code eligibility. Pages must state the correct category before drawing a conclusion.
Claim language follows source strength
When the source is a defined cart observation, the page may say that a documented cart result occurred. When the source is a study, the page may say that the study reported its stated finding in its stated population. When a record is unavailable, blocked, or preliminary, the page should say so directly.
The site avoids universal words such as guaranteed, proven for everyone, sitewide, best, or independently verified unless the evidence is specific enough to support them. In most cases, narrower language is both more truthful and more useful because it tells the reader what they can check for themselves.
Evidence before labels
The strength of a verification claim comes from the record behind it, not from the appearance of a badge. For a current VIP10 checkout status, the desk looks for a dated event containing the code tested, visible cart context, observed discount result, event time, and retained evidence.
Third-party badges, testimonials, promotional statements, community sentiment, and search-result descriptions may provide context, but they do not replace those required event fields.
Limitations are required, not optional
Every substantive page should carry the limitation that matters to its source type. Cart pages state that final checkout conditions remain merchant-controlled. Research pages state study and population limits. Archive pages distinguish old formats from evidence-led records. Policy pages identify what the site does not do.
The standard is designed for a deeper audit. A reviewer should be able to find a source, identify the relevant claim, see the evidence category, and understand the limitation without relying on an unexplained score, badge, or author credential.
A standard has to be usable in public
A written standard is only meaningful if a reader can apply it to a real page. For research material, that means the reader should be able to identify the source type, model or population, measured outcome, date, and limit. For a code check, the reader should be able to identify the product, cart state, visible response, timing, and whether the result stopped before payment. Abstract language about “rigour” is not enough.
The test is therefore practical: could an outside reader tell what would have changed the conclusion? A product becoming unavailable, a code displaying a rejection, a missing screenshot, an ambiguous gateway, or a later source can all change the record. A standard that names those conditions is stronger than one that only describes successful outcomes.