Every Pawpry page must say what kind of evidence it uses, match claims to the exact subject and source, preserve conflicts and unknowns, give readers an observable decision process, and avoid claiming testing or expertise that was not actually recorded.
1. Define the decision before collecting sources
Research begins with a reader task: measure a dog for a harness, distinguish a seat extender from a restraint, maintain a particular coat layer, or audit an exact product. The task defines the facts needed and prevents a page from becoming a generic collection of dog advice.
2. Use a source hierarchy
| Source class | Best use | Built-in limit |
|---|---|---|
| Current official manual or exact-model page | Dimensions, components, instructions, warnings, care, stated tests. | The manufacturer owns the claim; marketing is not independent validation. |
| Independent standard, certification, or research record | Test method, pass criteria, current certification, studied outcomes. | Applies only to the tested model, version, size, method, and date. |
| Veterinary or professional organization guidance | Care boundaries, welfare, handling, general safety, when to seek help. | Usually cannot answer exact-product performance. |
| Breed parent club and recognized standard | Function, physical standard, coat, temperament context, recommended health screens. | Describes a population or breeding ideal, not an individual diagnosis or product size. |
| Retailer, review, forum, or social post | Question discovery and leads to primary documentation. | Not sufficient for hard specifications, safety status, or universal conclusions. |
Sources are checked for availability, subject match, date relevance, and whether the visible page actually supports the proposed statement. A URL in a bibliography does not make an unsupported claim acceptable.
3. Match the exact entity
For products, record brand, full model, generation or version when known, size or SKU where relevant, market, and checked date. Similar names do not prove equivalence. Old test results cannot be assigned to a redesign as though it was tested, and redesign language cannot erase a relevant historical result. Both records should be presented with scope.
For breeds, use the recognized breed name and parent-club source. For mixed-breed or individual advice, do not infer ancestry from appearance. Measurements, observed behavior, health, coat, and task remain primary.
4. Assign an evidence identity
- Hands-on review: requires possession of the exact product, original dated evidence, a stated protocol, sample limitations, and disclosure of any supplied item. Pawpry currently makes no hands-on claim on its launch product records.
- Official-docs or specification audit: reports exact-model documentation, conflicts, missing facts, and decision implications without claiming physical use.
- Compatibility guide: maps published dimensions, variants, and constraints; it does not guarantee fit.
- Source-reviewed guide: synthesizes relevant primary and authoritative guidance into an observable process.
- Editorial page: explains site purpose, policy, or disclosure rather than a product outcome.
5. Separate fact, inference, and rule
A manufacturer-published girth range is a fact attributed to that manufacturer. “This is a candidate size” is a bounded inference. “Verify shoulder freedom and escape resistance” is a Pawpry decision rule supported by fit guidance. Pages should make those layers legible instead of writing all three in the voice of a laboratory conclusion.
6. Preserve conflicts and unknowns
When sources disagree, Pawpry checks whether they refer to different dates, versions, sizes, markets, or methods. If the conflict cannot be resolved, the page records both claims, identifies the owners, and tells the reader what cannot be concluded. Missing load ratings, undefined “crash tested” language, uncertain generation match, or absent dimensions remain unknowns; they are not filled with estimates.
7. Build useful depth
Length is not the goal. A high-quality decision page should normally include:
- A direct answer and clearly defined scope.
- The measurements or observations needed before shopping.
- A decision table, sequence, or worksheet that changes what the reader does.
- Known specifications with exact source ownership.
- Limits, unresolved questions, and people or dogs for whom the approach is unsuitable.
- Safe-use, stop-use, and escalation conditions proportional to risk.
- Relevant internal links without duplicating another page’s job.
- Reviewed or checked dates and accessible source links.
8. Medical, behavior, and safety boundaries
Pawpry can help a reader notice rubbing, altered gait, distress, damaged hardware, ingestion risk, or a documentation gap. It cannot diagnose disease, prescribe treatment, certify collision performance, or replace professional assessment. Pages direct readers to veterinarians, qualified groomers, reward-based behavior professionals, manufacturers, or independent certification records when the decision exceeds editorial guidance.
9. Breed-page rules
Breed standards are used only where they change a real equipment or care check—such as recognized coat variety, broad body proportion, or working movement. Stereotypes are avoided. Weight ranges are attributed, and when a standard does not specify weight the page says so. Health-screening recommendations are not converted into claims that every dog has a disease or that a CHIC number guarantees health.
10. Product-page rules
An exact-model product page contains documented specifications, a quick decision answer, explicit limitations, model-history notes when relevant, related how-to guidance, and sources. Pawpry does not add star scores without a defined scoring protocol and evidence dataset. Prices and stock are omitted unless a process exists to keep them current.
11. Images and diagrams
Visuals must help the decision and accurately depict the subject. Dog images must remain unmistakably canine at article scale, with a rounded muzzle, black nose where anatomically appropriate, dog-like ears, visible paws or legs when the composition requires them, tail, and believable proportions. Measurement visuals must place the tape or tool correctly and must not show unsafe restraint or handling. Generated or illustrative images are not evidence of product testing.
12. AI-assisted editorial work
Automated tools may assist with drafting, code, structured records, link checks, and visual concepts. They do not replace source verification. Public claims must remain traceable to checked sources or be clearly identified as Pawpry editorial guidance. Invented experience, citations, credentials, model facts, or safety outcomes are prohibited. A human site owner remains responsible for publication decisions and corrections.
13. Corrections and update triggers
A material correction should update the underlying record, reviewed or checked date, affected summary, structured data, and related pages. Triggers include a changed size chart, new generation, revised manual, certification change, unavailable source, identified factual error, or safety notice. Cosmetic edits should not falsely refresh a review date.
Correction requests should include the Pawpry URL, exact disputed text, primary source, model or SKU, and date checked. The editor should preserve the original evidence state long enough to understand what changed, then document the resolution in the record.
14. Affiliate separation
Affiliate eligibility, commission rate, retailer relationship, and stock do not determine inclusion or conclusions. A product may be documented without an affiliate link; a linked product can receive an unfavorable or unresolved verdict. The affiliate disclosure explains the commercial relationship in reader-facing terms.