Treat every amount as a number-and-unit pair
A printed amount is incomplete without its unit. Milligrams, micrograms, grams, and International Units describe different measurement conventions, so a content record should never store the numeral alone. Capture the label's original number, exact unit symbol, ingredient row, serving basis, punctuation, and source image. Preserve capitalization as printed, then map the symbol to a controlled unit field for sorting. This approach prevents 10 mcg from becoming an unexplained 10 in a comparison table and makes transcription errors easier to find.
Unit typography deserves deliberate review. The abbreviation mcg is commonly used for micrograms on consumer labels, while scientific references may use the micro symbol. A blurry image, seller transcription, or lost character can change what a record appears to say. Do not repair an unclear symbol from context alone. Mark it unreadable, seek a clearer panel from the same package version, and record where the replacement image came from. The evidence trail matters more than completing a row quickly.
Normalize compatible metric prefixes transparently
Metric mass units can be aligned for a documentary comparison when both labels clearly declare the same substance on the same basis. A system may retain the printed value while adding a normalized mass value, the conversion rule used, and a machine-readable unit. The table header must state the display unit, and the original declaration should remain visible. Normalization is an editorial convenience for like-for-like fields; it is not a judgment about how much is appropriate or whether one product is preferable.
The comparison must stop when the ingredient identity, serving definition, or amount basis differs. A mass value for a named nutrient is not automatically comparable with a mass value for a blend, extract, compound, or entire serving. Decimal movement is also a common source of error, especially when copying between milligrams and micrograms. Require automated dimensional checks and a second human review of every normalized value. If either check fails, publish the original units in separate columns rather than forcing a common figure.
Keep International Units outside generic mass conversion
International Units are not a universal mass unit. An IU statement can depend on the specific substance and the labeling framework that applies to it. A single sitewide formula that turns every IU value into milligrams or micrograms would therefore be misleading. Store IU as its own unit type, preserve any parallel declaration printed by the manufacturer, and identify the package and jurisdiction. Do not calculate a mass equivalent merely because a search result or unrelated product shows one.
When one label uses IU and another uses a metric mass unit, the safest default is a split comparison. Explain that the declarations do not share a directly normalized unit in this record. If an authoritative rule explicitly provides a substance-specific labeling relationship, that relationship belongs in a separately reviewed methodology with its source, effective date, scope, and rounding rule. This general literacy guide does not perform such conversions and never uses them to produce intake instructions or a product ranking.
Separate unit facts from serving and package facts
An amount stated in milligrams per serving cannot be compared with milligrams per capsule, scoop, packet, millilitre, or package until each label's denominator is confirmed. Record serving size and servings per container as separate fields beside the ingredient amount. Front-panel callouts may use yet another basis, such as an amount for the entire pack or a multi-item serving. Quote those statements independently instead of merging them into the facts-panel value.
Unit-bearing numbers can also appear in net quantity, other-ingredient descriptions, suggested-use text, and marketing claims. Their location changes what they describe. A robust extraction schema includes statement type and panel location, not just value and unit. Reviewers should be able to reconstruct why a number was classified as a declared ingredient amount rather than package mass. This prevents a large net-weight figure from entering an ingredient comparison and keeps informational fields distinct from directions.
Publish unit handling as part of the methodology
A trustworthy comparison discloses which units it accepts, which metric prefixes it can normalize, how it handles decimals and rounding, and when it refuses conversion. Show the printed value beside any normalized display value and label the latter clearly. Maintain test cases for leading decimals, thousands separators, mixed capitalization, unreadable symbols, and labels that state two units. Version the methodology so older pages can be traced to the rules used when they were produced.
Before publication, verify the package match, serving basis, ingredient identity, unit transcription, arithmetic, and source date. Use an explicit status such as directly comparable, metric normalization applied, different basis, IU retained, or unit unclear. These statuses give readers a useful explanation without implying a medical conclusion. The resulting page answers a narrow question about what the printed units mean for documentation, while leaving substance-specific interpretation and individual decisions outside its scope.
- Capture each printed number together with its exact unit symbol.
- Record ingredient identity, serving basis, and panel location.
- Retain the original declaration beside any normalized metric value.
- Normalize only compatible mass units on an identical basis.
- Keep IU outside generic mass-conversion logic.
- Flag unclear symbols instead of reconstructing them from context.
- Run dimensional and arithmetic checks on normalized fields.
- Publish the unit-handling rule and review date.
Limitations
- A shared unit does not establish that two ingredient declarations use the same basis.
- International Units do not have one universal mass conversion.
- Package images may be too unclear to distinguish a unit symbol reliably.
- This guide explains documentation conventions and does not calculate intake or suitability.
MIHEN / Sources
Sources
Sources mapped to the sections in this guide.
