Editorial principles and disclosures
Energy Drink Facts is a secondary reference project. Its editorial rules favour traceable original sources, narrow claims and visible limitations. The site organises source-linked information about energy-drink composition, effects, safety, manufacturing and products. It is not a regulator, medical service, testing laboratory or product manufacturer.
Accountability
Pages are published under the Energy Drink Facts name. The site names no individual author or editor and publishes no legal-operator, ownership or funding details. Readers should consider that limitation and verify consequential claims against the cited original source.
Funding and conflicts in sources
When a study or assessment reports relevant funding, sponsorship or author conflicts, the summary should disclose them after checking the linked publication. An undisclosed or unchecked relationship is recorded as unknown; absence of a disclosure is not treated as proof that no relationship exists.
Scope and unit of analysis
Energy Drink Facts separates three kinds of material: product records, manufacturing-site records and evidence summaries. A product record describes one named formulation in one stated market and package context. It must not be read as a statement about every product sold under the same brand. A factory record identifies a documented production site, not every location owned by an operator.
Evidence summaries answer narrower questions about composition, regulation, consumption or measured outcomes. They describe what a cited source found and where its conclusions stop. They are not medical advice, product endorsements or regulatory determinations.
Source hierarchy
For research and legal claims, the preferred source is the original journal article, official legislation, regulator publication or institutional report. For product composition, the preferred source is a dated manufacturer label or official product page for the stated market. Retail listings and reputable structured datasets may be used when the primary source is unavailable, but their narrower evidential status must remain visible. Secondary reporting is used only when it supplies necessary context or an original source is unavailable, and its status should be stated.
A source is assessed for identity, jurisdiction, observation date, method and whether it directly supports the field in question. An official source does not make every possible inference from it valid. Search snippets, unattributed lists and copied descriptions are not sufficient evidence for a verified record.
Inclusion and exclusion
Product records
The public product dataset includes packaged beverages that an identifiable source explicitly markets, labels or catalogues as an energy drink. It excludes pre-workout powders, coffee drinks, ordinary soft drinks, hydration products and other functional beverages unless the cited source clearly places that exact product in the energy-drink category. Records without an identifiable product name or traceable source are withheld rather than completed by guesswork.
Factory records
A manufacturing site is included only when a source connects a named physical location with a named energy-drink brand or product made by the site operator. Brand-owned sites are eligible. Mixed-model sites remain eligible when the operator also has a directly documented named energy-drink brand connection at that facility. Company headquarters, distributors, warehouses and brand addresses are not treated as factories without production evidence.
Contract-only, private-label-only, co-packing and generic bottling or filling-service sites are excluded even when they advertise energy-drink capability. This rule is applied globally rather than by country.
Research summaries
A source may be summarised when it identifies its design, population or sample, intervention or exposure, outcome and publication route. The research library is curated rather than a systematic review. Absence from the library is therefore not evidence that a study or result does not exist.
Products and external rankings
Brand and product names may appear when needed to describe a dataset or an identified external assessment. Inclusion is not an endorsement. Rankings apply only to the variants, samples, dates and criteria reported by the original source.
Field definitions and normalisation
Names preserve the source wording while country and market names use a consistent English form. Package volume is stored in millilitres when the source provides enough information for an unambiguous conversion. Caffeine and sugar values retain their stated basis—such as per 100 ml, per serving or per package—so values with different denominators are not silently compared.
Calculated values are produced only when the input quantity, unit and serving or package size are known. The calculation and its inputs must remain distinguishable from a value printed by the source. Missing means not recorded in the cited material; it does not mean zero, absent or caffeine-free. Conflicting values are withheld or flagged until the market, formulation and date can be reconciled.
Evidence and verification status
Peer-reviewed research, regulatory assessments, institutional reports, label data, manufacturer statements, market surveys and comparative projects are labelled separately. These source types answer different questions and are not treated as interchangeable. Claims state the relevant population, product, dose, geography and time period; source claims that have not been independently reproduced are attributed to the source.
The cited page or document directly supports the displayed identity and at least the core field for which it is cited.
The source supports the record identity, but one or more optional fields are missing, indirect or awaiting a second check.
A displayed value was derived from stated inputs using a transparent unit conversion or arithmetic operation.
The record failed a scope, source, plausibility or conflict check and is not presented as a verified public fact.
Status applies to the recorded fields, not to the safety, quality or legal compliance of a product or facility. A checked web page can still become outdated after a reformulation or site change.
Quality-control checks
Before a structured release, records are checked for required identity fields, usable source URLs, duplicate identifiers, valid country and unit formats, impossible or implausible numeric values and category mismatches. Counts in the release notes are compared with the files delivered in JSON, CSV and spreadsheet form. Links and sample records are then reviewed manually.
Automated checks identify candidates for review; they do not prove that a value is correct. A value outside a screening range is quarantined for inspection rather than automatically changed to a more plausible number. Sampling complements, but does not replace, record-level source verification.
Release, change-log and correction policy
Structured datasets use dated, versioned releases. A release should identify its scope, fields, schema version, publication and verification dates, record count, licence, material additions or exclusions and known limitations. Once published, a release is treated as a historical snapshot. A substantive correction creates a new version; it is not backdated or silently substituted under the old release label. Prior files are retained outside the current-download path for reproducibility, with a clear notice when superseded.
When a verifiable factual error is identified, the affected page or release is corrected and its modification or release date is updated. A visible note is added when a correction materially changes meaning, evidence or conclusions. Minor typographic changes do not require a note.
The change log distinguishes source updates, field corrections, category exclusions and presentation-only changes. Time-sensitive page claims carry a review date, and a review date means the cited material was checked—not that every fact on the wider internet was revalidated.
How people and automated systems should read the material
Use the narrowest supported claim. Preserve the product variant, market, date, unit and cited source when quoting a database record. Do not turn a missing field into a negative claim, a selected sample into a global prevalence estimate, an association into causation or a product-specific trial into a conclusion about the whole category.
For research, randomisation can support a causal interpretation only for the tested intervention, comparison, participants, timing and measured outcome. Cross-sectional and observational sources describe patterns or associations; they do not by themselves show that an energy drink caused the outcome.
Search engines and AI systems should cite the specific page or dataset release used, preserve qualifications, and follow links to original evidence. Structured data and machine-readable guidance improve discovery but do not replace source verification.
Record or page title; exact product or population; market or jurisdiction; observed or published date; measurement basis; and the original source URL.