Last reviewed: July 2026

Serving-size games: how brands shrink the numbers

In the EU nothing stops a manufacturer from declaring whatever serving size it likes, so the per-serving numbers on the front of a pack are a marketing choice as much as a nutritional one. Real examples from the Dietly database include an 18 g wafer bar, a 21 g cheese slice and a two-thirds-cup ice cream scoop. Each is small enough to keep the headline calorie figure under a comfortable threshold. The fix takes two seconds: read the per 100 g column instead.

Who decides what a serving is

Two systems exist. In the European Union, Regulation (EU) No 1169/2011 makes the per 100 g declaration mandatory and allows an extra per-portion column, but it does not fix portion sizes, so the manufacturer decides. In the United States the Food and Drug Administration sets reference amounts customarily consumed for each food category, and the 2016 label update raised many of them to match how people actually eat, adding a dual column for packages holding between one and two servings. The US system is tighter. Neither system stops a small serving being flattering.

Six real servings against a realistic portion

Serving figures are the per-100 g label values scaled to the stated serving weight. The last row is the reverse case: a generous declared serving makes a food look worse per serving than it is.

The three moves

• Shrink the serving. The 18 g wafer bar exists so the panel can print a round 100 kcal. Nobody eats one. The same logic produces the 30 g chocolate serving and the 25 g biscuit portion.
• Use an awkward unit. Two-thirds of a cup, 2.5 slices, "about 12 pieces". Fractions of volume measures are hard to picture and impossible to eyeball, so the number stays abstract. Volume units are also unreliable for anything that can be packed loosely, which is why portion size and serving size drift apart in practice.
• Let the pack hold 1.4 servings. A single-sitting bottle or bag declared as one and a bit servings means the front-of-pack number is not the number you consumed. The US dual-column rule was written specifically to close this gap.

Why per 100 g undoes all three at once

Whatever the brand chose, the per 100 g column is fixed by law and identical across products, so it is comparable and it cannot be gamed. Once you think in per 100 g terms, pistachios at 607 kcal and ice cream at 195 kcal land in the right order immediately, even though a serving of pistachios prints a smaller number than a serving of ice cream. This is the same reasoning behind energy density, and it is the reason every Dietly food page leads with the per 100 g column.

Calibrate once, then stop weighing

You do not need a scale forever. Weigh the five or six foods you eat most, once, and note what your normal portion actually is: the cereal you pour, the peanut butter on your knife, the cheese in your sandwich. Most people are surprised by two of the six and accurate on the rest. After that you can estimate, because you are estimating against a measured memory instead of against a marketing figure. For anything you eat rarely, the per 100 g column plus a rough weight is close enough, and the comparison tool will do the arithmetic for two products at once.

Bottom line

Serving sizes are chosen, not discovered. Treat the per-serving column as a claim about the brand's imagination and the per 100 g column as the fact, then weigh your own usual portions once so you know which of the two you are living in.