How to Build a Marketing Claim Evidence Ledger

How to Build a Brand-Claim Evidence Ledger

A brand-claim evidence ledger records every material public claim in its exact context, the meaning a reasonable audience could take from it, the evidence available before publication, the supported scope, necessary exceptions or disclosures, any endorsement relationship, the accountable owner, and the date or event that makes the claim stale.

The ledger makes attractive copy reviewable. It does not create a legal safe harbor or turn weak evidence into strong evidence.

flowchart LR
    A["Exact public copy and context"] --> B["Express and implied meaning"]
    B --> C["Claim class and materiality"]
    C --> D["Matching evidence"]
    D --> E["Supported scope and exceptions"]
    E --> F["Approval, owner, and expiry"]
    F --> G["Publish, narrow, hold, or remove"]
    G --> H["Correction propagation"]

Atomize the copy

Begin with the smallest statement that can be reviewed independently. "Fast, safe, affordable, and trusted" is not one claim. It may contain a speed claim, safety claim, comparative price impression, and social-proof claim.

Capture the exact words, image, demonstration, placement, audience, offer, channel, date, price context, nearby testimonial, and linked qualification. A sentence that looks subjective in isolation can create an objective implication beside a graph, before-and-after image, or customer result.

Assign a stable claim ID. Record every route and asset where the claim appears, including landing pages, advertisements, proposals, decks, videos, captions, sales scripts, app-store text, partner materials, and automated messages. This reuse map becomes the correction path later.

Write what the audience could reasonably understand

Do not review only what the writer intended.

The FTC's Advertising FAQs explains that US advertising analysis considers the point of view of a reasonable consumer, express and implied claims, omissions, materiality, context, and whether the advertiser had sufficient support before the ad ran.

For each claim, write the likely audience meaning in plain language. "Up to 50 percent faster" raises questions about the baseline, task, population, conditions, and how often the maximum occurs. "Customers love us" may imply a current and representative basis. "Professional quality" may be subjective in one context and imply specific training, workmanship, or compliance in another.

Ask what fact would change the customer's decision. Claims about performance, safety, price, effectiveness, outcomes, guarantees, qualifications, environmental impact, typical results, or comparisons often deserve heightened review.

Match the evidence to the claim

Record the evidence source, method, date, sample, population, exclusions, limitations, custodian, and current location. Then state exactly which product, service, audience, geography, period, channel, and conditions it supports.

A timestamp can support a defined response-time claim if the start and stop events are clear. It cannot prove a customer's problem was resolved. A controlled product test may support performance under tested conditions. It does not automatically support real-world use outside those conditions. A customer survey may support reported perception for the sampled population. It does not prove the technical cause of the perception.

Evidence should exist before publication for the claim being made. A plan to collect evidence later is a reason to hold or narrow the copy.

The ledger should make unsupported scope expansion visible. If the evidence covers one product version in one period, the claim should not silently extend to every version and future period.

Separate testimonials from substantiation

A testimonial must reflect a real and honest experience, but sincerity does not prove the objective result inside the statement.

The FTC's current Consumer Reviews and Testimonials Rule Q&A discusses fake or false reviews and testimonials, insider relationships, incentives, review suppression, and the distinction between hosting reviews and using testimonials in marketing. The FTC's Endorsement Guides Q&A addresses material connections, disclosure, expert and consumer endorsements, monitoring, and changing circumstances.

For every endorsement, record the person's identity, consent, actual experience, relevant expertise, relationship to the company, compensation or incentive, approval of the final use, date, continuing accuracy, and withdrawal path.

Then evaluate typicality separately. If a testimonial reports an exceptional result, "results may vary" does not establish what customers generally should expect. The evidence owner must support the underlying objective claim and the expected typical result or use appropriately limited language after qualified review.

Record disclosures as part of the claim

A disclosure should communicate information the audience needs to avoid a misleading impression. Record its exact text, placement, format, prominence, timing, language, accessibility, and the devices or media tested.

The FTC's Disclosures 101 emphasizes that material connections should be obvious and hard to miss. A buried profile note or a disclosure requiring the audience to click "more" may not perform the same job as language next to the endorsement.

A disclosure should clarify rather than contradict. "Guaranteed to save 30 percent" cannot be repaired by a distant statement that no savings are guaranteed. Narrow the headline to what the evidence supports.

Accessibility matters to whether people can perceive and understand the qualification. The W3C's WCAG 2.2 Recommendation includes requirements concerning text alternatives, color, contrast, reflow, labels, focus, language, error identification, and status messages. Claim review should include the media and interaction through which the disclosure appears.

Assign an owner, expiry, and correction path

Every claim needs an accountable business owner with authority to hold, narrow, correct, or remove it. Review roles may include editorial, technical, research, privacy, accessibility, regulatory, commercial, and legal owners depending on the claim.

Set an expiry date or refresh trigger tied to the evidence. Triggers may include a product change, new service process, pricing change, market comparison change, customer population change, new research, complaint pattern, endorsement withdrawal, regulatory change, or correction to a cited source.

When a claim fails review, update every reuse location. The correction record should show what changed, why, when, who approved it, and which assets remain outstanding. A canonical page correction is incomplete if the old claim remains in an advertisement, video description, sales script, or partner deck.

Use four decisions

The ledger should end with a decision to publish, narrow, hold, or remove. "Needs more research" without an owner and date is not a decision.

Publish only when evidence and review support the likely meaning. Narrow when the support is real but more limited than the copy. Hold when evidence, consent, disclosure, or review is missing. Remove when the claim is false, stale, misleading, unauthorized, or cannot be corrected in context.

[[How to Differentiate a Service Business Without Empty Claims]] identifies candidate claims from customer and market evidence. [[How to Turn a Brand Promise Into an Operating System]] establishes whether the operation can deliver them. E095's [[How to Audit an Online Pricing Study]] shows why method and denominator control a public conclusion. E096's [[When Is Variable Pricing Fair]] adds the customer and governance questions around price language. E103's [[How to Evaluate Predictive AI Marketing Claims]] applies the same discipline to model performance.

This guide was developed with AI assistance from the preserved E022 interview, the linked ledger schema, and current FTC, SBA, USPTO, NIST, and W3C sources. Dalton Anderson remains the author. It provides United States advertising issue spotting, not legal advice, claim approval, testimonial clearance, accessibility certification, or a substitute for qualified review. Editorial, research, advertising, legal, accessibility, and founder review are required. Publication is not authorized.