Research is not hands-on testing

We identify the exact model, read manufacturer documentation, look for independent testing and examine owner reports. If we have not physically tested a product, we say so. Another reviewer's observations are attributed to that reviewer, never presented as our own experience.

Different kinds of evidence stay separate

Manufacturer specifications tell us what is claimed or supported. Independent reviews can describe observed performance. Owner reports can raise patterns and questions, but an anecdote alone cannot establish a failure rate. Native PetDecider reader reviews will remain separate from external reports and editorial scores.

Four questions guide a review

Function fit
Does the product address the defined need, and what evidence supports that conclusion?
Daily upkeep
What needs cleaning, refilling, replacing or checking during normal use?
Controls & access
What identification, power, connectivity and compatibility dependencies affect use?
Ownership cost
What verified recurring supplies, accessories or optional services affect cost over time?

Scores are editorial judgments, not scraped averages

Numerical ratings require a documented category-specific rubric, supporting evidence and an explanation of uncertainty. Missing evidence is shown as “Not rated,” not as zero. No score is published in the current research profiles.

A ranking must define the comparable product class and which researched models it includes. A manually filled access bowl is not assigned an overall winner against a scheduled dispenser without explaining the different jobs they perform.

Owner evidence needs a traceable sample

Before describing common themes, we record where reports came from, which period they cover, how many unique records were reviewed and what limits the sample has. We do not infer a verified purchase or representative failure rate from an unverified public post. Selected anecdotes are labeled as selected reports.

Commercial disclosure

The current research pages use ordinary manufacturer links. No affiliate tracking, paid placement or live inventory feed is connected. If this changes, relevant links and sponsorships will be disclosed. Commissions will not determine an editorial score or ranking.

Images and publication status

Our homepage uses a labeled AI-created editorial scene for atmosphere. Product profiles use clearly labeled concept diagrams while exact model photographs are pending. Neither is evidence of hands-on testing. Approved, accurate product imagery is required before a profile can be marked ready for publication. Drafts and incomplete research remain excluded from indexing and the public sitemap.

Corrections and updates

Material changes should record what changed, the source and the editorial update date. A price refresh alone must not imply that the entire review was re-researched. A public contact channel and correction-submission workflow have not been connected in this development version.

Editorial policy prepared October 8, 2026. This is the working standard for the site's first research profiles.