How to Review a Press Release Draft for Unsupported Claims

By Kevin Mahoney

Before a press release goes anywhere near a distribution network, read every sentence and ask one question: “If a regulator, a journalist, or a skeptical customer asked me to prove this, could I?” If the answer is no, the claim comes out or gets qualified. That's the entire review in one sentence. The rest of this article is the checklist I actually use to get there, plus a claim-to-evidence grid you can run your own drafts through before they leave your desk.

Why This Review Step Exists

A press release is advertising copy wearing a news hat. The Federal Trade Commission treats it that way too: advertisers need a “reasonable basis” for a claim before it runs, meaning objective evidence that supports what's being said (FTC, Advertising FAQs: A Guide for Small Business).

There's a second, separate risk that has nothing to do with the FTC: search engines. Google's spam policies define “misleading functionality” as intentionally creating pages that trick users about what they'll get, and treat deceptive claim patterns as something that can get a page suppressed or a site penalized, independent of any legal exposure (Google, Spam Policies for Google Web Search).

So the review isn't a courtesy pass. It's the step that decides whether the release is safe to distribute at all — and none of what follows replaces legal review on a specific claim. It's the baseline check before a release goes further, not a substitute for counsel when a claim is genuinely borderline.

The Claim-to-Evidence Grid

I run every draft through the same exercise: pull out each factual claim, name what kind of claim it is, and write down exactly what evidence would need to exist to support it. If that evidence doesn't exist in hand, the claim doesn't ship as written. Here's the grid, organized by the claim types that show up most often in press release drafts.

• Superlative claims (“best,” “#1,” “leading,” “top-rated”). What would support it: a named, checkable ranking source (an independent list, a market-share figure, a review aggregator score) current as of the release date. Without a citable source behind it, a superlative is an opinion dressed as a fact, and it's one of the easiest things for a reader — or a search engine — to flag as unsupported.
• Comparative claims (“more effective than,” “outperforms,” “faster than”). What would support it: a head-to-head comparison using the same test conditions for both products, ideally from a source independent of the company making the claim. Comparative advertising itself is legal as long as it's truthful (FTC, Advertising FAQs); what makes it defensible in practice is that the comparison is apples-to-apples.
• Health and medical claims (“supports immune health,” “reduces inflammation,” “clinically proven to…”). What would support it: “competent and reliable scientific evidence” — testing conducted and evaluated by qualified experts, using methods generally accepted in that field. The FTC has been explicit that this bar applies even to claims that never use the word “disease”: a structure/function claim can still imply a disease benefit through context, product naming, or imagery, and gets held to the same standard once it does (FTC, Health Products Compliance Guidance).
• “Clinically proven” / “clinically tested” language specifically. What would support it: an actual clinical result, not just a clinical test. The FTC's guidance is blunt on this point — the phrase “clinically tested ingredient” implies not just that testing happened, but that the results proved a benefit. If the study exists but was inconclusive, or tested something adjacent to what the release claims, the phrase doesn't belong in the draft.
• Results and outcome claims (“increased visibility by X%,” “generated leads within days,” “achieved top rankings”). What would support it: a specific, attributable data source you can point to — first-party reporting, a named study, or a verifiable case result published with the client's permission. Vague outcome language with no source is one of the fastest ways to turn a release into an unsupported promise.
• Testimonial and endorsement claims. What would support it: the endorsement itself, plus a disclosure adjacent to it whenever the endorser has a financial or business relationship with the company being described. A glowing quote with no disclosure, sitting next to a specific results claim, compounds two problems into one paragraph — this is the territory the FTC's Endorsement Guides (16 CFR Part 255) govern.

For each row where the evidence doesn't exist yet, there are three honest options: cut the claim, soften it to what the evidence actually supports (a documented process instead of a promised outcome, for example), or hold the release until the evidence is sourced. There's no fourth option where the claim stays as written because it “sounds true.”

Applying It to Superlatives

Superlatives are the most common flag I catch, because they're the easiest words to reach for under a deadline. “Leading provider,” “top choice,” “most trusted” — none of these are illegal on their own, but every one of them is a claim that requires backing the moment it's published. Before it survives a review, it needs an answer to: leading by what measurement, according to whom, as of when? If there's no clean answer, the fix is almost always a rewrite toward something specific and true — a stated year founded, a stated number of clients served, a stated service area — rather than a ranking word nobody can source.

Applying It to Medical Claims

Medical and health claims get the strictest read, because the evidence bar is the highest and the audience is the most vulnerable to being misled. The review question here isn't just “is this true” — it's “does this imply more than the evidence shows.” A supplement release that says a product “may support joint comfort” is making a narrower claim than one that says it “relieves joint pain,” even though both might describe the same product. The second version needs clinical evidence behind it that the first doesn't. When in doubt, the draft goes back to the narrower, better-supported version — never the other direction.

What This Looks Like in Practice

On the production side, this is exactly the work: a search-led plan for the release, purposeful content built to survive this kind of review, distribution management once it's cleared, and reporting the client can actually check afterward. What that process doesn't include is a promise about what happens after the release goes out. Downstream pickup, indexing, search position, and any resulting traffic, leads, or sales aren't things any release — reviewed or not — can guarantee, and a paid distribution placement is not the same thing as earned editorial coverage a publication chose to run on its own judgment. The full terms are public if you want the exact language: MBK's press release service terms.

None of that changes if you're not working with an agency at all. The same grid works whether you're reviewing your own draft, a freelancer's draft, or something a wire service sent back for revision. The standard doesn't move based on who wrote it.

The Next Practical Decision

If you're staring at a draft right now, the fastest path forward is this: list every factual claim in the release on one page, run each one through the grid above, and mark it green (evidence in hand), yellow (needs a rewrite to match what's actually supportable), or red (cut it). A release with a few yellow marks and zero red marks is close. A release with any red marks doesn't go out until those are resolved — not because a rule says so, but because an unsupported claim is the one thing in a press release that can outlast the release itself, in a way that's much harder to walk back than it was to avoid.