# How to evaluate product claims before you repeat them Canonical: https://namedproof.com/knowledge/en/evaluate-product-claims By NamedProof (Organization) — https://namedproof.com/about Published: 2026-10-17 · Sources checked: 2026-10-17 · Language: en Expert assessments Turn a product claim into a testable question, match evidence to the exact wording, check comparisons and calculate percentages without overstating the result\. Evaluate a product claim by stating exactly what it promises, for whom and under which conditions\. Match the evidence to that wording, inspect the comparison and calculation, and separate observed results from wider inference\. If the evidence covers only a version, sample or workload, keep that limit in the claim you repeat\. ## Capture the promise and the decision Save the exact wording, source and date before rewriting a claim into something easier to defend\. Faster, safer and more accurate each need a comparison and a measure\. A feature description needs different evidence from a claim that the feature improves a customer outcome\. Write the decision you would make if the claim were true\. A purchasing team may need a tool to complete a task within a time limit, while a marketing team may want to reuse a superiority claim publicly\. The cost of a wrong answer changes how much investigation is justified; it does not change what the available evidence actually shows\. In US advertising policy, the FTC’s substantiation statement discusses having a reasonable basis for objective claims before dissemination and matching support to an advertised level of proof\. This is a US regulatory context, not a universal legal checklist\. The practical guide below evaluates evidence and does not decide legal compliance\. (F04) ## Match the method to the claim A practical claim\-to\-evidence comparison\. These are starting questions, not a complete test protocol\. | Claim | Evidence to request | Common scope gap | | --- | --- | --- | | A feature exists | Current documentation and, where needed, a reproducible check | A higher\-priced plan is presented as available on every plan | | The product is faster | Defined workload, comparator, settings and repeated timing results | A selected task becomes a general speed claim | | Users achieve an outcome | Study design, participants, baseline, period and exclusions | A selected testimonial becomes a typical result | | The product is best | Named alternatives, consistent criteria and justified weighting | One favorable measure becomes an overall recommendation | | A risk is absent | Appropriate investigation with explicit coverage and limits | Failure to observe a problem becomes proof that none exists | ## A hypothetical speed claim with two different percentages Suppose a tool completes a fixed task in 8 minutes and a comparator takes 10 minutes under the same hypothetical setup\. The task takes 2 minutes less, or 20% less time: \(10 − 8\) ÷ 10 × 100\. Throughput for repeated identical tasks would be 25% higher if these times held: 10 ÷ 8 − 1\. These are teaching values, not results from actual products\. For the hypothetical 8\-minute versus 10\-minute task, 20% less time and 25% higher throughput answer different questions\. Saying 25% less time would be wrong for these values, even though 25% higher throughput follows from the inverse relationship under the stated assumption\. Keep the metric and baseline attached to the number\. A single run also leaves repeatability unresolved\. Background processes, task choice and setup could affect the observation\. Before using the claim, inspect how the test controls those factors and whether the result is meaningful for your workload\. An exact calculation can still sit on weak inputs\. ## Trace the evidence chain Move from the claim to the underlying record\. If the supporting report quotes a vendor’s chart, ask for the chart’s method and data\. If an expert interpreted the result, distinguish the measurement from the expert’s explanation\. An expert name does not create observations that were never collected\. Inspect evidence that could challenge the claim\. Ask what failed, what was excluded and whether the same selection rules applied to the comparator\. A favorable average can conceal an important failure case, and a large sample does not repair a comparison built around unequal conditions\. Google’s review\-publishing guidance encourages useful measurements, comparisons and first\-hand supporting evidence\. Those are useful questions for a reader as well as a publisher\. Following the guidance does not establish that a particular test is valid or that its review will rank\. (F02) ## Choose wording proportionate to the finding Use a conclusion that names what the evidence establishes\. For example, a document check may confirm that a feature is listed; a controlled test may show performance under a specific workload\. If the intended claim is broader, record that the current evidence does not establish the broader wording\. Missing support is not automatically proof that a claim is false\. A result can be contradicted, inadequately supported, untested or outside the available scope\. Distinguishing these states prevents an evidence review from turning a gap into an accusation\. A hypothetical final note could say that the 8\-minute result was observed for the defined task on one tested configuration, while performance across other tasks was not established\. That sentence is less marketable than a universal speed promise, but it gives the buyer a usable boundary\. ## Make the next investigation narrow enough to complete List the unresolved condition most likely to change the decision\. It may be a missing plan restriction, an untested file type or a comparator configured differently\. Ask for that evidence before commissioning a broad verdict on the whole company\. When briefing an expert, provide the exact claim, intended use, available records and the decision threshold\. Request a written explanation of the method and limitations, including what the investigator could not establish\. The useful outcome is a justified answer to your question, which may be unfavorable or inconclusive\. ## Method NamedProof editorial guide\. Prepared with AI\-assisted research and synthesis; the cited primary\-source passages were checked for this guide\. This is not a commissioned expert assessment and has not received human expert review\. Examples labelled hypothetical are teaching examples, not client results\. ## Original contribution A claim\-to\-method comparison and a fully labelled calculation distinguishing 20% less time from 25% higher throughput\. ## Limitations - This is an evidence\-evaluation guide, not legal, medical or financial advice\. Regulated claims require appropriately qualified review\. - The numerical example is hypothetical; no product performance or NamedProof customer outcome is asserted\. ## Sources ### F04: FTC Policy Statement Regarding Advertising Substantiation https://www.ftc.gov/legal-library/browse/ftc-policy-statement-regarding-advertising-substantiation Sources checked: 2026-10-17 US FTC policy statement dated 23 November 1984, reasonable\-basis and prior\-substantiation sections\. US advertising\-policy context; this article is a practical evidence guide, not legal advice or a statement of universal legal duties\. Official source read; details in scope and limitations ### F02: Write high quality reviews https://developers.google.com/search/docs/specialty/ecommerce/write-high-quality-reviews Sources checked: 2026-10-17 Google Search Central review\-publishing guidance, updated 10 December 2025\. Recommendations for useful review pages, not a guarantee of ranking or proof that a particular review is accurate\. Official source read; details in scope and limitations ## Related guides - [What is an independent expert assessment? Scope, evidence and deliverables](https://namedproof.com/knowledge/en/independent-expert-assessment) - [Evidence\-based product reviews and user opinions: how to use both](https://namedproof.com/knowledge/en/evidence-based-product-reviews) - [How to verify sources and citations, one claim at a time](https://namedproof.com/knowledge/en/verify-sources-citations) - [Research methods, uncertainty and limitations: how to read a review](https://namedproof.com/knowledge/en/research-methods-uncertainty-limitations) ## Available languages - [English](https://namedproof.com/knowledge/en/evaluate-product-claims)