These standards describe how we are improving DeskFitPro's buying guides. They are not a certification that every legacy article has completed a full fact-check.
Sources must support the actual claim
A citation must be real, accessible, and relevant to the statement next to it. We check the study population, intervention, outcome, and limitations before drawing a conclusion. Evidence for an ingredient does not establish a benefit for every finished product containing it. A mechanism or customer testimonial does not establish a clinical outcome.
When a claim cannot be supported, it should be removed or narrowed. Phrases such as “research suggests” do not make an unsupported claim acceptable. Certifier and regulatory claims require evidence for the exact product and scope; a brand's reputation is insufficient.
Explain how a recommendation was reached
A useful comparison names the buying decision, meaningful differences, tradeoffs, and reasons to choose another option or skip a purchase. Rankings should have a stated basis. Listing order and star ratings are not proof that a product is best.
Research-based comparisons must be described as such. First-hand testing claims require a record of the product, method, observations, and limitations. We should not invent sample sizes, review counts, test durations, or personal-use stories. Our own testing and another publisher's testing must be clearly distinguished.
AI assistance and checks
AI tools assist with research, drafting, editing, and site maintenance. Automated checks look for selected content and technical defects; browser checks verify presentation. Neither constitutes a complete fact-check, clinical review, or guarantee of human review for each page. Passing those checks does not establish compliance with Google's policies.
Amazon catalog checks and prices
We use Amazon's Creators API to check listing identities and available catalog facts. A returned ASIN is not evidence of medical effectiveness, certification, authentic customer feedback, or guaranteed stock at checkout. Titles, selected variants, seller terms, and product labels can change.
The scheduled refresh attempts price checks every six hours. A displayed API price includes its successful check time and expires after 24 hours without a fresh check. Older figures are hidden. Ratings and review counts are not supplied by this integration and must not be invented or treated as current API data.
An unavailable listing may remain for comparison with a dated note. It is not automatically replaced with another ASIN. Correcting an identity means checking that the label, link, image, and explanation refer to the same product.
Health claims require particular care
A guide should separate nutritional requirements from claims that supplementation improves a symptom. It should distinguish upper intake limits from recommended doses, describe relevant interaction risks, and avoid diagnosing readers. Where evidence does not justify a purchase, the recommendation should say so.
Affiliate relationships
We earn from qualifying Amazon purchases. No manufacturer has paid for placement in our guides. Commercial links are marked as sponsored or nofollow, and affiliate disclosures appear on review pages. Readers should be able to reach a no-purchase decision from the information we provide.
Corrections and updates
When we identify a substantive error, we correct the affected text, product cards, comparison tables, and metadata, and add a dated explanation. A price refresh is not a new clinical evidence review. A last-updated date does not mean that every source or product was newly tested.
Our remaining catalog review is ongoing. Send a page URL and supporting source through our contact page if you find something that needs correction.
Google guidance
We use Google's spam policies, review guidance, and helpful-content guidance as reference points for useful, trustworthy publishing. A traffic decline near an update does not identify a specific violation, and fixing our own checks does not guarantee ranking recovery.
Updated September 8, 2026: removed unsupported universal verification promises and claims about the specific targets of Google's June update.