A nutrition panel presents its figures as though they were measured from the item in your hand. They are averages drawn from sampled production, adjusted by rounding conventions and permitted tolerances that vary by jurisdiction.
The figures describe a batch, not a package
Manufacturers determine values either by laboratory analysis of representative samples or by calculation from the recipe using established composition tables for each ingredient.
Agricultural products vary naturally between harvests, growing conditions and suppliers, so the true content of any individual package differs from the printed average.
Regulators accept this by allowing a tolerance band around declared values rather than requiring exactness, and the width of that band differs between nutrients and between countries.
Rounding rules compress small quantities
Most labelling regimes specify how values must be rounded, and small amounts are typically rounded down to zero below a defined threshold.
This is how a product can legitimately declare zero of something while containing a small quantity of it, which matters most for nutrients consumed across many servings.
The rules exist to prevent false precision on quantities that measurement cannot reliably distinguish, but they interact awkwardly with serving sizes.
Serving size does more work than any other number
Everything on the panel is expressed per serving, and the serving is a defined portion rather than the amount a person is likely to eat.
A package can therefore contain several servings while presenting as a single item, and every figure on the panel reads far lower than the package total.
Comparing two products means comparing per-unit-weight columns where they are provided, because comparing per-serving figures can compare different quantities of food.
Energy figures come from assigned factors
Energy is not measured by burning each product. It is calculated by multiplying the mass of protein, fat, carbohydrate and other components by standard energy values.
Those factors are averages that ignore how much energy the body actually extracts, which depends on food structure, cooking and individual digestion.
Whole nuts, for example, yield less energy than their calculated value because some fat remains within intact cell structures and passes through undigested.
What the panel is still good for
Its strength is comparison between similar products rather than accounting for what a body received. Ranking three breads by salt or fibre is exactly what it does well.
The ingredient list, ordered by weight, often carries more information than the panel about what a product actually is.
Treating the numbers as a reliable ranking tool and an unreliable ledger keeps the label useful without expecting precision the system was never built to provide.