Convert the claim into a sentence that names the subject, property, comparison, market, time, and level of certainty. Replace this brand is durable with a question about a named model, failure type, use condition, and period. Replace the brand is American with separate questions about headquarters, ownership, design, manufacture, and model-level origin. This prevents one credible document from being stretched across claims it never addressed.

• Subject

  Name the legal entity, brand, product line, model, seller, policy, or claim owner.

• Property

  Define what is being asserted: specification, performance, origin, ownership, authorization, warranty, review experience, or value.

• Scope

  State market, model, seller, user group, conditions, comparison set, and time period.

• Evidence threshold

  Decide what would confirm, contradict, or leave the claim unknown before reading persuasive material.

• Decision relevance

  Explain how the finding would change the shortlist, seller route, price tolerance, or need for more research.

The first source-quality question is not whether the source looks authoritative. It is whether the source was capable of answering the claim you actually wrote down.

The FTC’s advertising guidance states that advertising claims should be truthful, not deceptive or unfair, and evidence-based. Its substantiation policy focuses on whether the advertiser possesses the amount and type of support the claim communicates. For a shopper or editor, that means the evidence must fit both the explicit words and the reasonable implication of the claim, not merely mention a related ingredient, technology, or product category.^[1][2]

1. Confirm the exact source

   Open the underlying page, document, filing, dataset, or policy rather than relying on a search snippet, citation chain, or generated summary.

2. Identify who produced it and why

   Record the publisher, author or responsible entity, commercial relationship, date, and whether the source is describing itself, measuring something, or reporting an experience.

3. Read the method and sample

   For tests and studies, check the exact model, comparator, sample, conditions, endpoints, exclusions, and whether the method can support the advertised conclusion.

4. Match geography and channel

   Policies, models, formulas, voltage, sellers, warranty, and legal claims can differ by country, retailer, or marketplace.

5. Check recency and change risk

   Record publication and check dates, then schedule review based on how quickly ownership, policy, model, price, support, or guidance can change.

6. Separate observation from inference

   Write what the source states or measures, then label the narrower editorial conclusion and any remaining alternative explanations.

1. State the direct answer

   Answer the user’s actual question in a bounded sentence before presenting background or adjacent facts.

2. Attach only supporting sources

   Cite the material that supports the direct answer or visible body and explain what each source contributes and does not prove.

3. Label conflicts and unknowns

   Show where sources disagree, where a relationship is only inferred, and which product-, seller-, or market-specific fact remains unresolved.

4. Separate commercial relationships

   Do not let sponsorship, affiliate economics, reciprocal links, brand representatives, or access to samples change the evidence standard or disappear from disclosure.

5. Set the next review

   Record the checked date, likely change triggers, owner, and next review date so the page does not present a once-true fact as timeless.

For public companies, SEC EDGAR can provide primary filings about operations and corporate events. For U.S. trademarks, USPTO search can help identify records and assignment information. Neither database alone proves who manufactures a product, operates a website, controls every licensed use, or provides a warranty, so combine the record with the exact relationship being researched.^[3][4]