Why the Numbers on Your Snack Label Don't Add Up

Try this the next time you are holding a packet of biscuits. Take the protein, carbohydrate and fat in grams, multiply protein and carbs by 4, multiply fat by 9, and add them up. You will almost never land exactly on the calorie figure printed at the top of the same panel.

That gap is not a misprint, and it usually is not a company trying to fool you. It is the product of three things most shoppers never see: how the number was produced in the first place, how it is rounded, and how much legal room a manufacturer has before the number is considered wrong. Understanding all three makes you much better at reading a label - and much harder to mislead.

The label is a calculation, not a measurement of your packet

The first assumption worth dropping is that someone tested the exact packet in your hand. Nobody did.

Singapore's guide to nutrition labelling for food products, issued under the Ministry of Health and Health Promotion Board, recommends that declared values be established by direct chemical analysis using AOAC official methods, ideally at an accredited laboratory. But that analysis is done on a sample, at one point in time - not on every batch that follows.

In practice many manufacturers calculate rather than test. As a 2020 study in Food Chemistry notes, European law explicitly permits declared values to be average values derived from a manufacturer's analysis, or from a calculation using known average values of the ingredients, or from generally accepted composition data. Chemical analysis is expensive and slow; spreadsheets are neither.

So the number on the pack is best read as a careful estimate of a typical packet, not a reading taken from yours.

The calorie line runs on a 120-year-old shortcut

The 4-4-9 arithmetic you just tried has a name. It is the Atwater general factor system, developed at the end of the nineteenth century, and according to the FAO's technical report Food energy - methods of analysis and conversion factors it assigns 17 kJ/g (4 kcal/g) to protein, 37 kJ/g (9 kcal/g) to fat and 17 kJ/g (4 kcal/g) to carbohydrate, regardless of which food they came from.

That last part is the catch. The same FAO report sets out food-specific factors that vary enormously: protein ranges from about 2.44 kcal/g in some vegetable sources to 4.36 kcal/g in eggs, and fat from 8.37 to 9.02 kcal/g. Applied across a typical mixed diet, the general factors produced energy values roughly 5 per cent different from the food-specific ones.

Five per cent on a 150 kcal snack is about 7 kcal. Not worth losing sleep over - but worth knowing that the calorie line is a modelled figure, not a measured one.

Rounding quietly absorbs the rest

Then the numbers get tidied. The Singapore labelling guide is specific: energy, cholesterol and sodium are rounded to the nearest whole number, and the remaining nutrients to one decimal place.

Rounding conventions are also why "0 g" is a slippery word. The European Commission's December 2012 guidance on tolerances allows fat, protein and carbohydrate to be declared as "0 g" at or below 0.5 g per 100 g, and saturates as "0 g" at or below 0.1 g per 100 g. A pack can honestly print a zero while containing a small, genuine amount - and across a multi-serving pack those small amounts add up.

The tolerance band is wider than most people assume

This is the part that surprises people. Labels are not held to the exact printed figure. Singapore's guide sets out nutrient verification criteria used in post-market surveillance:

  • Naturally occurring nutrients (vitamins, minerals, protein, fibre, unsaturated fats): must be at least 80 per cent of the declared value.
  • Nutrients added to fortified foods: at least 100 per cent of the declared value.
  • Nutrients we are advised to cut back on - calories, sugars, total and saturated fat, trans fat, cholesterol, sodium: must be under 120 per cent of the declared value.

Read that last line again. A snack declaring 200 kcal and 300 mg sodium per serving can, in principle, test at just under 240 kcal and just under 360 mg and still be compliant. The EU's guidance is built on similar logic, allowing roughly plus or minus 20 per cent for sodium at or above 0.5 g per 100 g and for saturates at or above 4 g per 100 g.

Does the gap show up in real products?

Sometimes, yes. The Food Chemistry study - Albuquerque and colleagues, 2020 - chemically analysed 209 prepacked foods bought in Portugal, including snacks, bakery items and potato products. Only half of the samples declaring 1.25 g of salt per 100 g or more fell within the permitted tolerance for salt. For saturated fat, 27 per cent of samples sat outside the limits. Fat performed best, and cereal products were fully compliant for both fat and salt.

Two honest caveats: that was a European sample collected in 2015-16, not a Singapore one, and it says nothing about any specific brand on our shelves. But it does puncture the idea that the printed number is exact.

How to use a label anyway

None of this makes nutrition panels useless. It changes how much precision you should read into them.

  • Compare per 100 g, not per serving. Serving sizes are chosen by the manufacturer; 100 g is fixed. It is the only fair basis for comparing two packets.
  • Treat small differences as noise. Between two crackers at 480 mg and 505 mg sodium per 100 g, there is no meaningful difference. Between 480 mg and 900 mg, there is.
  • Be sceptical of a bare "0 g". It means "below the reporting threshold", not "none".
  • Read the ingredients list too. Order by weight is far harder to round away than a decimal place, and it tells you what the food actually is.

The panel is a reliable guide to the general shape of a food - whether it is salty, whether it is mostly refined starch, whether there is real protein or fibre in it. It is not a precision instrument, and it was never built to be one.

Where we sit on this

We would rather you read our panel sceptically than take it on faith. HexaCrunch crackers are baked brown rice crackers made with cricket protein, and we print the numbers per 100 g precisely so you can line them up against whatever else is in your trolley. If you want to run that comparison yourself, the Try-All 3-Pack is the easiest place to start.

Sources

Back to blog