Our methodology

How the Food score works

Built on published guidelines (FEDIAF, WSAVA). This is not veterinary advice — always speak to your vet about your own dog.

How we stay neutral. The Food score is identical for every visitor. Affiliate partnerships and priority placement never change it, and price is never an input — it is shown separately. The quiz's Match score may read the Food score, but nothing ever writes to it. We earn commission on some links; that never changes a score. Read our affiliate disclosure.

1. Principles

  • Neutral. The score is the same for every visitor. Boosts and price never change it.
  • Transparent. Every point comes from a published rule on this page. Rules marked [house] are ChewsyDog judgements; the rest come from FEDIAF or WSAVA.
  • No guessing. Data we have not collected yet is excluded and the food shows a "Data incomplete" badge. Where a maker was asked and did not disclose, the item scores 0 — following WSAVA's advice to be cautious when a maker cannot or will not provide information.
  • Not scored: prescription and veterinary diets (shown with "use under veterinary guidance") and products under an active recall (shown with a recall banner).
  • Grain-free is neutral. Grain-free foods score neither up nor down — the FDA's DCM investigation found the data insufficient to establish causality. Hypoallergenic claims are not scored either.

2. Weights, bands and completeness

ComponentWeight
Nutrition35%
Ingredients30%
Processing15%
Maker standards & safety20%

A component only counts towards the total when we have at least half of its data. The total is a weighted average of the counted components, rescaled to what's present — and Nutrition must count, or there is no score at all.

Bands: 85–100 Excellent · 70–84 Very good · 55–69 Good · 40–54 Fair · 0–39 Poor.

Completeness: below 85% data completeness a food shows a "Data incomplete" badge; below 50% its score is labelled Provisional. Until we have verified a food against the manufacturer's own label or website, its score is shown as Provisional with no band. Complementary (mixer/topper) foods are capped at 50.

3. Nutrition (35%)

All nutrient checks use dry matter so wet and dry foods are compared fairly: nutrient DM = as-fed % × 100 ÷ (100 − moisture %). Carbohydrate (NFE) is estimated as what's left after protein, fat, fibre and ash.

When moisture is missing on a dry-style food we assume 10% (UK/EU rules only require moisture on the label above 14%) and flag it. When it's missing on wet, raw or fresh food, those items are excluded.

  • Complete vs complementary (25 pts): complete foods score full points; complementary foods score 0 and are capped at 50 overall.
  • Life-stage fit (15 pts): protein and fat DM must meet the FEDIAF 2025 minimum for the food's life-stage group (adult 18/5.5, growth 20/8.5, all life stages 25/8.5 g per 100 g DM). Large and giant puppy foods above the FEDIAF late-growth calcium maximum also fail.
  • Protein DM (25 pts): 0 below the minimum, rising linearly to full marks at 32 (adult) or 35 (growth) [house curve].
  • Fat DM (10 pts): full marks between 10 and 22, tapering outside that range [house].
  • Fibre (5), ash (5) and NFE carbohydrate (15): moderate fibre and ash score best; lower estimated carbohydrate scores higher because dogs have no established minimum carbohydrate requirement (NRC 2006) — this rewards energy from protein and fat, and is not a safety claim [house].

Implausible label values (for example fibre above 30% as fed, or protein below 10% DM) are excluded and flagged as data errors rather than scored.

4. Ingredients (30%)

We parse the manufacturer's own ingredient list and technological additives:

  • First ingredient (35 pts): a named species ("chicken", "dried duck") scores highest; an animal source with no species ("poultry", "fish") scores less; category terms like "meat and animal derivatives" score lowest.
  • Declared animal content (25 pts): the sum of declared percentages on animal ingredients. The 4/14/26% cut-offs follow FEDIAF naming rules ("with" ≥ 4%, "rich in" ≥ 14%, "dinner" ≥ 26%).
  • Open labelling (20 pts): catch-all category terms (cereals, oils and fats, various sugars, meat meal and similar) reduce the score; none scores full marks.
  • Added sugars and bulk fillers (10 pts): added sugar, caramel, molasses, syrup or honey score 0; bulk by-products score part marks. Sugar beet pulp counts as fibre, not sugar.
  • Artificial additives (10 pts): artificial colours, synthetic antioxidants (BHA, BHT, propyl gallate, ethoxyquin), unspecified "EC permitted" additives and artificial flavours reduce the score [house]. These additives are authorised in the EU — this is a preference for transparency and natural preservation.

Dried and fresh meat percentages are not equivalent; we use the declared % as printed and say so on the product page. Vegetarian and vegan complete foods score low on the first two items by design — their nutrition is still judged on its own merits.

5. Processing (15%)

Deliberately a small part of the score — evidence that one cooking method is nutritionally superior is limited [house]. Gently cooked/fresh (85), wet, baked, cold-pressed and air-dried (80), extruded kibble (75) and semi-moist (65) all score well.

Raw (50) and freeze-dried raw (55) score lowest because WSAVA, the FDA and BSAVA all report pathogen risk to dogs and households from raw meat diets; documented HPP treatment adds 15 points. Where the method is unknown on a dry food, we score it as extruded and flag it.

6. Maker standards & safety (20%)

Brand-level questions following the WSAVA "selecting a pet food" checklist. Every answer needs an evidence link and a verification date:

  • Qualified nutritionist (30 pts): a PhD animal nutritionist or ACVN/ECVCN board-certified nutritionist on staff scores full marks; a consultant scores part marks.
  • How adequacy is verified (25 pts): feeding trials score highest, then finished-product lab analysis, then formulation-only.
  • Quality control (25 pts): a certified site (for example BRCGS or ISO 22000) plus documented ingredient and finished-product testing scores full marks.
  • Recall history, last 5 years (20 pts): no recalls scores full marks; safety recalls reduce it. Sources: FSA alerts, FDA recalls and the maker's own notices.

7. Data provenance and verification

The initial dataset was compiled from public listings. Before a score is published as final, each product is verified against the manufacturer's own label or website, and the product page cites the manufacturer only. Until then, scores are shown as Provisional.

Spotted something wrong on a product page? Use the "Report an error" form on that page and we'll check it against the manufacturer's label.

8. Sources

9. Changelog

Every rule change bumps the score version, re-scores every product, and is listed here.

  • v1.0.0 — initial published ruleset.