Marketing Claims Pre-Screen

Run marketing material — pitch decks, one-pagers, website copy, ad creative, campaign emails, product sheets, customer-facing reports — through the common pitfall categories of advertising-substantiation and consumer-protection principles and produce a severity-ranked flag list: the exact language at issue, which pitfall category it hits, and what a legal, compliance, or brand reviewer would likely require before it goes out. The point is to arrive at review with a clean draft instead of burning two review cycles on findable problems.

This is a pre-screen, not legal advice. It does not replace review by your legal team, compliance function, regulatory affairs group, or outside counsel, and a clean pre-screen is not a determination that material is compliant. It exists to make the human review faster and to teach the drafting team what keeps getting flagged.

Why this skill exists

Marketing finishes the launch page Thursday night; legal flags "clinically proven", an unlabeled case study, and "#1 rated" on Friday; the campaign ships Monday. Repeat every quarter — in software, healthcare, consumer products, financial services, professional services, nonprofits, and everywhere else that publishes claims. Most claim findings in routine materials are pattern-level: the same handful of pitfall categories account for the bulk of comment cycles, and regulators and self-regulatory bodies across sectors keep citing the same failures (unsubstantiated claims and results shown without their qualifying context above all). Catching pattern-level issues before the reviewer sees the draft saves the reviewer for actual judgment calls.

Inputs

Required:

• The marketing material: deck, one-pager, website or landing-page copy, ad creative, email campaign, product sheet, proposal, or excerpts (text or file)

Optional but high-value:

• Audience and channel: who receives it, one-on-one sales use vs. broad public distribution, prospective vs. existing customers, consumers vs. businesses (the standards that apply, and how strict they are, depend on this)
• Industry and jurisdiction: some categories — health and medical, financial products, alcohol, children's products, environmental claims, employment, credit — carry sector-specific rules on top of general principles; non-US markets differ. The screen uses general principles and labels where a sector-specific framework would govern
• The organization's existing disclosure pages, footers, terms, or legends, so the screen can check whether required context exists elsewhere in the document
• Prior review feedback on similar materials
• Whether a substantiation file exists for the key claims (the screen can't see it, but knowing it exists changes the likely ask)

Missing-input behavior: If no material is provided, ask once. If audience or industry is unknown, screen anyway under the assumption "broad public distribution to prospective customers, in a category where claims are scrutinized" — the strictest common case — and label the assumption.

How it works

1. Read the full document first, including footnotes, footers, disclosure pages, and linked terms — many apparent problems are cured by a disclosure elsewhere, and the screen must check before flagging.
2. Sweep for the seven pitfall categories:
   • Results without the context that makes them true. Any headline number — savings, ROI, growth, uptime, speed, returns, outcomes — shown without the basis, period, costs, and typical-case context needed to read it accurately. Includes the sneaky variants: "up to 80% savings" where the typical customer sees 15%; a best-case figure in the summary with the average buried in an appendix; a pre-fee or pre-cost figure presented as what the customer actually gets; "illustrative" numbers that are really results.
   • Cherry-picked results and case studies. Selected customers, projects, periods, or metrics without the context a reader needs — case studies showing only winners, "representative customers" that aren't representative, a time period that conveniently starts after a bad stretch, a metric chosen because the others went the wrong way.
   • Projections and hypothetical results presented as outcomes. ROI calculators, modeled savings, forecasts, targets ("targeting 40% cost reduction"), pilot results extrapolated to scale, pro-forma blends — all need their assumptions, criteria, and risks stated, and need to be clearly labeled as projections rather than results.
   • Testimonials, endorsements, and reviews. Customer quotes, expert or influencer endorsements, media mentions, partner statements, star ratings — these raise disclosure questions (was the person compensated or otherwise connected? is the experience typical? was the quote edited?) and, where paid, generally need a clear material-connection disclosure.
   • Unsubstantiated claims and superlatives. "Best-in-class," "#1," "proven," "clinically proven," "the only," "market-leading," "guaranteed," "risk-free," "consistently outperformed," "most secure" — any statement of fact the organization must be able to substantiate on demand. Flag each with what substantiation would be needed.
   • Misleading comparisons. Competitor comparisons using outdated competitor data, apples-to-oranges bases (list price vs. discounted price; their base tier vs. your premium tier), unnamed or undefined benchmarks, "vs. the industry average" with no source.
   • Missing or inadequate disclosures. Results presented without the standard accompaniments (as-of dates, basis, "results vary" or equivalent where applicable, conditions behind the number); third-party awards or ratings without who-gave-it/when/criteria/whether-paid; "free," "unlimited," or "no fees" with undisclosed conditions; regulatory misstatements ("FDA-approved," "government-certified," "regulator-endorsed") where the status is something less.
3. Quote exactly. Every flag includes the precise language and its location (slide/page/section/URL). A reviewer can't act on "somewhere it says something superlative."
4. Rank by severity:
   • High — pattern matches the bright lines that draw enforcement across regimes (unsubstantiated health, safety, or financial claims; undisclosed paid endorsements; results shown in a way that misstates what a typical customer gets; regulatory-status misstatements)
   • Medium — likely to draw comments; fix is usually added disclosure, a source, or reframing
   • Low — judgment-call territory; flag for the reviewer's attention
5. State the likely reviewer ask per flag — the kind of remediation a legal or compliance reviewer typically requires: add the typical-case figure with equal prominence, add the assumptions-and-risks legend, cut the claim or attach substantiation, label the case study and disclose selection criteria, add the material-connection disclosure, name and date the comparison source. Phrase as "a reviewer would likely require…" — not as a legal determination.
6. Note what the screen can't see: whether substantiation actually exists, whether a quoted person was compensated, whether a comparison's underlying data is current, whether sector-specific rules apply. These become questions for the reviewer, listed explicitly.

Output format

# Marketing Claims Pre-Screen — [Document name]
Screened [date] | Assumed: [audience, channel, industry — stated or assumed]
**This is a pre-screen to accelerate legal/compliance review. It is not legal
advice and is not a substitute for review by your legal, compliance, or
regulatory function.**

## Summary
[n] flags: [n] High / [n] Medium / [n] Low. The 2–3 items most likely to
block distribution, in one paragraph.

## Flags (severity-ranked)
### Flag 1 — [HIGH/MED/LOW] — [Pitfall category]
**Location:** [slide/page/section/URL]
**Exact language:** "[verbatim quote]"
**Issue:** [why this pattern draws scrutiny, in plain English]
**Likely reviewer ask:** [the typical remediation]

(repeat per flag)

## Cured-Elsewhere Notes
Potential issues checked and found addressed by existing disclosures [location].

## Questions for the Reviewer
What this screen cannot determine: substantiation files, compensation behind
quotes, currency of comparison data, sector-specific rules in play.

## Clean Categories
Pitfall categories swept with no findings — listed so the reviewer knows the
scan was complete, not selective.

High-frequency flag patterns (the screen's cheat sheet)

Language patterns that account for most routine findings:

• Any %, multiple, dollar savings, ROI, growth rate, or outcome figure → check for basis, period, typical-case counterpart, as-of date, whether costs are included
• "Up to", "as much as", "starting at", "from $X" → is the typical case disclosed with comparable prominence?
• "Target", "expected", "projected", "could save", "potential" + any figure → projection/hypothetical territory
• "Since [date]", "over the last N years" → verify the period chosen and whether a predecessor's results are being imported
• "Select customers", "representative", "case study", "success story" → selection-criteria and full-picture context check
• Quotes from anyone outside the organization → testimonial/endorsement analysis (compensated? connected? typical? edited?)
• "Award-winning", "ranked", "rated", "#1", "top 10" → third-party rating disclosure check (who, when, criteria, paid?)
• Superlatives and absolutes: "best", "top", "leading", "proven", "guaranteed", "consistently", "uniquely", "the only", "no other" → substantiation flag
• "Certified", "compliant", "approved", "registered", "accredited" → fine if accurate and current; "approved/endorsed by [regulator]" when the regulator doesn't endorse → always a finding
• Comparisons to named competitors or "the industry" → source named, current, and apples-to-apples?
• Customer, partner, or press logos on a page → implied endorsement and permission question
• "Free", "unlimited", "no fees", "cancel anytime" → conditions disclosed nearby?
• "Green", "sustainable", "carbon-neutral", "ethically sourced" → specific, substantiated, and not broader than the evidence?
• Health, safety, or efficacy language ("clinically proven", "doctor recommended", "prevents") → highest-scrutiny category; substantiation and sector rules question

Guardrails

• Not legal advice — say it twice. The disclaimer appears in the header and the framing never drifts into "this is compliant" / "this violates the rule." The vocabulary is "matches a pattern that draws scrutiny" and "a reviewer would likely require."
• A clean screen is not clearance. If few or no flags are found, say explicitly that absence of flags is not a compliance determination.
• Quote exactly, locate precisely. Never paraphrase flagged language; never flag language the document doesn't contain.
• Check before flagging: a headline figure with the typical-case context right beside it is not a finding — read the whole document, footnotes and footers included.
• Don't rewrite the claims. Suggest the type of fix; drafting compliant claim language is the reviewer's and the marketing team's call, and this skill never composes substitute figures, statistics, or disclosure numbers itself.
• Never invent rule citations, cases, or enforcement actions. Refer to advertising-substantiation and consumer-protection principles and their known pattern areas generally; if asked for specific cites, recommend counsel confirm.
• Jurisdiction and sector humility: the screen is built around general principles that recur across advertising regimes; sector-specific frameworks (financial promotions, health and drug claims, alcohol, children's products, environmental marketing, employment, credit) and non-US jurisdictions differ — label the analysis as pattern screening, not the applicable framework.

Quality checks

• Disclaimer present in header and conclusion?
• Every flag: exact quote + location + category + severity + likely ask?
• All seven categories swept, with clean categories listed?
• Cured-elsewhere check actually performed (footnotes, footers, and linked terms read)?
• No compliant/non-compliant verdicts anywhere in the text?
• Reviewer-questions section converts the screen's blind spots into a checklist?

Example prompts

"Attaching our new product launch deck before it goes to legal Monday. Enterprise SaaS, deck goes to prospective customers. Slides 8–11 have the ROI case studies and there's a customer quote on slide 14 I'm nervous about. Pre-screen it — exact language, severity, and what legal will probably make us change — so I can fix the obvious stuff this weekend."

"Pasting the new landing-page copy for our supplement line before the redesign ships. Public audience, US consumer. Screen it — especially the 'clinically proven to boost energy' line marketing loves — and tell me where sector rules would take over from general principles."

"Our fundraising team drafted this donor one-pager claiming '100% of your gift goes to programs' and '3x more effective than comparable charities.' Screen it under the strictest assumptions and label the framework caveat."

"Our fund's investor deck goes to compliance Friday — gross returns on page 6, 'top-decile manager' on page 2, and a placement-agent quote on page 9. Pre-screen it and note where the financial-promotions framework would go beyond the general principles."

Works well with

Run this on business-one-pager and stakeholder-update-generator outputs before they leave the building; claim language flagged here usually traces back to answers in questionnaire-response-drafter libraries worth cleaning at the source.