Editorial standards & research methodology

We prioritize decisions with meaningful consequences for safety, recurring cost, reliability, privacy, or the daily care routine. Reader questions, product changes, and gaps in existing coverage can all prompt new research.

A comparison should pair plausible alternatives for the same buyer or clearly explain why a premium-versus-value or same-brand decision is useful. We do not create mismatched comparisons simply because two product names attract searches.

A shortlist may consider availability, category relevance, model generation, and evidence quality. Popularity can help identify what readers encounter, but it is not evidence of quality.

Primary sources—manuals, support documents, official specifications, regulatory filings, subscription terms, and current retailer listings—form the baseline for factual claims. We record the relevant model, region, source, and check date when those details affect interpretation.

Manufacturer statements are attributed as claims. Independent measurements, structured user-feedback analysis, and our own observations are labeled separately so readers can tell what each conclusion rests on.

When sources conflict, we show the disagreement or withhold a conclusion. Missing information is not silently converted into a negative score, and a retailer rating is never treated as proof of a core function.

We only say “tested” when PetMetric directly evaluated the product. The page should explain the unit or variant, evaluation period, conditions, procedures, and important limitations. Research based only on documentation or external evidence is described as research, not testing.

A manufacturer-provided or loaned sample is disclosed. Providing a sample does not guarantee publication, a positive result, or advance approval. Manufacturers may check factual details, but they do not approve our verdict.

Where repeatability matters—such as feeder portions, battery behavior, notification timing, or noise—we favor repeated observations over a single successful attempt and avoid implying laboratory precision we did not achieve.

Scores summarize evidence; they do not replace it. Criteria are defined for the category and should connect to a real owner decision. The underlying criterion results and limitations matter more than a decimal-point difference in the total.

A product can be strong overall and still be wrong for a specific pet or home. Best Picks are organized around use cases, while comparisons explain where each alternative has an edge and whether the evidence is sufficient to name one.

Price, subscription terms, regional coverage, firmware, and availability change. We avoid permanent “best” language when the conclusion depends on volatile facts, and we tell readers what to verify before buying.

Pages carry a checked or updated date appropriate to their content. We prioritize review when a product is recalled or discontinued, a new generation launches, a major firmware release changes behavior, subscription terms move, or a substantiated reader correction identifies a material issue.

A date means the page was reviewed to the scope described; it is not a guarantee that every retailer field remains current. Price, stock, seller, shipping, and regional compatibility should always be confirmed at the destination.

When a material change affects a verdict, score, safety statement, or recommendation, we update the page and describe the change rather than quietly preserving an outdated conclusion.

We welcome corrections that include the page URL, the disputed statement, and a source or reproducible observation. We aim to acknowledge submissions promptly and verify substantive corrections within five business days, though complex or manufacturer-dependent questions may take longer.

Confirmed factual errors are corrected. Material changes receive a visible note or changelog entry with what changed and when. Minor spelling or formatting fixes may be made without a separate notice when they do not alter meaning.

If we cannot verify a proposed correction, we do not present it as fact. We may ask for more detail, preserve the existing wording, or revise the page to state the uncertainty more clearly.