Methodology
How PetScored scores pet supplements.
A public, deterministic rubric. Two reviewers with the same label, the same evidence library, and the same source records compute the same score. AI may parse labels, but it does not decide scores. The composite is five sub-scores on one calibrated 0-10 scale, so a 7.8 means the same thing whether it is a dog food, a cat food, or a supplement.
The composite
Five sub-scores, weighted.
Composite = Evidence Quality 30% + Active Dose 25% + Ingredient Clarity 15% + Safety & Manufacturing 15% + Value 15%. Bands: 8.5-10 Strong, 7.5-8.4 Good, 6.5-7.4 Mixed, below 6.5 Limited.
Evidence Quality - 30%
Does the primary active have credible evidence for its claimed benefit, ideally in the target species? We tier the active the headline claim rests on against a cited evidence library of canine and feline trials and meta-analyses, then adjust for whether a combination product headlines a genuinely well-evidenced active or a halo claim sits on a trace ingredient. This is the supplement aisle's biggest weakness and our biggest lever: a best-selling glucosamine chew can be beautifully dosed and still score mid because the 2022 meta-analysis found a marked non-effect for glucosamine and chondroitin in osteoarthritis.
Active Dose - 25%
Does a labeled daily serving deliver the dose the evidence used? We evaluate the delivered dose at a 10 kg reference dog or 4 kg reference cat, crediting per-active disclosure, meeting (or reaching 50-99% of) the studied dose, weight-scaled feeding guidance, and the bioavailable studied form. The pill is not the dose. A proprietary blend that hides per-active amounts cannot be credited for a dose it will not show, so Active Dose is capped at 4.0.
Ingredient Clarity - 15%
Can a buyer see exactly what they are getting, in a form they can evaluate? We credit every active named with its specific form, no amount-masking proprietary blends, disclosed inactive ingredients, and stated sourcing or standardization. A proprietary blend that masks per-active amounts is penalized, so a dose-hiding label scores below a transparent one.
Safety & Manufacturing - 15%
The pet-supplement analog of brand trust plus purity. Supplements are lightly pre-market regulated, so third-party signals carry weight: the NASC Quality Seal, independent third-party testing (which supersedes in-house batch testing), named cGMP manufacturing, and FDA/CVM recall and warning-letter history. An ingredient unsafe for the target species floors the composite.
Value - 15%
Cost per day at the reference serving, never cost per pill or per bottle, benchmarked against the median for that supplement type (joint, gut, skin and coat, calming, multivitamin, liver, immune, longevity). When a blend hides its doses we still band the cost per day, with a visible caveat that the day's serving may not be an effective dose.
Hard rules
Floors, caps, and refusals.
Refuse to score: explicit disease treatment, cure, or prevention claims; prescription products; and unsettled-status actives such as CBD. The engine returns a refusal, not a low score.
Floor the composite: an ingredient unsafe for the target species (for example xylitol in a dog product, or an allium-derived ingredient) floors the composite and sets a species-unsafe flag, regardless of the other dimensions.
Cap a dimension: a proprietary blend that hides per-active amounts caps Active Dose at 4.0 and flags the product, because we cannot verify a dose the label will not show.
Hold, do not guess:if a product's headline active is not in our cited evidence library, we add a sourced entry (pinned to a primary PMID or DOI) before we score a product that leads with it. We do not assign a tier we cannot source.
The evidence backbone
Every tier is cited.
Evidence Quality reads from a versioned library that keys each common active to a claim and a species, assigns an evidence tier, and pins it to at least one primary source (a PubMed PMID or a DOI). The single most useful thing this library does is let the rubric say what the marketing will not: in companion-animal osteoarthritis, the strongest evidence is for omega-3, while glucosamine and chondroitin show a marked non-effect. That separation of execution from evidence is the whole point.
The library covers omega-3 (joint and skin), glucosamine and chondroitin, undenatured type II collagen, green-lipped mussel, probiotic strains (E. faecium SF68, B. animalis AHC7, B. longum BL999 for anxiety), alpha-casozepine, L-theanine, L-tryptophan, SAMe and silybin, and general multivitamins, with CBD held as not-yet-scored. It is re-reviewed on the rubric changelog cadence.
Common questions
What the score does and does not mean.
Why does a popular, well-made joint chew only score Mixed?
Because evidence is weighted heavily and dose transparency matters. Glucosamine and chondroitin for osteoarthritis are contested - a 2022 systematic review and meta-analysis found a marked non-effect - so a chew built on them can be fully disclosed, tested, and weight-scaled and still land in the middle. Omega-3, which has clear meta-analytic support, scores higher on the same five rules.
Why do proprietary blends score so low?
A proprietary blend hides the per-active amounts, so we cannot verify any active reaches a studied dose. Active Dose is capped at 4.0, Ingredient Clarity takes a penalty, and Value is flagged approximate. We refuse to be impressed by a blend that hides the only number that matters.
Do your scores tell me a supplement will help my pet?
No. We score the population-level evidence for an active doing what is claimed and whether the label delivers the studied dose - not whether a product will help an individual animal. PetScored does not give veterinary advice; dosing and use are decisions for you and your veterinarian.
What do you refuse to score?
Products sold with explicit disease treatment, cure, or prevention claims (an unapproved drug claim), prescription products dispensed under veterinary supervision, and actives whose legal status or species safety is unsettled, such as CBD. We also hold a product out until its headline active is in our cited evidence library.
Browse
See the rubric in action.
Browse all scored supplements, or jump to the best for what you are treating. Scores are our opinion under this published rubric, not veterinary advice. We score the evidence and the disclosed dose, not whether a product will help an individual pet.