A food label carries two columns and only one of them is comparable between products. Per 100 grams is fixed by regulation; per serving is a size the manufacturer chose, and it varies between similar products in ways that are not accidental. Comparing two cereals on their per-serving figures compares two different portion decisions rather than two foods.
That is why per 100 g is mandatory in the EU and UK, and why it is the column worth reading first.
How much can a serving size differ?
Enough to reverse a comparison. Two products with the same per-100 g energy can show a 40 per cent difference per serving if one declares 30 g and the other 45 g — and the smaller declaration usually belongs to the product that looks worse on the honest column.
The same effect runs through drinks, where a bottle frequently contains more than one declared serving. A 500 ml bottle declaring a 250 ml serving halves every figure on the label for a container most people finish in one sitting.
The defence is arithmetic rather than suspicion: multiply the per-serving figures by the number of servings in what you will actually eat, or ignore them and use per 100 g.
What do the reference intakes mean?
A single adult figure that applies to almost nobody exactly. The percentages on front-of-pack labels are calculated against a reference intake of 2,000 calories a day, which is a population average rather than a personal target.
Someone whose daily expenditure is 2,600 is reading percentages that understate their allowance by a third; someone at 1,700 is reading percentages that overstate it. The direction of error is systematic rather than random.
The figure is still useful as a relative measure — a product at 30 per cent of the reference for one portion is energy-dense whatever your own total is.
Which numbers are worth reading?
Fewer than the label carries, and the useful ones depend on what you are deciding.
| Reading for | Look at |
|---|---|
| Energy balance | Energy per 100 g, then per portion |
| Protein target | Protein per 100 g against the protein figure |
| Salt | Salt per 100 g — 1.5 g+ is high, 0.3 g or less is low |
| Added sugar | Ingredients list order, not the sugar row |
That last row is the one label reading most often gets wrong. The sugars figure includes naturally occurring sugars — the lactose in milk, the fructose in fruit — so a plain yoghurt and a sweetened one can show similar numbers for entirely different reasons. The ingredients list, ordered by weight, is what separates them.
Why does the ingredients list matter more than it looks?
Because it is ordered by weight and cannot be presented selectively. If sugar appears in the first three ingredients, the product is substantially sugar regardless of what the front of the pack says.
The technique worth knowing is ingredient splitting: dividing one sweetener into several — sugar, glucose syrup, fruit juice concentrate, honey — so that none individually ranks high enough to appear near the top. The total is unchanged and the list reads better.
Counting the distinct sweeteners is a faster read than adding them up, and it is usually enough to answer the question.
What about cooked versus raw weights?
They differ by a lot and labels are inconsistent about which they use. Dry pasta roughly doubles to two and a half times in weight when cooked, and rice more than doubles — so 100 g on a packet and 100 g on a plate are entirely different quantities.
Meat goes the other way, losing 20 to 30 per cent of its weight in cooking. A 200 g raw chicken breast is around 150 g cooked, and a label that quotes per 100 g raw will overstate what a cooked portion delivers if read the wrong way round.
The portion article covers the dry-to-cooked ratios for the staples where this matters most.
What do front-of-pack colours add?
A fast read of four figures — fat, saturates, sugars and salt — against thresholds set per 100 g. Red does not mean avoid and green does not mean unlimited; the colours are density measures, so a small portion of a red-flagged food can contribute less than a large portion of a green-flagged one.
The system also says nothing about protein, fibre or micronutrients, which is why a product can be entirely green and nutritionally uninteresting. It is a screen for the four things most often over-consumed rather than a summary of what a food contains.
Questions people ask
Is per 100 ml the same idea? Yes, for liquids, and it is directly comparable between drinks in the way per-serving is not.
Are the figures accurate? Within a legal tolerance that is wider than most people assume — typically several per cent, and larger for micronutrients. They are averages of a production run rather than measurements of the packet in your hand.
Does the energy figure account for fibre? Partly. Different regulatory regimes treat fibre and polyols differently, which is one reason the same product can carry slightly different figures in different markets.
Should I weigh food to use these numbers? For a few days, to calibrate what a portion looks like. Permanently, only if the precision is genuinely serving a purpose — the deficit article covers where the real error in a plan usually sits, and it is rarely in the weighing.
One column is regulated and the other is a decision. The calories per serving calculator converts a recipe into both, the macro calculator turns a target into grams, and the protein calculator gives the one figure with an evidence-based number attached.