Dry matter basis: how to compare pet food nutrient levels, with a worked calculation

01The short answer

Dry matter basis restates a nutrient as the share it would hold with all water removed, so foods of different moisture, such as kibble and canned food, can be compared fairly. Divide the as-fed value by (100 − moisture %) and multiply by 100. A can listing 9% protein at 78% moisture works out at 40.9%, ahead of a kibble listing 26% protein at 10% moisture (28.9%), although the label suggests the opposite.

The conversion corrects for water, not for calories. Nutrient standards such as FEDIAF’s are stated per 100 g dry matter and per 1000 kcal ME ¹, so finish with an energy check.

02What dry matter basis means

Every pet food is water plus dry matter, and only the dry matter carries protein, fat, fibre, minerals and vitamins. The guaranteed analysis on a US label is stated as fed, for the product as sold: minimum crude protein and crude fat, and maximum crude fibre and moisture. Under AAFCO’s 2023 model regulations, the guarantees for protein, fat, total carbohydrate, dietary fibre and moisture are each mandatory minimums or maximums ².

That matters for the arithmetic. A guarantee is a limit, not a measured average, so any conversion from a label is an approximation. Water dilutes every nutrient, which is why a can’s protein looks low beside kibble until the water is taken out.

03The dry matter basis calculation

The logic is a rescaling. The nutrient was measured in 100 g of food, but only 22 g of that is dry matter, so dividing by 22 restates it per 100 g of dry matter. Any dry matter calculator runs these steps, so one result checked by hand confirms it.

The formula also runs backwards: as-fed % = dry matter % × (100 − moisture %) ÷ 100. Formulators use it to turn a dry matter target into the as-fed guarantee the label must declare.

04Comparing wet and dry food fairly

Table 1 uses two hypothetical labels and repeats the calculation for each line.

Table 1. Illustrative labels, calculated by Guires PNL. Assumes 6% ash (kibble) and 2% ash (can), carbohydrate by difference, and modified Atwater factors of 3.5, 8.5 and 3.5 kcal/g for protein, fat and carbohydrate ⁵.

On the label the kibble leads by 17 percentage points; with the water removed the can leads by 12. On an energy basis the can still leads, with 98.4 g protein per 1000 kcal against 73.9 g, but the gap narrows from 1.42 times to 1.33 times, because the can’s dry matter is richer in fat. Since dogs and cats eat to meet their energy needs, the per-1000 kcal figure is the one to compare with a standard ¹.

05Reading the new US label

AAFCO members approved new model pet food regulations on 31 July 2023, which AAFCO called the first major label update in more than 40 years, with six years of enforcement discretion counted from the availability of the printed 2024 Official Publication ³. The Pet Nutrition Facts box gives calories per household unit and the calories from protein, fat and carbohydrate in that unit, and “kcal ME” will no longer appear on labels ².

The box is still stated as sold, so formats still need moisture correction. Hill and Dzanis give a shortcut to protein per 1000 kcal straight from the box ²:

Because the energy basis already removes water, this route lets a buyer compare a can and a kibble without converting moisture at all. See AAFCO label modernisation for the rest of the box.

06Common mistakes

• Comparing formats as fed. Convert both foods first. Two cans of similar moisture can be compared directly.
• Treating guarantees as actual values. A minimum protein of 26% may mean 28% in the bag. Use laboratory moisture and proximate results where you have them; laboratories measure moisture directly by drying to constant weight ⁶.
• Converting protein only. Use the same moisture figure for fat, fibre and minerals so the whole profile is comparable.
• Mixing moisture figures. One food, one moisture value, from one batch or source.
• Stopping at dry matter. Energy density decides how much is fed, so add g per 1000 kcal ME ⁴.

07The energy step carries its own error

An energy estimate from label values is itself a prediction. In 451 Brazilian foods compared with in vivo values, the modified Atwater factors used in Table 1 came in low: by 8.6% on average for dry dog foods and 16.9% for wet dog foods ⁵. The unmodified Atwater factors (4, 9 and 4 kcal/g) were closest in that dataset, but were still 8.25% to 11.99% out for wet foods ⁵. A US study found errors fell when separate equations were fitted for dry and wet foods ⁴.

A low energy estimate inflates every per-1000 kcal figure, and inflates it more for wet food. When two products are close, ask for analysed energy, or at least note which equation was used. Our energy calculation guide compares the methods.

08Questions brands ask

Sources

1. FEDIAF (2025). Nutritional Guidelines for Complete and Complementary Pet Food for Cats and Dogs. Source ↗
2. Hill RC, Dzanis DA (2024). Updates to pet food labels and the effect on nutritional evaluation. Today’s Veterinary Practice, May/June 2024, 34–38. Source ↗
3. AAFCO (2023, 31 July). AAFCO membership approves new model pet food and specialty pet food regulations [news release]. Source ↗
4. Jewell DE, Jackson MI (2023). Predictive equations for dietary energy are improved when independently developed for dry and wet food which could benefit both the pet and the environment. Frontiers in Veterinary Science 10:1104695. Source ↗
5. Marchi PH, Amaral AR, Príncipe LA, et al. (2025). Accuracy of predictive equations for metabolizable energy compared to energy content of foods for dogs and cats estimated by in vivo methods in Brazil. Animals 15(10):1477. Source ↗
6. Brociek RA, Li D, Broughton R, Gardner DS (2025). Nutritional analysis of commercially available, complete plant- and meat-based dry dog foods in the UK. PLOS ONE 20(9):e0328506. Source ↗

Evidence reviewed by the Guires Pet Nutrition Lab scientific team. This article is general information for product developers and is not legal or veterinary advice.