Why do grams and cups give different nutrition results?

Nutrition data lives in grams

Food composition databases report nutrients per 100 g of food, because weight doesn't care how the food was cut, poured or packed. 100 g of grated cheese is 100 g however fluffy the pile, while a cup of the same cheese can hold noticeably different amounts depending on how finely it is grated and how firmly it is packed. A range of roughly 80 g to 120 g is a reasonable illustration, not a standard.

To use a volume measure, you multiply the millilitres by the density of that particular food. The densities below are illustrative approximations, not fixed values.

grams = millilitres × density (g per mL)
Water:        250 mL × 1.0  = 250 g
Rolled oats:  250 mL × 0.34 = about 85 g
Honey:        250 mL × 1.43 = about 360 g

Same cup, three weights. If your software treats every cup as 250 g, the oats are overstated about threefold and the honey comes up about 30% short. The 0.34 g/mL for rolled oats is an approximate bulk density, which varies with flake size, how the oats are packed and the air between the flakes.

What one cup weighs

These are approximate weights and energy values for a levelled 250 mL metric cup, scaled from USDA FoodData Central entries. A US customary cup is approximately 236.6 mL, so American figures come out about 5% lower.

Flour scooped straight from the bag packs in more than flour spooned into the cup, so even the same food varies by how it's measured. Spoons vary by country too. Standard measuring tablespoons are generally 15 mL in the UK and US and 20 mL in Australia.

Measured units and estimated units

Every food in a nutrition database has its values stored against one base measurement, usually 100 g. Any other unit has to be converted to grams and scaled from that base, and software can do it in two ways.

• Food-specific units. These use a published or manufacturer-provided weight for a particular food, such as a cup of cubed watermelon or a labelled scoop of protein powder. They generally provide a more appropriate conversion than a generic estimate, although actual portions can still vary.
• Generic estimated units. These apply an assumed weight to a volume measure without accounting for the food's specific density. They may introduce substantial errors, particularly for foods that are very light, bulky or dense.

A cup of oats logged at an estimated 250 g comes to about 950 kcal, when a real cup of oats is nearer 320 kcal. For a very small portion, such as a splash of milk, the practical impact of a small conversion error may be limited. For anything light and fluffy, or dense and sticky, the error can be large, so good software tells you which kind of unit you're looking at.

Fix it once with a custom measurement

If a client always measures their cereal in the same mug, weigh one mugful and save it as a measured unit on that food, for example "mug, rolled oats" at 60 g. Saving a measured weight for that mug provides a more consistent conversion for future entries, although the amount can still vary depending on how full or tightly packed the mug is.

You can also move the base itself. A protein powder labelled per 30 g scoop can have the scoop as its base, so the label figures go straight in. How to add a food that isn't in the database covers the rest.

For the recipe maths, see how to calculate calories in a recipe. For greater precision, weigh energy-dense ingredients such as oils, nut butters, cheese, oats and rice. Volume measures may be sufficient for some lower-energy foods or liquids when exact tracking isn't necessary, provided the appropriate food-specific conversion is used.