Our methodology
How we analyze food for dogs and cats.
What sets our analysis apart
Most comparison tools assign a single grade per product, from a handful of sub-scores. We go much further:
Every product is examined against 20+ nutritional criteria: protein, fat, carbohydrates, minerals, key ratios, additives, cooking, traceability. A level of detail unmatched by mainstream comparison tools.
A grade per profile, not a single grade
The same food is scored separately for each profile: puppy (small or large breed), adult, active dog, neutered, senior, weight loss, for dogs and cats alike. The needs differ, so does the grade.
A scientific basis: g/Mcal
We don't score the percentages printed on the bag, but the dose actually ingested per day, relative to energy (grams per 1000 kcal), as FEDIAF, NRC and AAFCO do. It's the only fair way to compare two foods of different densities.
Thresholds anchored in the references
Our thresholds are based on official animal-nutrition recommendations (FEDIAF, NRC, AAFCO) and scientific literature, not the market average. An A has to be earned: the scoring genuinely separates the very good from the mediocre.
Redhibitory alerts
A genuinely risky excess (calcium, phosphorus, unbalanced ratio, controversial additive) automatically caps the grade, even when the rest of the composition is good. Safety takes priority over the average.
The criteria we evaluate
Each product is evaluated on a set of nutritional criteria grouped into three families:
Here is the full set of criteria evaluated for a kibble. Each is scored individually, then weighted; the criterion-by-criterion detail appears on every product page.
Nutritional profile
-
Protein density (RPC)
Protein amount relative to the food's energy, not the raw percentage on the label. It shows whether the calories come mainly from protein (the hallmark of a carnivore-appropriate diet) or from carbohydrates and fat. It is one of the most decisive criteria in the score.
-
Fat content
Fat level, the main energy source. The target depends on the profile: higher for a puppy or an active dog, lower for a neutered or weight-loss animal. Both excess and deficiency are penalized.
-
Carbohydrates
Carbohydrate share (starch, sugars), estimated by difference. Of little use to dogs and cats, it mostly binds the kibble; an excessive share is penalized, especially for cats (metabolic and diabetes risk).
-
Ash (total minerals)
Total mineral content (the residue after combustion). A high level can reveal a large share of bone, carcass or by-products, and therefore lower-quality protein.
-
Fiber
Dietary fiber: in the right amount it supports transit, satiety and a balanced gut flora. Too little, or too much, hinders digestion and nutrient absorption.
-
Calcium / phosphorus balance
Ratio between calcium and phosphorus. Even with correct individual levels, an imbalance disrupts bone mineralization; the references (FEDIAF, NRC) target roughly 1.1 to 1.5 in adults, with tighter bounds during growth.
-
Calcium level
Calcium actually supplied (relative to energy, see below). Scored on a "U" curve: deficiency hampers bone growth, excess is risky, particularly for large-breed puppies.
-
Phosphorus level
Phosphorus supplied. An excess strains the kidneys over the long term; this criterion weighs more for senior profiles and animals at renal risk.
-
Omega-6 / omega-3 balance
Ratio between omega-6 and omega-3. A good balance supports skin, coat and inflammation control; an excess of omega-6 relative to omega-3 is unfavorable.
-
Protein / ash ratio
Protein relative to ash (minerals). A high ratio indicates nobler protein and fewer mineral residues: a good indicator of the real quality of protein sources, beyond the protein level alone.
-
Fat / protein ratio
Balance between fat and protein. It helps tell a genuinely protein-rich food apart from one whose calories come mostly from fat.
-
Protein / phosphorus ratio (adjusted)
Protein-to-phosphorus ratio corrected by the share of animal protein. Plant proteins supply little phosphorus and artificially inflate the raw ratio: the adjustment neutralizes this bias, so that only mainly animal, noble protein earns a good score. It cross-checks and strengthens the protein/ash criterion rather than duplicating it.
Ingredient quality
-
Animal protein content
Share of animal-based protein versus plant-based protein. Dogs, and especially cats (obligate carnivores), make better use of animal protein, whose amino-acid profile is complete; the score rises with this share.
-
High glycemic index ingredients
Presence of high-glycemic-index ingredients (certain starches such as potato or tapioca in large proportion), which cause undesirable blood-sugar spikes, particularly in cats.
-
Gluten-free
Presence or absence of gluten (wheat, barley…). Not dangerous for most animals, but often used as a cheap plant protein and a problem for sensitive individuals.
-
Fresh meat / fish
Presence of fresh meat or fish in the recipe (not only dehydrated meals). Fresh material is a mark of quality, even if its real weight drops after cooking.
-
Taurine
Taurine, an amino acid essential for cats (heart function, vision, reproduction). Its addition or presence is rewarded; its absence in cat food is a genuine weakness.
-
Controversial additives
Controversial additives: chemical preservatives (BHA, BHT, ethoxyquin), artificial colorings and flavorings. Their presence lowers the score and can, in the most sensitive cases, cap it.
-
Legumes in ingredient list
Legumes (peas, lentils, chickpeas). When percentages are declared, we score the real load (cumulative dose); otherwise we combine the position in the ingredient list and the number of distinct sources. This count neutralizes "splitting": breaking peas into several entries (peas, pea protein, pea fiber…) so each looks minor no longer hides the total dose. The goal: to catch the dilated cardiomyopathy risk signal (FDA alert, 2018) on legume-heavy diets.
Transparency and manufacturing
-
Cooking method
Declared cooking method (high-temperature extrusion, low temperature, freeze-drying). Gentler processes preserve nutrients better; the absence of information is itself a transparency signal.
-
Fat traceability
Fat source: named (e.g. "chicken fat") or generic ("animal fats"). A named source is a mark of traceability and quality.
-
Calcium / phosphorus disclosure
Does the manufacturer disclose calcium and phosphorus levels? Without these values the mineral balance cannot be verified: the score is then corrected and capped (see redhibitory alerts).
Wet foods are evaluated on an equivalent set, adapted to their high moisture content (moisture, energy-based protein density, etc.). The full breakdown appears on every product page.
A grade tailored to each profile
A key point: the score is not a single number. It is calculated separately for each profile (adult, senior, puppy/kitten, neutered, active, weight loss) because nutritional needs vary widely with age, size and lifestyle. Reference thresholds are tuned for each one.
In practice, the same kibble can earn an A for a growing kitten and a C for a neutered indoor cat, because their needs have nothing in common. A universal grade cannot capture that nuance; ours can.
Label percentage vs scored value: an energy-based approach
Since September 2026, we no longer score nutrient percentages (protein, calcium, phosphorus, ash) as printed on the packaging. We first convert them to grams per 1000 kcal of metabolizable energy (g/Mcal), the scientific basis used by FEDIAF (Table III-4b), AAFCO and NRC.
Why the change?
A pet needs a fixed amount of energy per day, regardless of how calorie-dense its food is. Two kibbles showing the same calcium percentage can deliver very different real doses depending on their caloric density. The g/Mcal basis corrects this distortion by measuring what's actually ingested per day, not the label percentage.
A concrete example
A kitten covers a fixed energy need every day. With a low-calorie food it has to eat more grams to reach that need than with a very calorie-dense one. The result: at an identical calcium percentage on the bag (say 1.1%), the lower-calorie food actually delivers more calcium per day. It is this truly ingested dose, relative to energy (grams per 1000 kcal), that we evaluate, not the label percentage.
What you see on product pages: The displayed rate remains the raw percentage declared by the manufacturer (as printed on the package). Only the internal scoring uses the g/Mcal basis, not the display. This keeps the number consistent with what you can verify on the packaging.
Criteria concerned: Protein-to-Calorie Ratio (RPC), Calcium, Phosphorus and Ash, for both dry and wet food, dog and cat.
From the overall score to a letter
The total score (out of 100) is converted into a letter grade:
Redhibitory alerts
Certain criteria can trigger a redhibitory alert, regardless of the overall score. These are situations where a value exceeds a threshold we consider risky or unsuited to the profile. Those thresholds most often rest on the recommendations of reference organizations (FEDIAF, NRC, AAFCO) or on official alerts (such as the FDA's on legumes), but where the literature leaves room, they also reflect a deliberate editorial decision on our part.
The main alerts we watch for:
Nutritional imbalances
- Protein density (RPC) too low for the intended profile.
- Insufficient fat relative to the profile's needs.
- Calcium too high, or conversely below the recommended minimum.
- Phosphorus too high (renal caution), or insufficient.
- Unbalanced calcium/phosphorus ratio.
- Carbohydrate level above recommendations, especially in cats.
- Excessive ash: high mineral load, often a sign of bone or by-products.
- Excessive fiber, which dilutes nutritional density and digestibility.
Ingredients and transparency
- Presence of controversial synthetic preservatives (BHA, BHT, ethoxyquin).
- Dominant legumes (1st or 2nd position), split across several sources, or dominant without added taurine in cats (FDA alert, 2018).
- A recipe that discloses neither calcium nor phosphorus: with the mineral balance unverifiable, the grade is capped.
These alerts are clearly displayed on the product page. They do not change the criterion-by-criterion analysis, but cap the displayed score: while an alert is active, the product cannot exceed a certain grade, regardless of the quality of the rest of the composition.
Reading a score properly
A score is not a universal rating. Three key principles to use it well:
Compare within a category, not between categories
Scoring thresholds differ by food type (kibble, wet, raw, home-cooked, freeze-dried, treats). Industrial process, digestibility, water content and physiological needs vary dramatically from one category to another. A B on kibble and a B on wet food do not mean the same thing and cannot be directly compared.
A lower score may suit your pet better
A product graded C or B can sometimes be more appropriate than one graded A, depending on your pet's profile: age, weight, activity, spay/neuter status, condition (renal insufficiency, digestive sensitivity, food allergy…). The score reflects a generic quality for a given physiological profile - but your vet knows criteria no algorithm can anticipate.
A score never replaces veterinary advice
PetFoodStats is a decision-support tool, not a diagnosis or prescription. In case of doubt, a diagnosed condition or a significant diet change, consult your vet - they are the only professional qualified to recommend a diet based on the complete health assessment of your pet.
Where does the data come from?
Nutritional data (protein, fat, fiber, ash, moisture, etc.) is collected from information published by manufacturers on their packaging, official websites, and technical sheets.
PetFoodStats does not perform laboratory analyses. We work exclusively from declared data. This means our scores reflect what the manufacturer communicates, not an independent measurement of the actual composition.
When a data point is not disclosed by the manufacturer (for example moisture or metabolizable energy), depending on the criterion: we either apply a standardized estimate, or the criterion is removed from the calculation (neither bonus nor penalty) and the score is computed pro rata over the criteria actually available. These cases are flagged "n/c" on the product page.
What do we base our work on?
The thresholds and criteria used in our algorithm are informed by recommendations from leading animal nutrition organizations:
- FEDIAF (European Pet Food Industry Federation) - Nutritional Guidelines, current edition.
- NRC (National Research Council) - Nutrient Requirements of Dogs and Cats, the reference publication on nutritional needs.
- AAFCO (Association of American Feed Control Officials) - Dog and Cat Food Nutrient Profiles.
These organizations provide ranges that are sometimes broad, and do not cover every criterion we evaluate. When official recommendations do not allow for a precise threshold, we make an editorial judgement based on available literature and our experience. These choices are acknowledged and may evolve over time.
Official recommendations evolve with the state of knowledge. Our algorithm is regularly updated to reflect these changes and refine our own editorial choices.
What our score does not measure
In full transparency, here is what our scoring system cannot evaluate:
- Actual nutrient digestibility - two foods with the same protein level can be absorbed very differently by the body.
- Mineral and vitamin bioavailability - the chemical form of a nutrient influences its absorption.
- Organoleptic quality - palatability, texture and taste matter to the animal but cannot be measured from a label.
- Individual needs - an animal with a medical condition (kidney, cardiac, allergy) may require a specific food that a general score cannot reflect.
- Actual raw material quality - the label states the category ("chicken"), but not the quality of the batch used.
- Storage conditions and product freshness at the time of purchase.
Independence and commitments
PetFoodStats is a nutritional analysis tool for dog and cat food. No brand can pay to influence its score or ranking. If affiliate links to retail sites are present, they have no impact on analysis results.
Our goal is to provide a clear, factual framework to help pet owners compare the nutritional compositions of products available on the market.
The score we assign reflects an assessment of the composition declared by the manufacturer. It is not a value judgement on the brand, and does not replace veterinary advice.
If you notice an error in a product's data, please let us know. We correct confirmed errors as quickly as possible.
Our algorithm evolves. When new scientific data or official recommendations are published, we adjust our thresholds and criteria accordingly. A score may therefore change over time without the product's composition having changed.
Important notice
The information presented on PetFoodStats is provided for informational purposes only. It does not constitute veterinary or professional nutritional advice. For any medical condition or specific need, consult your veterinarian. The publisher disclaims all liability for dietary choices made based on this information.
Read the full disclaimer